Integrated Intelligent Research (IIR) International Journal of Data Mining Techniques and Applications

Volume: 05, Issue: 02, December 2016, Page No.195-197

ISSN: 2278-2419

Feature Based Underwater Fish Recognition Using

SVM Classifier
Sunil Kumar C1, Umagowri R2, Elangovan M3
Professor, Department of Computer Science and Engineering,
1

Mahendra Engineering College, Mahendhirapuri, Namakkal District, Mallasamudram, Tamilnadu, India.

2
Assistant Professor, Department of Computer Science and Engineering,
Mahendra Engineering College, Mahendhirapuri, Namakkal District, Mallasamudram, Tamilnadu, India.
3
Assistant Professor, Department of Computer Science and Engineering,
Mahendra Engineering College, Mahendhirapuri, Namakkal District, Mallasamudram, Tamilnadu, India.

Abstract: An approach for underwater fish recognition A new video surveillance object recognition algorithm is
based on wavelet transform is presented in this paper. presented in [7], in which improved invariant moments
This approach decomposes the input image into sub- and length-width ratio of object are extracted as shape
bands by using the multi resolutional analysis known as feature. Bayesian approach of dynamically selecting
Discrete Wavelet Transform (DWT). As each sub-band camera parameters to distinguish a given object from a
in the decomposed image contains useful information finite set of object classes is proposed in [8]. The
about the image, the mean values of every sub-band are Gaussian process regression is applied to learn the
assumed as features. This approach is tested on likelihood of image features given the object classes and
Underwater Photography - A Fish Database. The camera parameters.
database contains 7953 pictures of 1458 different species.
The database is considered for the classification based on II. PROPOSED SYSTEM
Support Vector machine (SVM) classifier. The result
shows that maximum recognition accuracy of 90.74% is The method for underwater fish recognition system is
achieved by the wavelet features. built based on extraction of DWT features in the first
step. These features are tested by the classifier where the
Keywords: Underwater fish recognition, wavelet output is obtained as the second step. These two steps are
transform, Haar wavelet, SVM Classifier. done for both the training and the testing input images.
The framework of our underwater fish recognition
I. INTRODUCTION system is as shown fig 1.

A model that captures the contextual information from A.FEATURE EXTRACTION

more than a hundred object categories using a tree The feature extraction is the most important step in all
structure is presented in [1]. A tree based context model machine learning systems. Likewise in our system the
that improves the object recognition performance and DWT features are extracted in each level of
also provides a coherent interpretation that enables a decomposition method. Before extracting the features,
reliable image querying system by multiple object first the input images undergoes the pre-processing step
categories. Group Sensitive Multiple Kernel Learning where the colour conversion of RGB to Gray colour
(GSMKL) method for object recognition to conversion is done. By using these pre-processed images
accommodate the intra-class diversity and the inter-class only the DWT features are extracted. The DWT features
correlation is presented in [2]. will be having discrete level of wavelet transform
features. This feature extraction is done for various levels
The distinct multicolored regions are detected using edge of decomposition process. The features extracted are
maps and clustering. An illumination and rotation saved as trained database that are used later in
invariant object recognition system is proposed in [3]. classification process.
The color desriptions from distinct regions covering
multiple segments are considered for object B.CLASSIFICATION
representation in [4]. A method used for the category of The same step is done for both the training set and testing
object recognition by civilizing the popular methods sets. The extracted features are stored in the database
from the following two aspects is proposed in [5]. according to their equivalent index class for the
Models that are used for capturing the contextual recognition. The features are extracted and are given as
information among the hundred object categories using a the inputs for the classifier as they need two types of
tree structure are proposed in [6]. It uses objects that inputs like trained database as one input and the tested
contain many instances of different object categories. image features as another input. Then the SVM classifier
1
Integrated Intelligent Research (IIR) International Journal of Data Mining Techniques and Applications
Volume: 05, Issue: 02, December 2016, Page No.195-197
ISSN: 2278-2419
used here will start to compare both the features and Table 1 explains the accuracies of the recognition that are
separate the image class according to the species in the obtained in this system for the underwater fish
present. recognition system. Figure 2 shows the sample input
images of the proposed system.

Table 1: Recognition accuracy of the underwater fish

recognition system

Level of Recognition accuracy (%)

Decomposi 0 1 3 4 6 9
tion 0
50 00 50 00 00
7.2 9.5 4.1 7.4 1.0 3.5
1
5 7 8 2 4 1
87. 78. 73. 66. 58. 51.
2
38 80 02 84 03 71
89. 81. 77. 70. 64. 58.
3
91 72 15 45 17 63
95. 89. 83. 75. 68. 61.
4
52 31 30 03 76 51
98. 94. 89. 83. 76. 69.
5
63 57 85 63 95 57

The approach uses up to 5th level of DWT decomposition.

The SVM classifier makes use of Euclidean distance as
distance measure for classification. Among the 1458
species used 1429 species are classified accurately and
Figure 1 Framework of Underwater fish Recognition only 39 species are misclassified. Performance of the
System proposed feature based underwater fish recognition using
SVM classifier is shown in the below Figure 3.
III. RESULTS

The underwater fish recognition system based on DWT

features are tested on Underwater Photography - A Fish
Database. The database contains 7953 pictures of 1458
different species. The images are tested and classified by
using the SVM classifier. The size of images present in
the database are of 256X256 in size and is of 1458
different species. So as to calculate the performance of
our proposed system, the given database is separated into
two parts namely training and testing set of database
images. The DWT features are extracted and are tested
and classified by using the SVM classifier as explained
above. Fig : Performance of the Proposed Underwater Fish
Recognition System

IV. CONCLUSION

In this paper, a system for recognition of underwater fish

is done based on DWT features and SVM classifier is
explained. The system uses the subband energies of
DWT as features that are used to represent the type of the
fish from the fish database. The system is tested with six
different training set that are separated from the database.
Result stats that the system gives a good recognition
accuracy of 98.63% for the features extracted at 5th level
Figure 2: Sample Input Images
2
Integrated Intelligent Research (IIR) International Journal of Data Mining Techniques and Applications
Volume: 05, Issue: 02, December 2016, Page No.195-197
ISSN: 2278-2419
of DWT decomposition and the overall recognition
accuracy is 90.74% of db.
Reference
[1] Antonio Torralba, and Alan, S, Willsky, Choi,
Myung Jin, (2012), "A tree-based context model for
object recognition." Pattern Analysis and Machine
Intelligence, IEEE Transactions on 34, no. 2, pp.
240-252.
[2] Ling-Yu Duan, Tiejun Huang, and Wen Gao, Yang,
Jingjing, (2012), "Group-sensitive multiple kernel
learning for object recognition." Image Processing,
IEEE Transactions on 21, no. 5, pp. 2838-2852.
[3] Kang, Sangseung, Kim, Jaehong, Lee, Jaeyeon,
Kim, Kyekyung, (2012), “Object recognition for cell
manufacturing system”, Biquitous Robots and
Ambient Intelligence, pp 512-514.
[4] C. A. Murthy and Naik, Sarif Kumar, (2007),
"Distinct multicolored region descriptors for object
recognition." Pattern Analysis and Machine
Intelligence, IEEE Transactions on 29, no. 7, pp.
1291-1296.
[5] Mei, Wu, Yanling, Li, Guangda, Zhou, Xiang-Dong,
Wang, (2010), “Object Recognition via Adaptive
Multi-level Feature Integration”, 12th International
Asia-Pacific on Web Conference, pp. 253-259.
[6] Antonio, B, Willsky, Alan S, Choi, Myung Jin,
Torralba, (2012), “A Tree-Based Context Model for
Object Recognition”, IEEE Transactions on Pattern
Analysis and Machine Intelligence, Vol. 34 , no. 2,
pp.240-252.
[7] Xiao, Zhi-Tao, Wu, Jun, (2010), “Video
surveillance object recognition based on shape and
color features”, Image and Signal Processing pp.451
– 454.
[8] Dencker, Tobias, Roschani, Masoud, Beyerer,
Jürgen, Huber, Marco F, (2012), “Bayesian active
object recognition via Gaussian process regression”,
Information Fusion, pp.1718 – 1725.