Noname manuscript No.

(will be inserted by the editor)

Automated Bank Cheque Verification Using Image

Processing and Deep Learning Methods

Prateek Agrawal · Deepak Chaudhary ·

Vishu Madaan · Anatoliy Zabrovskiy ·
Radu Prodan · Dragi Kimovski ·
Christian Timmerer

Received: date / Accepted: date

Abstract Automated bank cheque verification using image processing is an

attempt to complement the present cheque truncation system, as well as to
provide an alternate methodology for the processing of bank cheques with min-
imal human intervention. When it comes to the clearance of the bank cheques
and monetary transactions, this should not only be reliable and robust but
also save time which is one of the major factor for the countries having large
population. In order to perform the task of cheque verification, we developed
a tool which acquires the cheque leaflet key components, essential for the

Prateek Agrawal
University of Klagenfurt, Austria
Lovely Professional University, India
E-mail: [email protected]
Deepak Chaudhary
Lovely Professional University, India
E-mail: [email protected]
Vishu Madaan
Lovely Professional University, India
E-mail: [email protected]
Anatoliy Zabrovskiy
University of Klagenfurt, Austria
Petrozavodsk State University, Russia
E-mail: [email protected]
Radu Prodan
University of Klagenfurt, Austria
E-mail: [email protected]
Dragi Kimovski
University of Klagenfurt, Austria
E-mail: [email protected]
Christian Timmerer
University of Klagenfurt, Austria
Bitmovin Inc., CA, USA
E-mail: [email protected]
2 Prateek Agrawal et al.

task of cheque clearance using image processing and deep learning methods.
These components include the bank branch code, cheque number, legal as well
as courtesy amount, account number, and signature patterns. our innovation
aims at benefiting the banking system by re-innovating the other competent
cheque-based monetary transaction system which requires automated system
intervention. For this research, we used institute of development and research
in banking technology (IDRBT) cheque dataset and deep learning based con-
volutional neural networks (CNN) which gave us an accuracy of 99.14% for
handwritten numeric character recognition. It resulted in improved accuracy
and precise assessment of the handwritten components of bank cheque. For
machine printed script, we used MATLAB in-built OCR method and the ac-
curacy achieved is satisfactory (97.7%) also for verification of Signature we
have used Scale Invariant Feature Transform (SIFT) for extraction of features
and Support Vector Machine (SVM) as classifier, the accuracy achieved for
signature verification is 98.10%.

Keywords Cheque Truncation system · Image Segmentation · Bank Cheque

Clearance · Image Feature Extraction · Convolution Neural Network · Support
Vector Machine · Scale Invariant Feature Transform

1 Introduction

Cheques have an intriguing past, as evidently, it is still a mystery as to when

exactly the concept of cheque has emerged. Speculations were made that the
idea in itself was entertained during the Roman empire, however the very idea
could not catch on. Thus, its usage came in the business hubs in the 1500s
in regions near Holland and Amsterdam because of their richness in business
opportunities during that time. Therefore, with its gradual acceptance in the
commerce and enhanced usability it was slowly accepted as a means of mone-
tary transaction. The word ”cheque” was evidently originated in the 1700s as
a common practice for monetary transactions, certain unique numbers were
written on the piece of papers which were checked to keep track of and to pro-
ceed with the transaction. Financial transactions using cheques is still one of
the widely accepted way across the globe. Almost in every continent, banking
institutions have provision for their customers to use cheques for any financial
transaction. In 2018, people in United Kingdom used 346 million cheques for
payments and to acquire cash with totaling a value of £442 billion [1]. Few
sample cheques of different countries (Malaysia, Maldives, USA, Kenya, UAE,
and China etc.) are shown for reference purpose in Fig. 1:
Title Suppressed Due to Excessive Length 3

Fig. 1 Sample cheque images of different countries [2] [3] [4].

Although cheque systems in banks for transactions is one of the most pow-
erful tool, primarily due to its slow processing time, considerable complexities
in cheque clearing process and gradually with an immense growth in the tech-
nology, the electronic-transactions have created a big shadow on the cheque
system to cater the need of the banking sector. Therefore, we have proposed a
mechanism to make the entire process of bank cheque verification not only fast
but also reliable by using deep learning methods and image processing tech-
niques. Due to various technical constraints, it was not feasible for the banking
sector to implement the automation of the bank cheque clearing which required
sophisticated measures for authentication and verification for the transfer of
4 Prateek Agrawal et al.

Fig. 2 CTS-2010 compliant sample bank cheque

funds from accounts. Therefore, as a use case example, we have implemented

and tested our proposed model on cheque samples of different Indian banks
which follow the standard guidelines of the Reserve Bank of India (RBI). Be-
fore Jan 2019, the security features used in cheques at the country level by all
Indian banks were lacking technical and security edge such as non-availability
of void pantograph, Rupee symbol, micro-lettering, bank logo in Ultra Violet
(UV) ink etc. However taking a major step in modernization of cheque based
transactions, RBI made it mandatory to use the CTS-2010 cheque format as
a standard format for all regulated banks [5]. There are certain mandatory
security features in the CTS-2010 compliant cheques as shown in Fig 2 and
are mentioned below:

1 Bank Logo in UV ink

2 Printer name with CTS-2010

3 Void Pantograph

4 Micro-lettering

5 Date

6 Rupee Symbol

7 Signature

8 Account Number
Cheque Truncation System (CTS) is a way to authenticate the cheque
clearing process. The process flow of CTS is shown in Fig. 3. There are two
steps authentication of any bank cheque leaflet / form before making any fi-
nancial transaction using cheques. First step is to validate whether the leaflet
used, is genuine or fake. Deepak at al (2019) [6] proposed some methods to
validate the bank cheque leaflet. Second step to authenticate the cheque is
to verify about the details filled on that leaflet. In our research work, we fo-
cused on proposing some novel methods to automate the verification of the
bank cheques. For this, we extracted the key information from the cheque
leaflet which were required for verification of bank cheque and represented
Title Suppressed Due to Excessive Length 5

these features in the highlighted using bounding boxes. There are various tech-
niques already available for the identification, authentication and verification
of the bank cheques as shown in the research works presented by Dhanva et
al (2018) [7], Singh and Tigga (2012) [8], Gonge and Ghatol (2014) [9], Gonge
and Ghatol (2016) [10], Menon et al (2017) [11]. Unfortunately, the efficiency
of the works are not up to mark. For the task of verification, we considered
matching of Indian Financial System Code (IFSC), cheque number, account
number, legal amount, courtesy amount and signature of cheque issuer as input
components.

Fig. 3 Process flow diagram of proposed work.

Optical Character Recognition (OCR) is an automated text reading pro-

cess from images that includes machine printed (typographic) or handwritten
character recognition and numeral recognition. Image processing is a process
of analysing a digitised image for performing operations such as enhancement,
compression, segmentation, or editing. The fundamental stages of image pro-
cessing which almost all the image has to go through include the following
6 Prateek Agrawal et al.

components which are sequential in nature: acquiring the image, pre-processing

of the image, segmenting the image, interpreting the image, and recognising
the image. There are many ways to interpret any image. One effective way is
using convolution neural networks (CNN). The deep learning approach is an
advancement of machine learning with multiple layers allowing it to achieve
higher level of accuracy for pattern recognition. However, it demands a large
number of datasets for the desired training efficiency of the network. If we have
a enough large database, its implementation will assist in precise estimation of
the handwritten characters so as to give accurate algorithmic implementation
ability for the bank cheques. This may benefit the banking sector by reducing
the processing time and saving resources by automating the task of cheque
clearance. We implemented our research work using MATLAB-2018a.
The major contribution of our research was to identify the various possible
techniques to verify the bank cheque and then apply the best selected image
processing and deep learning methods to improve the performance of auto-
mated verification in terms of accuracy. After that, we compared our results
with existing work and showed the comparative analysis on how our innova-
tion covers all primary aspects to verify a bank cheque in a more efficient and
effective way in terms of accuracy. We describe the major contribution of our
innovation in the following steps:
– Captured the following key components of a scanned bank cheque image
using image segmentation: Date, Account Number, Cheque Number, Sig-
nature of account holder, Courtesy amount detail, legal amount detail,
associated bank information (IFS Code)
– Validated bank branch information, genuineness of the cheque and cus-
tomer account number by applying OCR method on IFSC segment, cheque
number segment and account number segment respectively.
– Obtained the pattern of signature and match it with the specimen signa-
tures of account holder using SIFT and SVM methods.
– Recognised the characters stated in the legal amount segment of the cheque
using CNN method and convert it into the place value system.
– Recognised the digits stated in the courtesy amount segment of the cheque
using CNN and verify it with the extracted and converted legal amount
information using the proposed algorithm.
This paper is organized as follows: Section 2 summarises the related work.
Section 3 presents the complete methodology and proposed algorithm along
with the models for verification of bank cheques. Section 4 presents the ex-
perimental results and analysis of our implemented model. Finally, section 5
concludes the research work and outlines future work.

2 Related Work

In order to perform the task of getting a logical response for the input of an
image, it requires a meaningful interpretation of the characters present in the
Title Suppressed Due to Excessive Length 7

image and, thus, we need to perform the character recognition for the various
patterns present in an image. Singh and Sachan (2018) [12], Wankhede and
Mohod (2017) [13], and Ramanathan et al. (2009) [14] presented their work to
retrieve the characters from the image Optical Character Recognition (OCR).
The segmentation of digits for which there are several techniques devised is
the major challenge faced in the recognition of characters. Chen (2017) [15]
proposed a general approach for handwritten digit segmentation using spectral
clustering and achieved an accuracy of over 97%.
One of the earliest researches on font recognition was done by Dickinson
(1960) [16], Liu and Shelton (1966) [17], Nagy and Shelton (1966) [18], Wang
et al. (1972) [19], Zramdini and Ingold (1998) [20] in which they suggested
a statistical approach for the Latin Language font recognition at the text
line level depending on the global typographical features. They were able to
achieve approximately 97% of recognition rate. For English language also, due
to its widespread usage and commercial usability, Zhu and Wang (2002) [21],
Dhande and Kharat (2017a) [22], Pal et al. (2006) [23], Dhande and Kharat
(2017b) [24], and Saenthon and Sukkhadamrongrak (2014) [25] proposed few
approaches for character recognition. Font recognition was done by using a
wide array of techniques and they were able to achieve a satisfactory accuracy
on text blocks.
Similarly, Ramanathan and Sharma (2017) [26] devised a method using
Support Vector Machine (SVM) for classification and feature extraction by
proposing a SVM-based multi-knowledge-based system, achieving an accuracy
rate of 93.54%, which marked the continued focus on the OCR or segmentation
for the character recognition in image processing methods. However, Stewart
et al. (2018) [27] proposed a segmentation free recognition techniques in which
they presented that for small training datasets, it is possible to manipulate
either manually or automatically that segmentation and stitching improves
handwriting recognition on dataset with few samples. In another attempt by
Peres et al. (2018) [28] in 2018 proposed a novel handwritten methodology in
which they used a hybrid CNN and SVM classifier with which they were able
to achieve an overall accuracy of 96.7% using the MNIST dataset [29].
For intelligent character recognition to create and match dynamic tem-
plates as well as the conversion of characters into text, Kajale et al. (2017) [30]
implemented the supervised machine learning methodology and achieved the
accuracy close to 95% for the alpha numerals and special characters. The pres-
ence of the noise in the form of lines, watermarks, diagrams, patterns or all of
these at the background is one of the important factors to create the perfor-
mance variation in different character recognition techniques. Goto and Aso
(2002) [31], Brisinello and Gibric (2018) [32], Brisinello et al. (2017) [33], and
Feng and Gao (2010) [34] performed research involving extraction of the infor-
mation from the colored or noisy background with satisfactory level of output.
Contrary to the notion that output will drop significantly if there is a change
in the background with varying levels of color change in the texts or intro-
duction of noise in the form of patterns and figures. Jia and Wei (2010) [35]
extracted the legal amount from the Chinese bank cheques and proposed a
8 Prateek Agrawal et al.

multiple channel binarization approach for extracting legal amount from color
Chinese bank check images with a combination of global and local threshold.
Sneha Singh et al. (2015) [36] proposed a method to recognise the legal amount
of Indian bank cheques using correlation coefficient method. They achieved an
average 76.4% word recognition accuracy on a self created dataset of 19215
word samples of 61 different words. Another researchers M Badrul Alam et
al. (2015) [37] developed a tool to recognise the handwritten legal amount
of Bangladesh bank cheques using ANN and achieved only 93.4% accuracy.
Similarly, Jayadevan et al. (2010) [38] proposed a lexicon-driven segmentation
recognition technique for the recognition of the legal amount words from the
Indian bank checks written in English using Modified Quadratic Discriminant
Function (MQDF) and claimed 97.20% accuracy.
Raghevendra and Danti (2015) [39] proposed a method to identify the bank
details associated with the cheques using bank after removing the background
noise. Researchers used the invariant geometrical features for accomplishing
the recognition of the checks with 93.8% of accuracy only. To improve the
security for verifying the cheque, another important feature of it is Magnetic
Ink Character Recognition (MICR) code. Even though if we are able to per-
form the pattern recognition as well as the legal amount recognition using the
above stated techniques one more component of a bank check that needs to be
properly identified for the sake or security as well as determining the validity
of the check is the Magnetic Ink Character Recognition (MICR), Shirai et al.
(2012) [40] proposed a three pronged solution to the problem starting with
the character color estimation and conversion into binary form to remove the
background, to separate each character using the character pitch estimation
and the removal of the unwanted noise & extraction of the required MICR
code using the sign fragments removal by the stroke tracing.
Lu et al. (2017) [41] proposed techniques to deal with noisy images as the
image that we have, may be a copy of the original image with some consider-
able noise such as shadow along with a spatial noise due to improper copying
of this image. Researchers used local adaptive threshold algorithm in order to
transform the gray-scale image into the binary image to capture the contours
of the texts in the image and applied a double-filtering algorithm to remove
the noise between the texts and the characters. Wiraatmaja et al. (2017) [42]
proposed a deep convolutional de-noising auto-encoder and claimed the im-
provement in accuracy by 26.78% as compared to a simple OCR in average
accuracy testing. Void Pantograph is another important feature which ensures
whether an original image of the cheque was taken for verification or not. It
refers to the process of creating a pantograph which is otherwise invisible to
naked eyes on the original print. However, when copied with a printer, it is
clearly visible making it a very important security features for security print-
ing toolkit. Aronoff et al. (2011) [43] formulated a method to automate the
optimization of void pantograph settings. Peymani and Soryani (2017) [44]
proposed a model to identify machine generated as well as hand written char-
acters for Farsi language images with an accuracy of 68% using an Arabic
language dataset [45]. Bhadwal (2019) [46], Singh (2018) [47] and Agrawal
Title Suppressed Due to Excessive Length 9

(2012) [48] also developed different applications ranging from object recogni-
tion to edge detection by using image processing and computer vision methods.
Ignat and Aciobniei (2016) [49] proposed method of feature extraction for of-
fline handwritten digit recognition by using rotation and filtering of edges for
the extraction of the digit characteristics. KNN and SVM classifiers were used
for classifying the features and claimed to achieve the accuracy about 99% on
the MNIST dataset.

Hou et al. (2017) [50] implemented a model using convolutional neural Net-
work for learning the handwritten numeric characters and perform scratchy
or non-scratchy decisions with error rate of 0.40% on MNIST database. Li et
al. (2018) [51] used grab cut segmentation based on image region by consid-
ering image texture and boundary information which helps them to perform
segmentation with minimal user interaction. Mityakov et al. (2017) [52] in ap-
plication of modern architectures of deep neural networks for solving practical
problems explained the advent of sophisticated networks like GoogleNet [53]
capable of automatic feature extraction and classification. Evidently watershed
segmentation is one of the most widely used algorithm in terms of achieving
feasible results however in cases of adhesion among components of the image,
it becomes difficult to obtain remarkable results. Ji et al. (2015) [54] proposed
an improved watershed algorithm based on distance transform for cell image
segmentation.

Shenoy et al. (2016) [55] proposed a performance based string matching

algorithm for Nvidia GPUs in which they are using Quick Search, Horspool
and Brute Force string matching approaches allowing a performance boost.
Whereas, Joshi et al. (2017) [56] proposed modified Rabin-Karp method by
parallelizing conventional Rabin-Karp algorithm. Zhou and Gao (2017) [57]
used double hash in order to perform multi-pattern matching and are able to
reduce the number of comparison by two hashes thus giving a slight perfor-
mance boost. Consequently all these approaches worked towards reducing the
space-time complexity and improving the performance for string matching.
Dhande & Kharat [22] [24] proposed a method to perform pattern recogni-
tion using vertical and horizontal projection methods along with convex hull
algorithm for feature extraction for cursive English handwritten characters
and implemented support vector machine (SVM) on CCC dataset [58]. Simi-
larly, Dash et al. (2015) [59] proposed a perimetric complexity based on gestalt
principle to develop a handwritten numeral recognition system for obtaining a
point-wise correspondence from the matching of shape context based on affine
transformation on a dataset of Odia numerals. Li et al. (2010) [60], Shen et al.
(2012) [61], Shirazi et al. (2017) [62], Kalaiyarasi and Karthikeyan (2014) [63],
Pornpanomchai et al. (2015) [64], Sahbi et al. (2013) [65], and Kaur (2016) [66]
proposed various approaches to recognize of logos present on the cheque leaflets
for identifying the issuing bank of the bank cheque.Few remarkable signature
recognition methods are highlighted as shown in table 1:
10 Prateek Agrawal et al.

Table 1 Methods of Signature Recognition.

Authors Techniques Datasets Efficiency
SVM classifier
based on RBF CASIA
Naseer and Dogru 98.75%
Kernel using dataset
(2017) [67]
SIFT
SVM classifier
based on RBF CASIA
96.25%
Kernel using dataset
SURF
HMM OnOffSignHindi-75 96.78%
Ferrer et al. (2018) [68]
SVM OnOffSignHindi-75 99.69%
Probabilistic Divergence
Mondal and Measure-
Self-Created 96.50%
Kundu (2017) [69] Kullback-Leiber
Divergence
Local Binary
Bangla Offline
Pattern Features
Jadhav and Chavan Signature 75.34%
using K-Nearest
(2018) [70] Dataset
Neighbor algorithm
Local Binary Pattern
Bangla Offline
Features using
Signature 90.36%
Support Vector
Dataset
Machine algorithm
A comparative
study of various
Jain and Patil
methods for offline N/A N/A
(2014) [71]
signature
verification
Handhyani et al
GLCM using Bootstrap Self-Created 100%
(2017) [72]
Local Directional
Morocho et al Patterns using Biosecure ID
96.10%
(2017b) [73] Support Vector Corpus
Machines
Discrete Wavelet
Transform using Self-Created 99.40%
Beresneva et al
SVM
(2018) [74]
Discrete Wavelet
Transform using Self-Created 98.44%
KNN
Discrete Wavelet
Transform using Self-Created 99.41%
Random Forest
Histogram of Templates
Morocho et al using Artificial BioSecureID
95.28%
(2017a) [75] Immune Recognition Dataset
System
Boudamous et al Crowdsourcing and
(2017) [76]; Gonge and Human-assisted Benchmark Dataset 99.6%
Ghatol (2014, 2016) [9] [10] Schemes
Karanjkar and Vasambekar
Artificial Neural Networks Self-Created 85%
(2016) [77]
CNN based model
MCYT-75, CEDAR,
Hafemann et al. and handcrafted
GPDS-160 and N/A
(2019) [78] feature extractor (CLBP)
the Brazilian PUC-PR
SVM : Linear, RBF
MCYT-75, CEDAR,
Hafemann et al. Model Agnostic
GPDS-160 and 94.84%
(2020) [79] Meta Learning (MAML)
the Brazilian PUC-PR

If we summarise the state of the art, many researchers worked on image

segmentation, character recognition and signature recognition separately and
also achieved remarkable performance. Few researchers considered bank cheque
validation as their research application but none of the researcher proposed any
complete solution to successfully implementing the automated cheque clear-
Title Suppressed Due to Excessive Length 11

ance process without any human intervention. While in our research, we pro-
posed the solution starting from pre-processing of scanned cheque image to
verification of cheque amount and signature. In further section, we described
the entire methodology to implement the model and presented the results
which are found satisfactory.

3 Methodology

The bank cheque verification tool required fulfillment of some key stages in
sequential manner. At first, the tool ensured the genuineness of IFSC present
on cheque then it verified the cheque number to ensure whether the cheque
was from the set of cheque leaflets assigned to the account holder or not.
After that, the amount was verified by checking whether the customer was
having sufficient amount in his account for which he released the cheque, and
finally the signature of the cheque issuer(s) were verified. Only after the com-
pletion of these necessary verification steps, the cheque clearance was passed
for transfer or withdrawal of funds.

Table 2 Bank cheque verification key features and their description.

Key feature Description
- Indian financial system code.
- It is an eleven character code where first four characters represent
IFS code
bank name and last six characters represents the branch. Fifth
character is kept reserved for future use.

- A unique number assigned to the user of the bank services, used for
Account number
monetary transaction related services.

- A six digit number on the bottom left corner of a cheque, it helps

Cheque number to identify a cheque and also verify if the cheque is from the list
of issued cheques to the user.

- The amount mentioned on the cheque leaflet in handwritten /

Legal amount
printed words.

- The amount mentioned on the cheque leaflet in handwritten /

Courtesy amount
printed numeric digits.

- Signature of the account holder.

- Every person has a unique signature which depends on the static
Signature
or dynamic features and it is hard to copy it.
- It is used as a key feature in authenticating a user’s identity.

The entire process of cheque clearance happened in a sequential manner

and each step was mandatory for the next step. All the key features presented
in the cheque verification are shown with their description in Table 2.
Different methods were used for extracting such information from the
cheque leaflet that depended on the viability and level of accuracy. For exam-
12 Prateek Agrawal et al.

ple, to extract information from the machine typographic script, we use OCR
method as it tends to produce results with higher level of accuracy whereas
for hand-written words and numerical values, we can use deep learning based
CNN. Similarly, for authenticating the signature(s), we extracted the features
using SIFT algorithm and for classification, SVM produced optimum results.
To implement these methods, we required noise free image therefore the better
idea was to segment the image so as to acquire only the relevant information.

Fig. 4 shows the broad steps using a flow-chart, involved in our research
process of cheque verification. These steps encompassed certain sub-steps each
of which were essential to meet the standards so as to extract the required
information from the bank cheque. The entire methodology for bank cheque
verification is summarized in Fig. 5.

Fig. 4 Flow Chart for Bank Cheque Verification.

Title Suppressed Due to Excessive Length 13

Fig. 5 Block diagram of bank cheque verification process

3.1 Image Acquisition

The image acquisition of a bank cheque is crucial for the CTS. Generally,
flatbed scanners are used to acquire such images. Due to orientation and ir-
regularities presented in the scanned image(s), we were unable to use the
acquired image(s) directly for the image processing operations therefore it
required some pre-processing steps.

Fig. 6 Raw scanned image of a sample cheque.

14 Prateek Agrawal et al.

3.2 Image Pre-processing

We used the cheque scanned images for our research. A sample scanned image
is shown in Fig. 6. As a scanned image obtained from the scanner can not be
used directly thus we needed to pre-process it, which involved two primary
operations, i.e, rotation and removal of unnecessary background information.
In this step, we rotated the scanned image with respect to the ‘Date Box’
(which is a common feature presented at the same part of every bank cheque)
and then removed the background noise and extra information. Efficiency of
the parameter identification considerably improves with removal of extra back-
ground information.

3.2.1 Rotation

As the scanned images may vary in terms of orientation therefore we used the
date box present in all standard cheque leaflets and the relative invariant na-
ture of the position was utilized. In order to perform the rotation of image, we
determined the point of rotation and degree of rotation. The primary compo-
nent essential for rotation to work was contour extraction, as we were able to
determine the position values of date box by using it which used as our anchor
for any set of operation related to length mapping. The process of contour
extraction was explained in the forthcoming segmentation sub-section. Also,
in order to perform the rotation, we used the midpoint of the image as the
rotation point, and we used date box in order to determine the angle necessary
for the rotation. Fig. 7 shows the same process.
Title Suppressed Due to Excessive Length 15

Fig. 7 Possible scenarios for rotation of cheque leaflet.

As shown in the figure among the possibilities for rotation, we represented

each quadrant with a, b, c, d and highlighted the quadrant consisting the
contour of date box with red color. We performed rotation by 90◦ , 180◦ and
270◦ depending upon the quadrant in which the date box is present. In Fig 7,
the scenario the order of rotation for possibilities was 270◦ , 180◦ and 90◦ . Once
the cheque orientation became correct, we proceed with the removal of unnec-
essary region available in the scanned cheque leaflet.

3.2.2 Removing Background Noise

There was lot of extra information in the scanned image which was to be
removed. For the task of removing of background information, we used the
date box present in the standard cheque template. The cheque dimensions of
any bank in a country are standard and are followed by the norms decided
by the government financial body. In India, RBI monitors and controls the
16 Prateek Agrawal et al.

functioning of all Indian banks. A sample cheque leaf with it’s standard di-
mensions regulated and decided by RBI, is shown in Fig. 8 [80]. All Indian
banks are bound to follow these standards and geometric dimensions strictly.
These geometric dimensions may vary for banks of different countries and can
easily be calculated as per the respective country agency standards. Using the
date box coordinates range and with the implementation of length mapping
between the date box and the whole cheque image, we determined the range of
the cheque as per the schematic representation of edge removal as mentioned
in Fig. 9. In reference to data box, we extracted the regions of interest for
courtesy amount, legal amount, signatures etc.

Fig. 8 CTS-2010 cheque form specification as per RBI guidelines [80].

Title Suppressed Due to Excessive Length 17

Fig. 9 Schematic representation for edge removal.

In the schematic representation, the black rectangle represents the scanned

image of the bank cheque whereas, the red rectangle represents the edge of the
cheque leaflet and the green rectangle represents the edge of the date box. The
purpose of the edge removal task was to remove the region between the red
and black rectangle and to retain the region in the red rectangle only. We used
following mathematical equation for length mapping of bank cheque leaflet as
represented below in Eq. (1).

(yA −yD )

AD
(AD+a) = y2 −yD



 AD = A −yD )
 (y
(AD+c) yA −y1
AB (xB −xA ) (1)


 (AB+d) = xB −x1
 AB = (xB −xA )

(AB+b) x2 −xA

Further the equation is simplified as Eq. (2).

 ((yA −yD )×(AD+a))

y 2 = (
 AD + yD )
y = (y − ((yA −yD )×(AD+c)) )

1 A AD (2)
x1 = (xB − ((xB −xAAB

)×(AB+d))
)
((xB −xA )×(AB+b))

x2 = ( + x )

AB A

The parameters involved in the equations are explained in the Fig. 9. Here,
dimension AB and AD are representing the length, breadth of the date box.
Whereas, a, b, c, d are the distance from the edges of the cheque leaflet. Also,
x and y are the horizontal and vertical coordinates of the date box. The key
parameters of these equations are the boundary of cheque leaflet which is the
coordinate range for unnecessary background information removal.
18 Prateek Agrawal et al.

3.2.3 Grayscale Operation

After removing the background noise of an RGB (a three channel) image, we

converted it into grayscale (a single channel) image. The mathematical formula
for this conversion is represented in Eq. (3) [81].

Gray Image = 0.2126 × R + 0.7151 × G + 0.0721 × B (3)

Here RGB represents the Red, Green and Blue channels of the image and
Gray Image represents the pixel value of the grayscale image.

3.2.4 Gaussian Filtering

After converting the image into grayscale format, next pre-processing step was
to remove the noise. For this, we applied Gaussian de-noising method which is
used in a linear form of smoothing filter for a signal. This method of filtering
is a form of convolution between the normal arrangement of values and the
image. The formula for 2-D Gaussian function is explained in Eq. (4).
2
−(x +y 2)
1
G(x, y) = ( 2
) × e 2σ2 (4)
2πσ
Here G(x, y) is the normal arrangement of the values. Whereas, σ repre-
sents the standard deviation of the arrangement.
We used the binary image format of the gray-scale image for precised and
accurate contour extraction, i.e., we converted the image into binary format
and then determined the boundary points for the region using the formula as
described in the Eq. (5).


f (i, j − 1) = 0

f (i, j) = 1
(5)

f (i, j) ≥ 1

f (i, j + 1) = 0


In this equation, f (i, j) function for binary image represents the pixel value
of the required point, and the functions f (i, j − 1) and f (i, j + 1) are repre-
senting the pixel values of the neighboring points for f (i, j).

3.2.5 Morphological Operations

After the removal of noise from the image, last step of pre-processing was to
perform morphological operations. There are much chances to have certain ad-
ditional values present in the image even after applying filters which hamper
the process of identification by using CNN. Therefore, we performed erosion
and dilation for morphological operations which were able to determine the
local minimum and local maximum. We applied dilation first followed by ero-
sion. This closing operation was capable of closing small cracks in the image
while not changing the shape of the it.
Title Suppressed Due to Excessive Length 19

3.3 Segmentation

To use only the desired part of the image for various operations, the good idea
is to do image segmentation. For this, we separated each pattern locally in the
optimum way. Thus, we performed segmentation to ensure that the tool would
use only the required information of image for its processing, while accessing
the entire process. We separated each key parameter of the cheque leaflet us-
ing segmentation for correct identification and verification. Other than contour
extraction in order to determine the position values for the date box as ex-
plained in previous section(s), we identified region of interest (ROI) as per the
CTS-2010 cheque leaf standard dimensions as shown in 2. In order to remove
the extra information from the desired region, we created a standard template
using pixel values as per RBI guidelines on CTS-2010 for bank cheques as
shown in Fig. 10. After the segmentation of bank cheque image, we separated
each segment of key parameters for OCR to determine the patterns for verifi-
cation. We also performed transfer learning approach in order to recognize the
handwritten numeric digits present in the date and courtesy amount of cheque.
For Indian banks, image segmentation work is considered automated as the
cheque dimensions of different banks are standardised as per CTS-2010 stan-
dards while for different country cheques, it should be done either manually
due to variation in cheque dimensions or separate dimensions will be defines
as per the country or bank guidelines. Therefore we consider our approach for
bank cheque leaf image segmentation as semi-automatic.

Fig. 10 Sample bank cheque image segmentation

Also, to identify the machine typographic characters of IFSC, cheque num-

ber and account number, we used OCR method while for the Similarly, signa-
20 Prateek Agrawal et al.

ture verification, we used SIFT for feature extraction and SVM for classifica-
tion.

3.4 CNN model for Legal amount and Courtesy amount identification

After extracting the image segments, we used CNN for handwritten numeric
digits recognition of courtesy amount as well as to convert the legal amount
into string. We used Deep Learning Toolbox (a MATLAB toolbox) for CNN
implementation on two convolution layers with six and twelve filters as shown
in Fig. 11. Whereas, in max pooling process, we used a window of 2x2 with one
thousand epochs and one stride. After that, we received numeric output from
the courtesy amount of bank cheques and converted it into string to compare
with the legal amount string. We followed IPV system so as to be written
in the words IPV format. For handwritten courtesy amount identification, we
used transfer learning for digit identification. We used MNIST dataset [82] of
handwritten digits which has sixty thousands examples in training set and ten
thousands examples in test set. All the images in the dataset are normalized
and centered in a fixed size. We used a dataset of two thousands samples (two
hundred for each digit number ranging from 0 to 9) from training dataset and
five hundred (fifty for each digit) samples from testing dataset. We distributed
the entire dataset in 80% for the training and 20% for the testing using CNN.
We trained with the training loss occurring with one thousand epochs, 0.01
base learning rate and seven iterations per epoch. Also for handwritten charac-
ters used in legal amount recognition, we used EMNIST dataset [83] of English
letters identification. EMNIST dataset contains 88800 training examples and
14800 testing examples. We used a dataset of 4160 training samples (160 for
each letter with mixed uppercase and lowercase alphabets in twenty six classes)
and 1040 (forty for each letter) samples for testing dataset in our work. For
CNN training, we distributed this training and testing dataset in a ratio of
80:20. Rest all the CNN training parameters are kept same as it was used for
MNIST dataset training. Table 3 shows the training parameters for MNIST
and EMNIST sample dataset training.

3.4.1 Proposed transfer learning algorithm for amount verification

After identifying the digits from courtesy amount image segment and charac-
ters from legal amount image segment using CNN, we arranged the acquired
numeric digits into the Indian Place Value (IPV) system for matching and
comparing it with the legal amount identified text. For this, we developed
an algorithm ”transfer learning method” to identify different nuances of hand-
written strings of numerals in accordance to the IPV chart. Then we converted
courtesy amount into strings for comparing it with the legal amount. It is very
essential to have a clear match between both the strings as only after that the
process of funds transfer is supposed to be initiated. For this we convert the
Title Suppressed Due to Excessive Length 21

Table 3 CNN model parameters for MNIST and EMNIST dataset.

Parameters Value
Number of convolution layers 2
Number of Max pooling layer 1
Number of filter at convolution layer 6 and 12
Kernel window size at convolution layer 3X3
Kernel window size at max pool layer 2X2
Number of stride at convolution layer 1
Number of stride at max pool layer 1
Number of neurons in output layer 2
Learning rate 0.01
Number of epochs passed 1000
Number of iteration (per epoch) 7
Number of input attributes for digits data (MNIST dataset) 784
Number of output attributes for MNIST dataset 1
Training data for MNIST dataset 2000
Testing data for MNIST dataset 500
Input attributes for letters data (EMNIST dataset) 784
Output attributes for EMNIST dataset 1
Training data for EMNIST dataset 4160
Testing data for EMNIST dataset 1040
Training / testing data 80%/20%
Activation function (before and after max pooling layer) ReLU
Batch size 64

Fig. 11 Proposed CNN architecture for handwritten digit recognition of courtesy amount.

numeric values into strings and then they are compared with the legal amount
strings. As the way of referring the amount IPV system be different due to
22 Prateek Agrawal et al.

local influence thus a dictionary function was used which kept all the possible
ways, a word can be written.
Algorithm 1: Algorithm for verifying legal amount with courtesy
amount.
Input: Input Courtesy amount (number), Legal amount (legal amount)
Output: (1) Courtesy amount to words (courtesy in words)
Output: (2) If Courtesy amount in words equal to Legal Amount
// Indexing of arrays ones[] and tens[] starts from one
1 ones[ ] ← {”One”, ”T wo”, ”T hree”, ...”N ineteen”}
2 tens[ ] ← {”T wenty”, ”T hirty”, ”F orty”, ...”N inety”}
3 S ← null
4 output ← null
5
6 Function convert(number):
7 if number > 19 then
8 if number%10 = 0 then
9 S = tens[(number/10) − 1]
10 else
11 S = tens[(number/10) − 1] + ” ” + ones[number%10]
12 else
13 S = ones[number]
14 end
15 return S
16 End Function
17
18 Function num2words(number):
19 if len(number) > 7 then
20 output ← output + crore
21 if len(number) > 5 then
22 output ← output + lakh
23 if len(number) > 3 then
24 output ← output + thousand
25 if len(number) > 2 then
26 output ← output + hundred
27 if number > 100 and (number%100) then
28 output ← output + ”and”
29 output ← output + convert(number%102 )
30 if len(number) = 2 and (number > 19) then
31 output ← convert(number)
32 if (len(number) = 2 and number ≤ 19) or (len(number) = 1) then
33 output ← ones[number − 1]
34 return output
35 End Function
36
37 Function compare(string 1, string 2):
38 if string 1 = string 2 then
39 Result ← ”M atching”
40 else
41 Result ← ”N ot M atching”
42 end
43 return Result
44 End Function
45
46 crore ← output + convert(number/107 ) + ”Crore”
47 lakh ← output + convert((number/105 )%100) + ”Lakh”
48 thousand ← output + convert((number/103 )%100) + ”T housand”
49 hundred ← output + convert((number/102 )%10) + ”Hundred”
50
// Courtesy amount to words
51 courtesy in words ← num2words(number)
52
// Comparing Courtesy amount in words with Legal Amount
53 result ← compare(courtesy in words, legal amount)
Title Suppressed Due to Excessive Length 23

We followed the algorithm as shown in algorithm 1 to convert the numeric

value of the courtesy amount as per the IPV numbering system and then
to match with the strings of legal amount. Once we converted the numeric
digits into string, next step was to match these strings. For matching the legal
amount with courtesy amount, we used a dictionary function that verified the
characters between both string values. If the strings match each other, the
result was displayed as verified otherwise not-verified. Using this algorithm,
we compared the legal amount with the courtesy amount for verification.

3.5 Signature verification

Next, we moved to signature verification step. In the dataset 114 images used
for our experiments, four images do not contain signatures. Therefore, we used
110 handwritten offline signatures from scanned cheque images for signature
verification task. These 110 signatures were created by 50 different users. Out
of these signature samples, 80 were genuine while rest 30 were forged sig-
natures. Each user’s signatures were found in one to three different cheque
images. Fig. 12 shows the overall signature verification sub-process of bank
cheque. For feature extraction, we normalized each signature image segment
in order to extract the features like resizing, thinning, rotating and cropping.
The normalization experimental results are represented in the Fig. 13. From
the figure, it is clearly shown that we identified the region of interest (of ac-
tual signature) by normalizing the signature image. We applied SIFT method
on these normalized images for feature extraction and contour identification.
SIFT works on a principle to reflect the structure of the image regions us-
ing key points and preventing the scale and orientation information of these
structures. We applied SIFT method into four steps: i) Scale-space extrema
counstruction (SSE), ii) Key point localization (KPL), iii) Orientation assign-
ment (OA), iv) Key point descriptor extraction (KDE).
SSE breaks down each image segment into Gaussian pyramid at different
levels. It further breaks each level into small sub-levels with different variations
using Laplacian of Gaussian (LoG) filters to get the prominent keypoints. Later
in next part, secured locations and key points are detected and calculated. OA
is responsible for calculating the scales and orientations of finalized keypoints.
KDE is responsible to generate a SIFT descriptor against each selected key
point.
After extracting SIFT descriptors and scales & orientations for each signa-
ture image, we applied SVM classifier on these two features dataset to match
and classify the signature patterns. SVM is a machine learning method under
supervised learning. It is used for both classification as well as regression. To
verify the genuineness of the signatures, we applied SVM classifiers on SIFT
extracted features. We distributed genuine and forged signatures data in uni-
form manner between training and testing dataset in ration of 80:20. For this,
88 signature data (with 66 genuine and 22 forged signature data) were taken
in training dataset while 22 signature data (14 genuine and 8 forged signature
24 Prateek Agrawal et al.

data) were taken as testing dataset. In SVM, a separating hyperplane works as

a discriminatory classifier which divides the data using a line or plane into two
parts in two-dimensional space [84]. This method produces the classification
results as data-points which are categorised ideally using optimal hyperplane.

Fig. 12 Account holder Signature classification process for bank cheque verification.

Fig. 13 Normalization of Signature: (a) Original segment of Signature. (b) resized image.
(c) morphological Operation: Thinning. (d) rotation of Image. (e) cropping.
Title Suppressed Due to Excessive Length 25

4 Results and Analysis

We experimented the problem stated in the research with the procedure elu-
cidated above in terms of performance and accuracy using MATLAB software
(deep learning toolbox and image processing toolbox). We used 114 bank
cheque images in our research work. Out of which, 112 images were from
IDRBT bank cheque dataset and 02 cheque images were scanned by us [85]
[86]. We performed training and testing on the developed system with these
cheque leaflets. We used different key parameter segments obtained for bank
cheque leaflet for training and testing.
For handwritten digit recognition using CNN, we achieved the accuracy
with a mean value of 99.14% as shown in Table 4. Similarly, the learning
progress of training networks for character recognition on different epochs
and iterations are shown Table 5. From the table, we can see that the mini
batch loss is reduced up to 0.0077 with 99.94% accuracy at 850 iterations. The
experimental results are represented in Fig. 17.
We achieved the collective loss less than 1% for the trained network. Af-
ter extracting the numbers from the courtesy amount images, we converted
these numbers into words using IPV system. Once we acquired the result in
words for the courtesy amount, we compared it with the legal amount using
the algorithm mentioned above and in case of matching the strings of cour-
tesy amount and legal amount, we proceed with the verification of signature.
We first normalized the segmented image, extracted the features using SIFT
method and then we applied SVM classifier on extracted features to match
and classify the patterns. Using SVM, we were able to achieve 98.1% accuracy.

Table 4 Accurate recognition rate for individual digits.

Digit Types Test Samples Recognition Accuracy (%)
0 50 98.10
1 50 98.98
2 50 99.20
3 50 99.13
4 50 98.96
5 50 99.55
6 50 99.23
7 50 99.11
8 50 99.20
9 50 98.96
Total 500 99.14 (average)

In order to acquire the required segments for different cheque leaflets, we

assigned different templates using pixel values. We took an example of axis
bank cheques and its multiple segments to demonstrate the different key pa-
rameters which are represented in the Fig. 14. We used OCR technique and
applied in-built OCR method for identifying all machine typographic text and
CNN for hand written courtesy amount text. We used SIFT and SVM as re-
26 Prateek Agrawal et al.

Fig. 14 Segmentation of a bank cheque image: (a) Original scanned cheque, (b) cropped
cheque leaflet, (c) gray-scale cheque image, (d) IFS code, (e) account number, (f) cheque
number, (g) courtesy amount, and (h) signature.

ferred in [87] for recognizing the signature patterns. Another sample cheque
with leaflet segmentation is presented in the Fig. 15. Considering a sample
cheque leaflet for experimenting the verification procedure, we got the exper-
imental outcome as shown in the Fig. 16. We executed the segmented images
from figure 15(a), 15(b), and 15(c) by OCR for text identification while for
figure 15(d) and 15(e), we implemented CNN respectively.
There are advances in the field of forensics still handwritten signatures are
considered as one of the most reliable feature for certifying authenticity due to
its unique style of writing. We can perform online as well as offline verification
of signatures and both have their benefits and shortcomings. Some of the highly
efficient methods for signature verification are summarized in the Table 1.
Module wise system performance comparison in terms of accuracy with the
state of the art methods, is summarised in Table 6.
Title Suppressed Due to Excessive Length 27

Fig. 15 Segmented regions of SBI bank cheque: (a) Original scanned cheque, (b) cropped
cheque leaflet, (c) gray-scale cheque image, (d) IFS code, (e) account number, (f) cheque
number, (g) legal amount, (h) courtesy amount, and (i) signature.
28 Prateek Agrawal et al.

Fig. 16 Image segments on left hand size and their OCR results (shown as value of s1) for
machine printed and handwritten characters on right hand side for (a) IFS code, (b) cheque
number, (c) account number, (d) legal amount, and (e) courtesy amount.
Title Suppressed Due to Excessive Length 29

Table 5 Experiment results of CNN for character recognition.

Time elapsed Mini-batch Mini-batch
Epoch Iteration
(hh:mm:ss) accuracy (%) loss
1 1 00:00:00 10.89 2.4308
8 50 00:00:22 97.88 0.0755
15 100 00:00:44 98.58 0.0556
22 150 00:01:06 98.96 0.0423
29 200 00:01:29 99.18 0.0298
36 250 00:01:51 99.38 0.0264
43 300 00:02:13 99.34 0.0254
50 350 00:02:35 99.50 0.0212
58 400 00:02:57 99.66 0.0169
65 450 00:03:20 99.56 0.0202
72 500 00:03:42 99.73 0.0168
79 550 00:04:04 99.76 0.0126
86 600 00:04:27 99.84 0.0108
93 650 00:04:49 99.80 0.0122
100 700 00:05:11 99.84 0.0099
108 750 00:05:33 99.90 0.0083
115 800 00:05:56 99.83 0.0103
122 850 00:06:16 99.94 0.0077

Fig. 17 Experimental loss in training.

5 Conclusion and Future Work

We developed the model to verify the bank cheques using OCR, CNN, SIFT
and SVM. We used OCR method to identify the machine typographic char-
acters with desirable accuracy and efficiency, whereas, we performed CNN to
give precise output for the handwritten digits written on the cheque leaflet.
We proposed and implemented the algorithm to convert numbers into words
to verify the cheque which is one of the major reason of its bounce-off and of
halt the monetary transaction. In order to achieve this, we applied OCR tech-
nique to recognise the machine printing digits and achieved 97.7% accurate
matching. Then, we used separate database to train the network and after
30 Prateek Agrawal et al.

Table 6 Module wise performance analysis of proposed system with other existing works.
Bank cheque Machine typo- Handwritten
Signature
Work details leaf image graphic text & text & digit
verification
segmentation digit recognition recognition
CNN & SVM method
Peres et al. [28] Manual NA for handwritten text NA
(96.5% accuracy)
Template matching &
correlation coefficient
Sneha et al. [36] manual NA NA
for handwritten text
(76.4% accuracy)
Modified quadratic
discriminat function
Jayadevan et al. [38] Manual NA NA
for handwritten text
(97.04% accuracy)
Pixel & geometrical
based CNN method
Ahlawat et al. [88] manual NA NA
for handwritten digit
(99.89% accuracy)
Kullback - Leiber
Mondal & Kundu [69] Manual NA NA Divergence method
(97% accuracy)
Model Agnostic
Hafemann & et al. [79] Manual NA NA Meta Learning method
(94.84% accuracy)
ANN method
Karanjkar et al. [77] Manual NA NA
(85% accuracy)
Gabor Filter &
Ramanathan et al. [14] Manual SVM classifier NA NA
(97% accuracy)
CNN method
Semi-automatic (standard FIST feature extraction
OCR method & transfer learning
Proposed system cheque template & SVM classifier
(97.7% accuracy) algorithm (99.14%
dimensions) (98.1% accuracy)
& 99.94% accuracy)

achieving a desired level of accuracy after training, we used different data sets
to test the trained model for matching the numbers into the words. Evidently,
we achieved an accuracy of 99.14% for the digit recognition which is an im-
provement over the previous accuracy of 99.05% for the CNN used for digit
recognition. Similarly, for character recognition using CNN, we were able to
achieve the accuracy up to 99.94%. For signature recognition, we used SIFT
and SVM classifier to determine the relevancy for verification of signature and
achieved 98.1% accuracy. our innovation will help the banking organizations
to automate the process of cheques clearance in faster and easier way in com-
parison to conventional and manual approach by saving time and manpower.
In our research work, we performed the verification of bank cheques using the
CNN, in an effort to automate the process of verification of the bank cheques
based on the identification of the handwritten numeric digits. However, in or-
der to successfully perform the clearance of bank cheque, we need to perform
the validation of bank cheques as well as ability to function online using a
database resolving the required queries in real time.
Title Suppressed Due to Excessive Length 31

In the future, verification process can be clubbed with validation process

of cheque leaflets by developing a system capable enough of sequentially per-
forming all the required steps for clearance of bank cheques in real time using
a customer-centric database with the required information of a bank customer
and also allowing the bank to keep track of all the transactions via cheques
in a centralized manner, making the process of cheques clearance completely
automated with minimal human intervention and in minimal processing time.
Currently, our innovation is applicable for the cheques written in English lan-
guage only. In future, similar models may be proposed for other languages
too.

Acknowledgements

This work has been partly supported by the European Union Horizon 2020 Re-
search and Innovation Programme under the ARTICONF Project with grant
agreement number 644179.

References

1. Cheque Market 2018. https://www.chequeandcredit.co.uk/information-hub/

facts-and-figures/cheque-market-2018, 2018. [Online; accessed October 4, 2019].
2. Malaysia Bank Template. https://www.chequepro.com/
Cheque-printing-writing-software-for-Malysian-banks.html, 2019. [Online;
accessed October 4, 2019].
3. Chrysanth Cheque Writer. https://freechequewriter.com/download/
cheque-writing-printing-software-for-maldives-banks, 2019. [Online; accessed
October 4, 2019].
4. Malaysia Bank Template. https://www.finder.com/how-to-write-a-check, 2019.
[Online; accessed October 4, 2019].
5. CTS2010. https://www.rbi.org.in/scripts/FAQView.aspx?Id=63, 2018. [Online; ac-
cessed October 4, 2019].
6. Deepak Chaudhary, Prateek Agrawal, and Vishu Madaan. Bank cheque validation using
image processing. International Conference on Advanced Informatics for Computing
Research (ICAICR’19), pages 148–159, 2019.
7. K Dhanva, M Harikrishnan, and P U Babu. Cheque image security enhancement in
online banking. Second International Conference on Inventive Communication and
Computational Technologies (ICICCT), pages 1256–1260, 2018.
8. HK Singh and AE Tigga. Impact of information technology on indian banking services.
1st International Conference on Recent Advances in Information Technology (RAIT),
pages 662–665, 2012.
9. S S Gonge and A A Ghatol. Combined dwt-dct digital watermarking technique software
used for cts of bank. International Conference on Issues and Challenges in Intelligent
Computing Techniques (ICICT), pages 776–783, 2014.
10. S S Gonge and A A Ghatol. An integration of svd digital image watermarking with aes
technique for copyright protection and security of bank cheque image. 2nd International
Conference on Contemporary Computing and Informatics (IC3I), pages 769–778, 2016.
11. D G Menon, S G Sreelakshmi, and A Shivdas. Green banking initiatives: A review
of indian banking sector. International Conference on Technological Advancements in
Power and Energy (TAP Energy), pages 1–7, 2017.
12. H Singh and A Sachan. A proposed approach for character recognition using docu-
ment analysis with ocr. Second International Conference on Intelligent Computing and
Control Systems (ICICCS), page 190195, 2018.
32 Prateek Agrawal et al.

13. P A Wankhede and W S Mohod. A different image content based retrieval using ocr
techniques. International conference of Electronics, Communication and Aerospace
Technology (ICECA), pages 155–161, 2017.
14. R Ramanathan, S Ponmathavan, L N V Thaneshwaran, A S Nair, and K P Soman.
Optical character recognition for english and tamil using support vector. International
Conference on Advances in Computing, Control, and Telecommunication Technologies,
pages 610–612, 2009.
15. C Chen. A general approach for handwritten digits segmentation using spectral clus-
tering. 14th IAPR International Conference on Document Analysis and Recognition,
pages 547–552, 2017.
16. W E Dickinson. A character-recognition study. IBM Journal of Research and Develop-
ment, pages 335–348, 1960.
17. N Liu, C and L Shelton, G. An experimental investigation of a mixed-font. IEEE
Transactions on Electronic Computers, EC-15(6):916–925, 1966.
18. G Nagy and G Shelton. Self-corrective character recognition system. IEEE Transactions
on Information Theory, 12(2):215222, 1966.
19. P Wang and C R Shiau. Machine recognition of printed chinese characters via trans-
formation algorithms. Pattern Recognition, 5(4):303–321, 1972.
20. A Zramdini and R Ingold. Optical font recognition using typographical features. IEEE
Transactions on Pattern Analysis and Machine Intelligence, 20(8):877–882, 1998.
21. Zhu Yong, Tan Tieniu, and Wang Yunhong. Font recognition based on global tex-
ture analysis. IEEE Transactions on Pattern Analysis and Machine Intelligence,
23(20):1192–1200, 2002.
22. P Dhande and R Kharat. Recognition of cursive english handwritten characters. In-
ternational Conference on Trends in Electronics and Informatics ICEI, pages 199–203,
2017.
23. U Pal, F Kimura, K Roy, and T Pal. Recognition of english multi-oriented characters.
International Conference on Pattern Recognition (ICPR’06), pages 1–6, 2006.
24. P Dhande and R Kharat. Character recognition for cursive english handwriting to recog-
nize medicine name from doctors prescription. International Conference on Computing,
Communication, Control and Automation (ICCUBEA), pages 1–5, 2017.
25. A Saenthon and N Sukkhadamrongrak. Comparison the training methods of neural
network for english and thai character recognition. Signal and Information Processing
Association Annual Summit and Conference (APSIPA), pages 1–4, 2014.
26. TT Ramanathan and D Sharma. Multiple classification using svm based multi knowl-
edge based system. International Conference on Advances in Computing & Communi-
cations, (ICACC’17), pages 307–311, 2017.
27. S Stewart, L Pinto, and Barrett B. Segmentation and stitching improves handwriting
recognition on datasets with few samples. 16th International Conference on Frontiers
in Handwriting Recognition (ICFHR), pages 465–470, 2018.
28. A A Peres, S M Vieira, and J R C Pinto. Hybrid neural models for automatic handwrit-
ten digits recognition. International Joint Conference on Neural Networks (IJCNN),
pages 1–8, 2018.
29. Y. LeCun, L. Bottou, Bengio. Y., and P. Haffner. Gradient-based learning applied to
document recognition. Proceedings of the IEEE, 86(11):2278–2324, 1998.
30. R Kajale, S Das, and P Medhekar. Supervised machine learning in intelligent char-
acter recognition of handwritten and printed nameplate. International Conference on
Advances in Computing, Communication and Control (ICAC3), pages 1–5, 2017.
31. H Goto and H Aso. Character pattern extraction from colorful documents with com-
plex backgrounds. Object recognition supported by user interaction for service robots
(ORSUISR), pages 180–183, 2002.
32. M Brisinello and R Gibric. Optical character recognition on images with colorful back-
ground. IEEE 8th International Conference on Consumer Electronics - Berlin (ICCE-
Berlin), pages 1–6, 2018.
33. M Brisinello, R Grbic, M Pul, and T Andelic. Improving optical character recognition
performance for low quality images. 59th International Symposium ELMAR-2017, pages
167–171, 2017.
Title Suppressed Due to Excessive Length 33

34. Wanli Feng and Shangbing Gao. A vehicle license plate recognition algorithm in night
based on hsv. 3rd International Conference on Advanced Computer Theory and Engi-
neering(ICACTE), 4:53–56, 2010.
35. H Jia, Shui and G Wei, Dong. Extraction of legal amount from color bank check images.
International Conference on Web Intelligence and Intelligent Agent Technology, pages
253–256, 2010.
36. Sneha Singh, Tharun Kariveda, Jija Das Gupta, and Kallol Bhattacharya. Handwritten
words recognition for legal amounts of bank cheques in english script. Eighth IEEE
International Conference on Advances in Pattern Recognition (ICAPR), pages 1–5,
2015.
37. Mohammad B. A. Miah, Mohammad A. Yousuf, Mohammad S. Mia, and Mohammad P.
Miya. Handwritten courtesy amount and signature recognition on bank cheque using
neural network. International Journal of Computer Applications, 118(5):21–26, May
2015.
38. R Jayadevan, U Pal, and K F. Recognition of words from legal amounts of indian bank
cheques. 12th International Conference on Frontiers in Handwriting Recognition, pp.
166-171. IEEE, pages 166–171, 2010.
39. S P Raghevendra and Danti A. A novel recognition of indian bank cheques based on
invariant geometrical features. International Conference on Trends in Automation,
Communications and Computing Technology (I-TACT’15), pages 1–5, 2009.
40. K Shirai, M Akita, M Okamoto, K Tanikawa, T Akiyama, and T Sakaguchi. Removal
of background patterns and signatures for magnetic ink character recognition of checks.
10th IAPR International Workshop on Document Analysis Systems, pages 190–194,
2012.
41. H Lu, B Guo, J Liu, and X Yan. A shadow removal method for tesseract text recognition.
International Congress on Image and Signal Processing, BioMedical Engineering and
Informatics (CISP-BMEI), pages 1–5, 2017.
42. C Wiraatmaja, K Gunadi, and IN Sandjaja. The application of deep convolutional
denoising autoencoder for optical character recognition preprocessing. International
Conference on Soft Computing, Intelligent System and Information Technology (IC-
SIIT), pages 72–77, 2017.
43. Jason Aronoff, Steven Simske, and Margaret Sturgill. Automated optimization of void
pantograph settings. HP Laboratories, pages 690–693, 2011.
44. K Peymani and M Soryani. From machine generated to handwritten character recog-
nition; a deep learning approach. 3rd International Conference on Pattern Recognition
and Image Analysis (IPRIA), pages 243–247, 2017.
45. Farsi-datatset. https://github.com/Kianenigma/DeepConvFarsiOCR, 2017. [Online; ac-
cessed October 4, 2019].
46. Neha Bhadwal, Prateek Agrawal, Vishu Madaan, Awadhesh Shukla, and Anuj Kakran.
Smart border surveillance system using wireless sensor network and computer vision.
International Conference on Automation, Computational and Technology Management
(ICACTM19), pages 183–190, 2019.
47. Sandeep Singh Chauhan, Prateek Agrawal, and Vishu Madaan. E-gardener: Building
a plant care taker robot using computer vision. 4th International Conference on Com-
puting Sciences (Feynman100-ICCS’18), pages 137–142, 2018.
48. Prateek Agrawal, Simranjeet Kaur, Harjeet Kaur, and Amita Dhiman. Analysis and
synthesis of an ant colony optimization technique for image edge detection. International
Conference on Computing Science (ICCS), pages 127–131, 2012.
49. A Ignat and B Aciobniei. Handwritten digit recognition using rotations. 18th Inter-
national Symposium on Symbolic and Numeric Algorithms for Scientific Computing,
pages 303–306, 2016.
50. J Hou, H Zeng, L Cai, J Zhu, J Cao, and J Hou. Handwritten numeral recognition
using multi-task learning. International Symposium on Intelligent Signal Processing
and Communication Systems (ISPACS), pages 155–158, 2017.
51. Y Li, J Zhang, P Gao, L Jiang, and M Chen. Grab cut image segmentation based on
image region. 3rd IEEE International Conference on Image, Vision and Computing,
pages 311–315, 2018.
34 Prateek Agrawal et al.

52. A. V. Mityakov, V. K. Varankin, and Y. S. Tatarinov. Application of modern architec-

tures of deep neural networks for solving practical problems. International Conference
on Soft Computing and Measurements (SCM), pages 389–390, 2017.
53. Christian Szegedy, Wei Liu, Yangqing Jia, Pierre Sermanet, Scott Reed, Dragomir
Anguelov, Dumitru Erhan, Vincent Vanhoucke, and Andrew Rabinovich. Going deeper
with convolutions. IEEE Conference on Computer Vision and Pattern Recognition
(CVPR), pages 1–9, 2015.
54. X Ji, Y Li, J Cheng, Y Yu, and M Wang. Cell image segmentation based on an
improved watershed algorithm. 8th International Congress on Image and Signal Pro-
cessing (CISP), pages 433–437, 2015.
55. Sourabh S. Shenoy, Supriya U Nayak, and Neelima B. Swift - a performance accelerated
optimized string matching algorithm for nvidia gpus. International Symposium on
Parallel and Distributed Computing, pages 80–87, 2016.
56. Omkar Sunil Joshi, Bhargavi R. Upadhvay, and Supriya Musica. Parallelized advanced
rabin-karp algorithm for string matching. International Conference on Computing,
Communication, Control and Automation (ICCUBEA), pages 1–5, 2017.
57. Zhou Yansen and Gao Cun. Research and improvement of a multi-pattern matching
algorithm based on double hash. 3rd IEEE International Conference on Computer and
Communications (ICCC), pages 1772–1776, 2017.
58. F. Camastra, M. Spinetti, and A. Vinciarelli. Offline cursive character challenge: a
new benchmark for machine learning and pattern recognition algorithms. International
Conference on Pattern Recognition (ICPR), pages 1–7, 2006.
59. K S Dash, N B Puhan, and G Panda. Gestalt configural superiority effect: A complexity
paradigm for handwritten numeral recognition. Eighth International Conference on
Advances in Pattern Recognition (ICAPR), pages 1–6, 2015.
60. Z Li, M Schulte, Austum, and M Neschen. Fast logo detection and recognition in
document images. 20th International Conference on Pattern Recognition (ICPR10),
pages 2716–2719, 2010.
61. L Shen, W Wu, S Zheng, Shen Linlin, Wu Wenfeng, and Zheng & Songhao. Tv logo
recognition based on luminance variance. IET International Conference on Information
Science and Control Engineering 2012 (ICISCE 2012), pages 1–4, 2012.
62. AA Shirazi, A Dehghani, H Farsi, and M Yazdi. Persian logo recognition using lo-
cal binary patterns. 3rd International Conference on Pattern Recognition and Image
Analysis (IPRIA 2017), pages 258–261, 2017.
63. C Kalaiyarasi and S Karthikeyan. Enhancing logo matching and recognition using
local features. International Conference on Information Communication and Embedded
Systems (ICICES2014), pages 1–6, 2014.
64. C Pornpanomchai, P Boonsripornchai, P Puttong, and C Rattananirundorn. Logo
recognition system. ICSEC15, pages 1–6, 2015.
65. H Sahbi, L Ballan, G Serra, and AD Bimbo. Context-dependent logo matching and
recognition. IEEE Transactions on Image Processing, 22(3):1–13, 2013.
66. Gurpreet Kaur and Prateek Agrawal. Optimisation of image fusion using feature match-
ing based on sift and ransac. Indian Journal of Science and Technology, 9(47):1–7, 2016.
67. T Naseer and N Dogru. Signature recognition by using sift and surf with svm basic on
rbf for voting online. International Conference on Engineering and Technology (ICET),
pages 1–5, 2017.
68. A Ferrer Miguel, Chanda Sukalpa, Diaz Moises, Chayan Kumar Banerjee, Anirban
Majumdar, Carmona Duarte Cristina, Acharya Parikshit, and Pa Umapada. Static
and dynamic synthesis of bengali and devanagari signatures. IEEE Transactions on
Cybernetics, 48(10):2896–2907, 2018.
69. P Mondal and N Kundu. An automated handwritten signature detection approach for
e-security purposes. Third International Conference on Science Technology Engineering
& Management (ICONSTEM), pages 409–413, 2017.
70. S K Jadhav and M K Chavan. Symbolic representation model for off-line signature veri-
fication. 9th International Conference on Computing, Communication and Networking
Technologies (ICCCNT), pages 1–5, 2018.
71. U A Jain and N N Patil. A comparative study of various methods for offline signature
verification. International Conference on Issues and Challenges in Intelligent Comput-
ing Techniques (ICICT), pages 760–764, 2014.
Title Suppressed Due to Excessive Length 35

72. T Handhyani, A R Yohannis, and L Hiryanto. Hand signature and handwriting recogni-
tion as identification of the writer using gray level co-occurrence matrix and bootstrap.
Intelligent Systems Conference (IntelliSys), pages 1103–1110, 2017.
73. D Morocho, A Morales, J Firrrez, and J O Garcia. Humans in the loop: Study of semi-
automatic signature recognition based on attributes. International Carnahan Confer-
ence on Security Technology (ICCST), pages 1–5, 2017.
74. A Beresneva, A Epishkina, and D Shingalova. Handwritten signature attributes for its
verification. IEEE Conference of Russian Young Researchers in Electrical and Elec-
tronic Engineering (EIConRus), pages 1477–1480, 2018.
75. D Morocho, A Morales, J Fierrez, and R V Rodriguez. Human- assisted recognition
based on comparative attributes. 14th IAPR International Conference on Document
Analysis and Recognition, pages 5–9, 2017.
76. F Boudamous, H Nemmour, Y Serdouk, and Y Chibani. Offline handwritten signature
identification using histogram of templates and artificial immune recognition system.
IEEE Conference of Russian Young Researchers in Electrical and Electronic Engineer-
ing (EIConRus), pages 1–5, 2017.
77. Shubhangi L. Karanjkar and P. N. Vasambekar. Signature recognition on bank cheques
using ann. IEEE International WIE Conference on Electrical and Computer Engineer-
ing (WIECON-ECE), pages 44–47, 2016.
78. Luiz G. Hafemann, Robert Sabourin, and Luiz S. Oliveira. Characterizing and evaluat-
ing adversarial examples for offline handwritten signature verification. IEEE Transac-
tions on Information Forensics and Security, 14(8):2153 – 2166, January 2019.
79. L. G. Hafemann, R. Sabourin, and L. S. Oliveira. Meta-learning for fast classifier adap-
tation to new users of signature verification systems. IEEE Transactions on Information
Forensics and Security, 15:1735–1745, 2020.
80. CTS-2010 cheque leaf dimensons and specifictions. https://rbidocs.rbi.org.in/
rdocs/content/PDFs/SCFR220210.pdf, 2010. [Online; accessed July 17, 2020].
81. Kanan Christopher and W. Cottrell Garrison. Color-to-grayscale: Does the method
matter in image recognition? PLoS ONE, 7(1):1–7, 2012.
82. Yann LeCun, Corinna Cortes, and Christopher J.C. Burges. The mnist database of
handwritten digits. http://yann.lecun.com/exdb/mnist/, 1998.
83. Gregory Cohen, Saeed Afshar, Jonathan Tapson, and Andr van Schaik. Emnist: an
extension of mnist to handwritten letters, 2017.
84. V. Madaan and A. Goyal. Predicting ayurveda-based constituent balancing in human
body using machine learning methods. IEEE Access, 8:65060–65070, 2020.
85. P. Dansena, S. Bag, and R. Pal. Differentiating pen inks in hand-written bank cheques
using multi-layer perceptron. 7th International Conference on Pattern recognition and
Machine Intelligence, pages 655–663, 2017.
86. Cheque Market 2018. http://www.idrbt.ac.in/icid.html, 2017. [Online; accessed
October 4, 2019].
87. Zhang Hongyang, Han Jianfeng, Jia Hui, and Zhang Yan. Extraction and matching
of binocular image based on sift algorithm. International Conference on Intelligent
Transportation, Big Data & Smart City (ICITBS), pages 665–668, 2018.
88. S Ahlawat, A Choudhary, A Nayyar, S Singh, and B Yoon. Improved handwritten digit
recognition using convolutional neural networks (cnn). Sensors, 20:3344, 2020.