See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/297791296

Detection and Description of Image Features: An Introduction

Chapter · February 2016

DOI: 10.1007/978-3-319-28854-3_1

CITATIONS READS

37 4,646

2 authors:

Mahmoud Hassaballah Ali Ismail Awad

South Valley University United Arab Emirates University
115 PUBLICATIONS 3,041 CITATIONS 96 PUBLICATIONS 2,626 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Ali Ismail Awad on 12 March 2018.

The user has requested enhancement of the downloaded file.

Detection and Description of Image Features:
An Introduction

M. Hassaballah and Ali Ismail Awad

Abstract Detection and description of image features play a vital role in various
application domains such as image processing, computer vision, pattern recognition,
and machine learning. There are two type of features that can be extracted from an
image content; namely global and local features. Global features describe the image
as a whole and can be interpreted as a particular property of the image involving all
pixels; while, the local features aim to detect keypoints within the image and describe
regions around these keypoints. After extracting the features and their descriptors
from images, matching of common structures between images (i.e., features match-
ing) is the next step for these applications. This chapter presents a general and brief
introduction to topics of feature extraction for a variety of application domains. Its
main aim is to provide short descriptions of the chapters included in this book volume.

Keywords Feature detection · Feature description · Feature matching · Image

processing · Pattern recognition · Computer vision · Applications

1 Introduction

Nowadays, we live in the era of technological revolution sparked by the rapid progress
in computer technology generally, and computer vision especially. Where, the last
few decades can be termed as an epoch of computer revolution, in which develop-

M. Hassaballah (B)
Department of Mathematics, Faculty of Science,
South Valley University, Qena 83523, Egypt
e-mail: [email protected]
A.I. Awad (B)
Department of Computer Science, Electrical and Space Engineering,
Luleå University of Technology, Luleå, Sweden
e-mail: [email protected]
A.I. Awad
Faculty of Engineering, Al Azhar University, Qena, Egypt
e-mail: [email protected]

© Springer International Publishing Switzerland 2016 1

A.I. Awad and M. Hassaballah (eds.), Image Feature Detectors and Descriptors,
Studies in Computational Intelligence 630, DOI 10.1007/978-3-319-28854-3_1
2 M. Hassaballah and A.I. Awad

ments in one domain frequently entail breakthroughs in other domains. Scarcely,

a month passes where one does not hear an announcement of new technological
breakthroughs in the areas of digital computation. Computers and computational
workstations have become powerful enough to process big data. Additionally, the
technology is now available to every one all over the world. As a result, hardware
and multimedia software are becoming standard tools for the handling of images,
video sequence, and 3D visualization.
In particular, computer vision, the art of processing digital images stored within
the computer, became a key technology in several fields and is utilized as a core part
in a large number of industrial vision applications [1]. For instance, computer vision
systems are an important part of autonomous intelligent vehicle parking systems,
adaptive cruise control, driver exhaustion detection, obstacle or traffic sign detection
[2, 3], and driver assistance systems [4]. In industrial automation, computer vision is
routinely used for quality or process control such as food quality evaluation systems
[5]. Even the images used in astronomy and biometric systems or those captured by
intelligent robots as well as medical Computer Assisted Diagnosis (CAD) systems
benefit from computer vision techniques [6]. A basic computer vision system contains
a camera for capturing images, a camera interface, and a PC to achieve some tasks
such as scene reconstruction, object recognition/tracking, 3D reconstruction, image
restoration, and image classification [7, 8]. These tasks rely basically on the detection
and extraction of image features.
Generally, feature extraction involves detecting and isolating desired features of
the image or pattern for identifying or interpreting meaningful information from the
image data. Thus, extracting image features has been considered one of the most
active topics for image representation in computer vision community [9]. Feature
extraction is also an essential pre-processing step in pattern recognition [10]. In
fact, image features can represent the content of either the whole image (i.e., global
features) or small patches of the image (i.e., local features) [11]. Since the global
features aim to represent the image as a whole, only a single feature vector is produced
per image and thus the content of two images can be compared via comparing their
feature vectors. On the contrary, for representing the image with local features, a set
of several local features extracted from different image’s patches is used. For local
features, feature extraction can often be divided into two independent steps: feature
detection and description. The main objective of a feature detector is to find a set of
stable (invariant) distinctive interest points or regions, while the descriptor encodes
information in spatial neighborhoods of the determined regions mathematically. That
is, the descriptor is a vector characterizing local visual appearance or local structure
of image’s patches [12].
In this respect, the number of extracted features is usually smaller than the actual
number of pixels in the image. For instance, a 256 × 256 image contains 65536
pixels, yet the essence of this image may be captured using only few features (e.g.,
30 features). There are many types of image features which can be extracted such as
edges, blobs, corners, interest points, texture, and color [13–16]. A large number of
feature extraction algorithms have been proposed in the literature to provide reliable
feature matching [17–19]. Many feature extraction algorithms are proposed for a
Detection and Description of Image Features … 3

specific applications, where they prove significant success and fail otherwise because
of the different nature of the other applications. A thorough comparison and a detailed
analysis of many extraction algorithms based on different application scenarios are
reported in [11, 20–22]. On the other hand, several trails have been done to make
these algorithms robust to various image artifacts such as illumination variation, blur,
rotation, noise, scale and affine transformation as well as to improve their execution
time performance to be applicable in real time applications [23, 24].
The use of local feature detection and description algorithms in some applications
such as large volume, low-cost, low-power embedded systems, visual odometry, and
photogrammetric applications is still limited or negligible to date due to the lack of a
worldwide industry standard [22]. Further, most of the aforementioned applications
have real-time constraints and would benefit immensely from being able to match
images in a real time, thus developing fast feature extraction algorithms is a must.
With all these factors and avenues to explore, it is not surprising that the problem of
image feature extraction, with various meanings of this expression, is actively pursued
in research by highly qualified people and the volume of research will increase in the
near future, which has given us the motivation for dedicating this book to exemplify
the tremendous progress achieved recently in the topic.

2 Chapters of the Book

This volume contains 15 chapters in total which are divided into two categories. The
following are brief summaries for the content of each chapter.

Part I: Foundations of Image Feature Detectors and Descriptors

Chapter “Image Features Detection, Description and Matching” presents a com-
prehensive review on the available image feature detectors and descriptors such as
Moravec’s detector [25], Harris detector [26], Smallest Univalue Segment Assim-
ilating Nucleus (SUSAN) detector [27], Features from Accelerated Segment Test
(FAST) detector [28], Difference of Gaussian (DoG) detector [29], Scale invari-
ant feature transform (SIFT) descriptor [29], and Speeded-Up Robust Features
(SURF) descriptor [30]. The mathematical foundations of the presented detectors
and descriptors have been highlighted. In general, the chapter serves as a good
foundation for the rest of the volume.
Chapter “A Review of Image Interest Point Detectors: From Algorithms to FPGA
Hardware Implementations” studies some image interest point detectors from the
hardware implementation viewpoint [31]. The chapter offers a review on the
hardware implementation, particularity, using Field Programmable Gate Array
(FPGA), for image interest point detectors [32]. The chapter emphasizes the real-
time performance of FPGA as a hardware accelerator. However, further researches
are demanded for improving the accelerator portability across different platforms.
4 M. Hassaballah and A.I. Awad

Chapter “Image Features Extraction, Selection and Fusion for Computer Vision”
addresses various research problems pertaining to image segmentation, feature
extraction and selection, feature fusion and classification, with applications in
intelligent vehicle, biometrics [33–35], and medical image processing. The chapter
describes different features for different applications from a holistic computer
vision perspective.
Chapter “Image Feature Extraction Acceleration” focuses on accelerating image
feature extraction process using hardware platforms [36]. It presents two focal-
plane accelerators chips, Application-specific Integrated Circuits (ASICs), that
aim at the acceleration of two flagship algorithms in computer vision. The chapter
offers the fundamental concepts driving the design and the implementation of two
focal-plane accelerator chips for the Viola-Jones face detection algorithm [37] and
for the Scale Invariant Feature Transform (SIFT) algorithm [29, 38].
Part II: Applications of Image Feature Detectors and Descriptors
Chapter “Satellite Image Matching and Registration: A Comparative Study Using
Invariant Local Features” is devoted for a comparative study for satellite image
registration using invariant local features. In this chapter, various local feature
detectors and descriptors, such as Features from Accelerated Segment Test (FAST)
[28], Binary Robust Invariant Scalable Keypoints (BRISK) [39], Maximally Sta-
ble Extremal Regions (MSER) [40], and Good Features to Track (GTT) [41],
have been evaluated on different optical and satellite image data sets in terms of
feature extraction, features matching, and geometric transformation. The chapter
documents the performance of the selected feature detectors for the comparison
purpose.
Chapter “Redundancy Elimination in Video Summarization” addresses the redun-
dancy elimination from video summarization using feature point descriptors such
as Binary Robust Independent Elementary Features (BRIEF) [42] and Oriented
FAST and Rotated BRIEF (ORB) [43]. A method for intra-shot and inter-shot
redundancy removal using similarity metric computed from feature descriptors
has been presented. Several feature descriptors have been tested and evaluated for
redundancy removal with a focus on precision and recall performance parameters.
Chapter “A Real Time Dactylology Based Feature Extractrion for Selective Image
Encryption and Artificial Neural Network” combines artificial neural network
with Speeded-Up Robust Features Descriptor (SURF) [30] for selective image
encryption in real time dactylology or finger spelling. Finer spelling is used in
different sign languages and for different purposes [44]. The integrity and the
effectiveness of the proposed scheme have been judged using different factors like
histogram, correlation coefficients, entropy, MSE, and PSNR.
Chapter “Spectral Reflectance Images and Applications” illustrates the use of spec-
tral invariant for obtaining reliable spectral reflectance images. Spectral imaging
can be deployed, for example, in remote sensing, computer vision, industrial appli-
cations, material identification, natural scene rendering, and colorimetric analysis
[45]. The chapter introduces a material classification method based on the invari-
Detection and Description of Image Features … 5

ant representation that results in reliable segmentation of natural scenes and raw
circuit board spectral images.
Chapter “Image Segmentation Using an Evolutionary Method Based on Allostatic
Mechanisms” proposes a multi-thresholding segmentation algorithm that is based
on an evolutionary algorithm called Allostatic Optimization (AO). Threshold-
based segmentation is considered as a simple technique due to the assumption
that the object and the background have different grey level distribution [46]. The
experimental work shows the high performance of the proposed segmentation
algorithm with respect to accuracy and robustness.
Chapter “Image Analysis and Coding Based on Ordinal Data Representation”
utilizes the Ordinal Measures (OM) [47] for image analysis and coding with an
application on iris image as a biometric identifier. Biometrics is a mechanism for
assigning an identity to an individual based on some physiological or behavioral
characteristics. Biometric identifiers include fingerprints, face image, iris patterns,
retinal scan, voice, and signature with broad deployments in forensic and civilian
applications [48].
Chapter “Intelligent Detection of Foveal Zone from Colored Fundus Images of
Human Retina Through a Robust Combination of Fuzzy-Logic and Active Con-
tour Model” proposes a robust fuzzy-rule based image segmentation algorithm for
extracting the Foveal Avascular Zone (FAZ) from retinal images [49]. The pro-
posed algorithm offers a good contribution toward improving the deployment of
retinal images in biometrics-based human identification and verification.
Chapter “Registration of Digital Terrain Images Using Nondegenerate Singular
Points” presents a registration algorithm for digital terrain images using nonde-
generate singular points. The proposed algorithm is a graph-theoretic technique
that uses Morse singularities [50] and an entopic dissimilarity measure [51]. The
experimental outcomes prove the reliability and the accuracy in addition to the
high computational speed of the proposed algorithm.
Chapter “Visual Speech Recognition with Selected Boundary Descriptors” is
devoted for visual speech recognition using some selected boundary descriptors.
Lipreading can be used for speech-to-text for the benefit of hearing impaired indi-
viduals. In the chapter, the Point Distribution Model (PDM) [52] is used to obtain
the lip contour, and the Minimum Redundancy Maximum Relevance (mRMR)
[53] approach is used as a following stage for feature selection.
Chapter “Application of Texture Features for Classification of Primary Benign and
Primary Malignant Focal Liver Lesions” focuses on the classification of the pri-
mary benign and primary malignant local liver lesions. Statistical texture features,
spectral texture features, and spatial filtering based texture feature have been used.
In addition, Support Vector Machine (SVM) [54, 55] and Smooth Support Vector
Machine (SSVM) [56] have been evaluated as two classification algorithms.
Chapter “Application of Statistical Texture Features for Breast Tissue Density
Classification” aims to classify the density of the breast tissues using statistical
features extracted from mammographic images. It presents a CAD system that is
formed from feature extraction module, feature space dimensionality reduction
module, and feature classification module. Different algorithms have been used in
6 M. Hassaballah and A.I. Awad

the classification module such as k-Nearest Neighbor (kNN) [57, 58], Probabilistic
Neural Network (PNN) [59], and Support Vector Machine (SVM) classifiers.

3 Concluding Remarks

Detection and description of image features play a vital role in various application
domains such as image processing, computer vision, pattern recognition, machine
learning, biometrics, and automation. In this book volume, cutting-edge research
contributions on image feature extraction, feature detectors, and feature extractors
have been introduced. The presented contributions support the vitality of image
feature detectors and descriptors, and discover new research gaps in the theoretical
foundations and the practical implementations of image detectors and descriptors.
Due to the rapid growth in representing image using local and global features, further
contributions and findings are anticipated in this research domain.

References

1. Klette, R.: Concise Computer Vision: An introduction into Theory and Algorithms. Springer,
USA (2014)
2. Raxle, C.-C.: Automatic vehicle detection using local features-a statistical approach. IEEE
Trans. Intell. Transp. Syst. 9(1), 83–96 (2008)
3. Mukhtar, A., Likun, X.: Vehicle detection techniques for collision avoidance systems: A review.
IEEE Trans. Intell. Transp. Syst. 16(5), 2318–2338 (2015)
4. Geronimo, D., Lopez, A., Sappa, A., Graf, T.: Survey of pedestrian detection for advanced
driver assistance systems. IEEE Trans. Pattern Anal. Mach. Intell. 32(7), 1239–1258 (2010)
5. Da-Wen, S.: Computer Vision Technology for Food Quality Evaluation. Academic Press, Else-
vier (2008)
6. Menze, B.H., Langs, G., Lu, L., Montillo, A., Tu, Z., Criminisi, A.: Medical computer vision:
recognition techniques and applications in medical imaging. LNCS 7766 (2013)
7. Koen, E., Gevers, T., Snoek, G.: Evaluating color descriptors for object and scene recognition.
IEEE Trans. Pattern Anal. Mach. Intell. 32(9), 1582–1596 (2010)
8. Szeliski, R.: Computer Vision: Algorithms and Applications. Springer, USA (2011)
9. Chen, Z., Sun, S.K.: A Zernike moment phase-based descriptor for local image representation
and matching. IEEE Trans. Image Process. 19(1), 205–219 (2010)
10. Andreopoulos, A., Tsotsos, J.: 50 years of object recognition: directions forward. Comput. Vis.
Image Underst. 117(8), 827–891 (2013)
11. Mikolajczyk, K., Tuytelaars, T., Schmid, C., Zisserman, A., Matas, J., Schaffalitzky, F., Kadir,
T., Gool, L.: A comparison of affine region detectors. Int. J. Comput. Vis. 65(1/2), 43–72 (2005)
12. Moreels, P., Perona, P.: Evaluation of features detectors and descriptors based on 3D objects.
Int. J. Comput. Vis. 73(3), 263–284 (2007)
13. Gonzalez, R.C., Woods, R.E.: Digital Image Processing, 3rd edn. Prentice-Hall Inc, USA
(2007)
14. John, C.R.: The Image Processing Handbook, 6 edn. CRC Press, Taylor & Francis Group, USA
(2011)
15. Li, J., Allinson, N.: A comprehensive review of current local features for computer vision.
Neurocomputing 71(10–12), 1771–1787 (2008)
 Detection and Description of Image Features … 7

16. Liu, S., Bai, X.: Discriminative features for image classification and retrieval. Pattern Recogn.
Lett. 33(6), 744–751 (2012)
17. Tuytelaars, T., Mikolajczyk, K.: Local invariant feature detectors: a survey. Found. Trends
Comput. Graph. Vis. 3(3), 177–280 (2007)
18. Mainali, P., Lafruit, G., Yang, Q., Geelen, B., Gool, L.V., Lauwereins, R.: SIFER: Scale-
invariant feature detector with error resilience. Int. J. Comput. Vis. 104(2), 172–197 (2013)
19. Zhang, Y., Tian, T., Tian, J., Gong, J., Ming, D.: A novel biologically inspired local feature
descriptor. Biol. Cybern. 108(3), 275–290 (2014)
20. Hugo, P.: Performance evaluation of keypoint detection and matching techniques on grayscale
data. SIViP 9(5), 1009–1019 (2015)
21. Bouchiha, R., Besbes, K.: Comparison of local descriptors for automatic remote sensing image
registration. SIViP 9(2), 463–469 (2015)
22. Bianco, S., Mazzini, D., Pau, D., Schettini, R.: Local detectors and compact descriptors for
visual search: a quantitative comparison. Digital Sig. Process 44, 1–13 (2015)
23. Takacs, G., Chandrasekhar, V., Tsai, S., Chen, D., Grzeszczuk, R., Girod, B.: Rotation-invariant
fast features for large-scale recognition and real-time tracking. Sig. Process: Image Commun.
28(4), 334–344 (2013)
24. Seidenari, L., Serra, G., Bagdanov, A., Del Bimbo, A.: Local pyramidal descriptors for image
recognition. IEEE Trans. Pattern Anal. Mach. Intell. 36(5), 1033–1040 (2014)
25. Morevec, H.P.: Towards automatic visual obstacle avoidance. In: Proceedings of the 5th Inter-
national Joint Conference on Artificial Intelligence, vol. 2, pp. 584–584. IJCAI’77. Morgan
Kaufmann Publishers Inc., San Francisco (1977)
26. Harris, C., Stephens, M.: A combined corner and edge detection. In: Proceedings of the Fourth
Alvey Vision Conference, pp. 147–151 (1988)
27. Smith, S., Brady, J.: Susan-a new approach to low level image processing. Int. J. Comput. Vis.
23(1), 45–78 (1997)
28. Rosten, E., Drummond, T.: Fusing points and lines for high performance tracking. In: Proceed-
ings of the Tenth IEEE International Conference on Computer Vision, vol. 2, pp. 1508–1515.
ICCV’05, IEEE Computer Society, Washington, DC (2005)
29. Lowe, D.G.: Distinctive image features from scale-invariant keypoints. Int. J. Comput. Vis.
60(2), 91–110 (2004)
30. Bay, H., Ess, A., Tuytelaars, T., Gool, L.V.: Speeded-up robust features (SURF). Comput. Vis.
Image Underst. 110(3), 346–359 (2008)
31. Schmid, C., Mohr, R., Bauckhage, C.: Evaluation of interest point detectors. Int. J. Comput.
Vis. 37(2), 151–172 (2000)
32. Possa, P., Mahmoudi, S., Harb, N., Valderrama, C., Manneback, P.: A multi-resolution FPGA-
Based architecture for real-time edge and corner detection. IEEE Trans. Comput. 63(10), 2376–
2388 (2014)
33. Jain, A.K., Ross, A.A., Nandakumar, K.: Introduction to Biometrics, 1st edn.Springer (2011)
34. Egawa, S., Awad, A.I., Baba, K.: Evaluation of acceleration algorithm for biometric identifi-
cation. In: Benlamri, R. (ed.) Networked Digital Technologies, Communications in Computer
and Information Science, vol. 294, pp. 231–242. Springer, Heidelberg (2012)
35. Awad, A.I.: Fingerprint local invariant feature extraction on GPU with CUDA. Informatica
(Slovenia) 37(3), 279–284 (2013)
36. Awad, A.I.: Fast fingerprint orientation field estimation incorporating general purpose GPU. In:
Balas, V.E., Jain, L.C., Kovaevi, B. (eds.) Soft Computing Applications, Advances in Intelligent
Systems and Computing, vol. 357, pp. 891–902. Springer International Publishing (2016)
37. Viola, P., Jones, M.: Rapid object detection using a boosted cascade of simple features. In:
Proceedings of the 2001 IEEE Computer Society Conference on Computer Vision and Pattern
Recognition (CVPR 2001), vol. 1, pp. 511–518 (2001)
38. Awad, A.I., Baba, K.: Evaluation of a fingerprint identification algorithm with SIFT features. In:
Proceedings of the 3rd 2012 IIAI International Conference on Advanced Applied Informatics,
pp. 129–132. IEEE, Fukuoka, Japan (2012)

View publication stats