Snowflake Logarithmic Metrics for Face Recognition in the Zernike Domain

Size: px
Start display at page:

Download "Snowflake Logarithmic Metrics for Face Recognition in the Zernike Domain"

Transcription

1 Global Journal of Pure and Applied Mathematics. ISSN Volume 3, Number 3 (27), pp Research India Publications Snowflake Logarithmic Metrics for Face Recognition in the Zernike Domain Ahmed J. Ali Faculty of Computer Science & Mathematics, University of Kufa, Najaf, Iraq. Zahir M. Hussain Faculty of Computer Science & Mathematics, University of Kufa, Najaf, Iraq. Professor, School of Engineering, Edith Cowan University, Joondalup, Australia. Mohammed S. Mechee Faculty of Computer Science & Mathematics, University of Kufa, Najaf, Iraq. Abstract Five power transform (a.k.a snowflake) distance metrics are constructed for the purpose of defining similarity measures for holistic face recognition. The power transform metrics are based on the logarithm of five distance metrics: Euclidean, City-Block, Soergel, Lorentzian and Minkowski metrics. The constructed power logarithmic metrics have been called, PL-City-Block, PL-Soergel, PL-Lorentzian and. Distance is considered in the Zernike domain of face images, a domain that contains a vector of Zernike moments for each face image. These distance metrics are further modified as similarity measures then normalized to be restricted over the range (,) so that a fair comparison could be made with the well-known image structural similarity measure () and Feature Similarity Index for Image Quality Assignment (). The purpose is discovering the similarity of a given face image with a face-database to check for best match. A measure for quality of recognition is proposed. This confidence is based on the difference between similarities given by the most likely match with that given by Corresponding author. zmhussain@ieee.org; z.hussain@ecu.edu.au

2 the second-likely confusing face. Simulation results show that the proposed distance measures have superior performance as compared to classical metrics, and the proposed similarity measures are all superior to the well-known and. AMS subject classification: Keywords: Computer vision, face recognition, Zernike moments, power transform, image similarity.. Introduction Face recognition has become a fundamental application after the advent of high performance computing. It is a multidisciplinary filed with many unresolved problems that involve several other fields especially mathematics, numerical analysis, statistics, computer science, and electronic engineering []-, [4]. One of the main streams in face recognition is to recognize a given face image in the sense of similarity with some image in a large face-database. This process involves a lot of unresolved difficulties [5]-, [7]. A powerful approach for image analysis and face recognition is through the use of image moments [8]-, [2], among which Zernike moments is the most important. Zernike moments has been proposed in 934 by Zernike [3], and the use of Zernike moment in image analysis was introduced by Teague [4]. Some researchers have proved Zernike moments to be robust under the presence of Gaussian noise. Due to the fact that Zernike moment functions have been designed using polar coordinates to represent image space, they have been used successfully in image recognition tasks which require rotation invariance. Zernike moments have some advantages over other moments like providing good feature representation of the image, providing more information about facial details, and reducing the dimension of the feature vector which improves computational complexity and accuracy of results [5, 6]. Despite the powerful features given by Zernike moments, they may not be decisive in recognizing a face image from a large face-database. However, this may be due to the inefficient exploitation of hidden features in Zernike moments. In this paper the distance features of Zernike moments are explored, where new similarity measures are designed to reveal the hidden powers of Zernike moments in face recognition. Power transform (a.k.a snowflake) metric is considered [7], from which five distance metrics are designed. Distance is measured in the Zernike domain of the face images. This domain is constructed using vectors of Zernike moments for face images. The power transform metrics are extensions for five well-known distance metrics: Euclidean, City-Block, Soergel, Lorentzian and Minkowski. Using the proposed definition for recognition confidence, numerical simulation of the proposed approach revealed excellent performance in face recognition with high confidence. The paper is organized as follows. In Section 2 some background theoretical basics of Zernike moments, metric distances, structural similarity and feature similarity are briefly reviewed. In Section 3 the proposed measures

3 Snowflake Logarithmic Metrics for Face Recognition 923 are designed, along with a definition of recognition confidence. In Section 4 simulation results are presented and compared to well-known methods. 2. Background In this section some related theoretical principles are explained. First, a brief on Zernike moments is presented, then the notions of metric space and metric distance are summarized. The structural distance between 2D objects is handled through the well-known structural similarity index (). 2.. Zernike Moments Zernike polynomials are a set of orthogonal polynomials defined over the interior of the unit circle. The two-dimensional Zernike moments of order p with repetition q of an image f with intensity function f(r,θ)are defined as follows [5, 8]: Z pq = p + π 2π R pq (r)e jqθ f(r,θ)rdrdθ, r (2.) where the radial polynomial R pq (r) is defined by the following form: R pq (r) = p q 2 k= ( )(p k)!r p 2k k!( p+ q 2 k)!( p q 2 k)! (2.2) Zernike moments have been designed using the polar coordinates (r, θ) R 2 inside the unit circle defined by {(r, θ) : r }. To approximate and compute these moments in discrete Cartesian form, a linear transformation of the image Cartesian coordinates (i, j) from the inside of the square {(i, j) : i, j =,,...,N } to the inside of the unit circle {(r, θ) : r } should be applied to get the following discrete Cartesian form: where r ij = x 2 i + y 2 i ; y i = Z pq = p + N N j= N i= 2j N ; x i = R pq (r ij )e jqθ f(i,j) (2.3) 2i N ; θ ij = tan ( y i x i ) (2.4) Table 2. shows low-order Zernike moments for different orders p and repetition q satisfying the conditions q p such that p q is an even integer. Features of an image f are represented by an n-dimensional vector of selected Zernike moments, V f (details in Section 3). In this work these moments are used to design similarity measures for face recognition; although they have other applications in face emotion recognition [9].

4 924 Ahmed J. Ali, Zahir M. Hussain, and Mohammed S. Mechee Table : Low-order Zernike moments p Dimensionality) Zernike moments (Z pq ) Z 2 Z 2 4 Z 2 Z Z 3 Z Z 4 Z 42 Z Z 5 Z 53 Z Z 6 Z 62 Z 64 Z Z 7 Z 73 Z 75 Z Z 8 Z 82 Z 84 Z 86 Z Z 9 Z 93 Z 95 Z 97 Z Z, Z,2 Z,4 Z,6 Z,8 Z, 2.2. Metric Distance In this subsection we consider five metrics (distance measurements) for the vectors in R n space. These metrics will be used later to design similarity measures Metric Space Let X be a set. A function d : X X R is called metric on X if, for all x,y,z X, the following properties hold:. d(x,y) (non-negativity); 2. d(x,y) = if and only if x = y (separation or self-identity axiom); 3. d(x,y) = d(y,x) (symmetry); 4. d(x,y) d(x,z) + d(z,y) (triangle inequality). A metric space (X, d) is a set X together with a metric d on X [7]. Important distance metrics considered in this study are Euclidean, Minkowski, City-Block, Soergel and Lorentzian. Let X and Y be vectors R n space, where X ={x,x 2,x 3,...,x n } and Y ={y,y 2,y 3,...,y n }. The above metrics are defined as follows [2, 2]: Classical Distance Measures The Euclidean distance between vectors X, Y is computed by: d Euc (X, Y ) = n (x i y i ) 2 (2.5) i=

5 Snowflake Logarithmic Metrics for Face Recognition 925 The Cityblock distance between two vectors X, Y vectors is computed by: d CB (X, Y ) = n x i y i (2.6) The Soergel distance between two vectors X, Y vectors is defined by the following form: ni= x i y i d Sg (X, Y ) = ni= (2.7) max(x i,y i ) The Lorentzian distance between two X, Y vectors is defined by the following form: d Lor (X, Y ) = i= n ln( + x i y i ) (2.8) i= The Minkowski distance between two vectors X, Y is defined by the following form: d MK (X, Y ) = n t (x i y i ) t ; t (2.9) i= The above measures can be used to investigate similarity between images through mean-squared error, however, this is a weak tool for image similarity (see [22]). A more powerful tool for similarity is explained below Image structural similarity measure () Wang et al. (24) [23] introduced a new image quality index named structural similarity index measure () based on the statistical structure of pixel intensities. The measure between two images X and Y is defined as follows: (X,Y) = (2µ xµ y + C )(2σ xy + C 2 ) (µ 2 x + µ2 y + C )(σ 2 x + σ 2 y + C (2.) 2) where µ x and µ y represent the local means of images X and Y, respectively, σ x and σ y represent the standard deviations, σ xy is the covariance of the two images, σ 2 x and σ 2 y represent the variances, respectively, while the constants C and C 2 are defined as C = (K L) 2 and C 2 = (K 2 L) 2 with K =.,K 2 =.3 and L = 255 [24] Feature Similarity Index for Image Quality Assignment () Zhang et al. (2) [25] proposed a novel image quality index called feature similarity index measure based on the phase congruency (PC) and the gradient magnitude (GM). The between images f (p) and f 2 (p) is defined as follows: p = S L(p)P C m (p) p PC (2.) m(p)

6 926 Ahmed J. Ali, Zahir M. Hussain, and Mohammed S. Mechee where p represents position in the image, PC m (p) = max(p C (p), P C 2 (p)) is the maximal phase congruency (P C) value at the position p between f (p) and f 2 (p), S L (p) = S PC (p) S G (p), S P C(p) is the similarity measure for PC (p) and PC 2 (p), S G (p) is the similarity measure for two gradient magnitude values at the position p between f (p) and f 2 (p); and is the whole image spatial domain [25, 26]. 3. Snowflake Logarithmic Distance Metrics In this section we will construct distance metrics based on the power transform metric (also known as snowflake metric). Power transform metric [7]: Given a metric space (X, d) and <α then, the power transform metric is a functional transform metric defined on on X by [d(x,y)] α. Theorem 3.. If (X, d) is a metric space, then (X, d) is also metric space with a new metric d(x,y) = ln( + d(x,y)). Proof. i. To show d(x,y) for all x,y X, since d is metric distance then, we have d(x,y) for all x,y X this implies to + d(x,y). Therefore, the new metric satisfies d(x,y) = ln( + d(x,y)). ii. d(x,y) = iffx = y; d(x,y) = ln( + d(x,y)) = d(x,y) = x = y. iii. d(x,y) = ln( + d(x,y)) = ln( + d(y,x)) = d(y,x). iv. To prove the triangle inequality, the monotone increasing property of the logarithmic function will be utilized as follows. Let x,y,z X, then d(x,z) = ln( + d(x,z)) ln( + d(x,y) + d(y,z)) ln( + d(x,y) + d(y,z) + d(x,y)d(y,z)) = ln(( + d(x,y))( + d(y,z)) = ln( + d(x,y)) + ln( + d(y,z)) = d(x,y) + d(y,z). Power logarithm metrics: Using the Theorem and definitions above, five metric distances are designed as follows. Power logarithm Euclidean distance: For <α, the distance between vectors X, Y is defined by: d PL Euc (X, Y ) = ln( + n (x i y i ) 2 ) i= α (3.2)

7 Snowflake Logarithmic Metrics for Face Recognition 927 Logarithm Cityblock distance: For <α, the PL-Cityblock distance between vectors X, Y is defined by: [ ( α n d PL CB (X, Y ) = ln + x i y i )] (3.3) Power logarithm Soergel distance: For <α, the PL-Soergel distance between two vectors X, Y vectors is defined by the following form: [ ( ni= α x i y i d PL Sg (X, Y ) = ln + ni= (3.4) max(x i,y i ))] Power logarithm Lorentzian distance: For < α, the PL-Lorentzian distance between two vectors X, Y is defined by the following form: [ ( α n d PL Lor (X, Y ) = ln + ln( + x i y i ))] (3.5) Power logarithm Minkowski distance: For <α, the distance between two X, Y vectors is defined by the following form: i= i= n d PL MK (X, Y ) = ln + t (x i y i ) t 3.. Normalized Distance Similarity in the Zernike Domain i= α ; t (3.6) Given a distance metric d between two n-dimensional feature vectors V x and V y, each vector is assumed here to contain a specific number of selected Zernike moments of the two images X and Y, the normalized similarity vector s is defined using doublenormalization of the distance vector d as follows: s = d o max(d o ) ; with d o = d max(d) (3.7) which has the property that s i s i s. The above normalization process is necessary for a fair comparison with and other similarity measures A Measure for Recognition Confidence A performance measure is an importance tool for comparison. Here a measure for the quality of recognition is designed based on the confidence provided by the face recognition approach. The confidence is defined as the complement of confusion between the recognized face and the second-likely face measured by how much near are there similarities with the test image.

8 928 Ahmed J. Ali, Zahir M. Hussain, and Mohammed S. Mechee Applying a similarity formula between a test image and a database of N face images will produce N similarity numbers ranging between and. Nearby numbers are confusing as they reduce confidence in recognition. Most important is the maximum similarity number, which gives the best match, and the second maximum number, which gives a measure of the amount of confusion. After normalization the maximum is always, while the difference m with the second maximum would vary according to the seriousness of confusion. See Figure. Now confidence in face recognition at level k of similarity with a specific test image is defined as follows: C k = where n k is the number of with k-level similarity s k >k. ( k) m n k (3.8) Figure : Similarity confidence. While testing a reference face image x versus a number of other faces, the confusion at decision (similarity) level k(k<) is defined as the number of faces giving similarities above k with the face x under test. 4. Implementation and Results 4.. Test environment The ORL database [27] contains faces of 4 people, each having different images. Some images were taken at different times. The facial expressions (e.g., open or closed eyes, smiling or non-smiling) and occlusion (glasses or no glasses) were included in some subjects. The tolerance for tilting and rotation is considered up to 2 degrees. Some variation in the scale of up to % is included. All images are grayscale and unified to the size of 2 92 pixels.

9 Snowflake Logarithmic Metrics for Face Recognition 929 The face94 database consist of 53 individuals; it contains 2 females, 3 male and 2 male staff images. The background of images is plane green. The images vary with intensity of expression, lighting and glasses [28]. The FEI face database has also been chosen: it contains a set of faces taken over the period from June 25 to March 26 at the Artificial Intelligence Laboratory of FEI in São Bernardo do Campo, São Paulo, Brazil. There are 2 people, with 4 face images for each individual, hence a total of 28 face images. The FEI face images are all colored, taken against a white background in an upright frontal position. A profile rotation of up to 8 degrees is included. Scale variation reaches %. The size of each face image is unified to pixels. The FEI database includes only faces of students and staff at FEI, whose ages range from 9 to 4 years, with distinct appearance, hairstyle, etc. There is equal numbers of males and females, each [29] Simulation Results The Face94, ORL and FEI databases contain 53, 4 and 2 face images, respectively. For the reference person, three images with three different poses have been used: two of them used as a test image, and one as a part of the training database (see Figure 3). In these experiments, best results are obtained with parameters α =. and t =.8 in Equations (2.9)-(3.5). Optimal results are obtained by using the following vector of Zernike features for an image f: V f ={ Z 2, Z 22, Z 3, Z 33, Z 4, Z 42, Z 44,..., Z 9, Z 93, Z 95, Z 97, Z 99 } (4.9) The implementation proceeds in three main parts; the diagram in Figure 2 explains the main steps. Figure 2: The proposed recognition process.

10 93 Ahmed J. Ali, Zahir M. Hussain, and Mohammed S. Mechee Figure 3: Examples of testing using face databases (a) face94 database (b) ORL database (c) FEI database.

11 Snowflake Logarithmic Metrics for Face Recognition 93.9 Person E Lor PL-Lor.9 Person E CBD PL-CBD (a) Lorentzian vs. PL-Lorentzian (b) Cityblock vs. PL-Cityblock.9 Person E Sg PL-Sg.9 Person E Euc PL-Euc (c) Soergel vs. PL-Soergel. (d) Euclidean vs.. Figure 4: Distance-based similarity metrics: Top left: Lorentzian metric vs. PL- Lorentzian; top right: Cityblock metric vs. PL-Cityblock; bottom left: Soergel metric versus PL-Soergel; bottom right: Euclidean metric versus. The first part presents a comparative study to test the performance of proposed distance metrics versus existing metrics. A person E is chosen as shown in Figure 3 to be a reference. Numerical results in Figure (4-5) show that the proposed metrics are more efficient than classical five distance metrics, where they give less confusion or more distance between the most likely face and the second-most likely face. The second part represents a comparative study on the performance of the proposed similarity metrics versus:

12 932 Ahmed J. Ali, Zahir M. Hussain, and Mohammed S. Mechee.9 Person E MK PL-MK Figure 5: Minkowski metric versus PL- Minkowski metric. i) the classical. ii) Feature Similarity Index Measure (). iii) Zernike moment with Euclidean distance, introduced by Singh et al. (2) [3]. The proposed metric measures are much more robust, with larger confidence. In addition, the metric is more efficient than other metrics in the sense of giving less confusion as shown in Figures (6 8). To test the confidence in recognition, the confidence measure given by Equation (3.8) has been implemented and applied. Figure 9 shows higher confidence for the proposed metrics as compared to other methods. Table 4.2 shows recognition rate for the proposed metrics versus other methods. Table 4.2 shows the time complexity of the proposed method versus other methods, where the laptop used here is hp EliteBook workstation 856w with Intel Core i7 272 QM CPU@2.2GHz (8 CPUs) and 8G RAM. Note that the actual time required for recognition using the proposed method can be much less than that in Table 4.2, as Zernike features for the database faces can be calculated offline. The third part represents a comparative study on the performance of Zernike MSE method (with measure based on features in Equation (4.2) under MSE) versus and under Gaussian noise. Peak Signal to Noise Ratio (PSNR) has been used in this test and defined as follows: PSNR = L2 P n (4.2) where p n is the Gaussian noise variance (power). A test image and a noisy version of the same image are shown in Fig., while Fig. shows the proposed method versus and under noise. It is clear that Zernike MSE can detect similarity at very

13 Snowflake Logarithmic Metrics for Face Recognition Person B.9 Person B Zernike[3] Zernike[3] (a) Person B (b) Person B2.9 Person B3.9 Person B Zernike[3] Zernike[3] (c) Person B3 (d) Person B4 Figure 6: Snowflake logarithmic similarity measures versus other methods using different faces from face94 database. low PSNR values, while and fail. For example, at the similarity level of.6, Zernike MSE works under PSNR = 2 db, while and cannot give such a level of similarity until PSNR is 8 db and 3 db, respectively. 5. Conclusion In this paper, five efficient similarity measures have been proposed for the purpose of holistic face recognition. Design of these measures is based on defining a measure of

14 934 Ahmed J. Ali, Zahir M. Hussain, and Mohammed S. Mechee.9 Person C.9 Person C Zernike[3] Zernike[3] (a) Person C (b) Person C Person C Zernike[3] Person C Zernike[3] (c) Person C3 (d) Person C4 Figure 7: Snowflake logarithmic similarity measures versus other methods using different faces from ORL face database. distance in the Zernike domain between a test image and a database of face images. It is shown that the choice of the relevant distance metric has a significant effect on the performance of the similarity metric. Classical distance metrics have been modified for this purpose using power transform (snowflake) metric. A measure for recognition confidence is also proposed. Various comparisons have been made using ORL, face94 and FEI face databases. Results show that the proposed similarity measures based on the proposed distance metrics outperform well-known existing methods like structural similarity measure () and Feature Similarity Index Measure () in recognition

15 Snowflake Logarithmic Metrics for Face Recognition Person E.9 Person E Zernike[3] Zernike[3] (a) Person E (b) Person E2.9 Person E3.9 Person E Zernike[3] Zernike[3] (c) Person E3 (d) Person E4 Figure 8: Snowflake logarithmic similarity measures versus other methods using different faces from FEI face database. rate and confidence. It is worth mentioning that the proposed similarity measure outperforms all other distance measures in giving larger distance between dissimilar vectors. Acknowledgements The Authors would like to thank the Ministry of Higher Education & Scientific Research (Iraq) and the School of Engineering, Edith Cowan University (Australia) for financial support of this project.

16 936 Ahmed J. Ali, Zahir M. Hussain, and Mohammed S. Mechee Confidence in Recognition Person E Zernike[3] Similarity Level, k Figure 9: Recognition confidence versus similarity level for various measures. Table 2: Recognition rate for the proposed method versus other methods. Method Recognition rate% ORL FEI database Face94 database Zemike moments[3] Proposed method with Proposed method with PL-Lorentzian Proposed method with PL-Cityblock Proposed method with Proposed method with PL-Soergel Table 3: Running time for the proposed method versus other methods. Method Running time (seconds) ORL database FEI database Face Zemike moments[3] Proposed method

17 Snowflake Logarithmic Metrics for Face Recognition 937 Figure : Original Image, Noisy Image, PSNR (db) = 2..8 Similarity Measures: A Comparison Zernike MSE PSNR (db) Figure : Zernike MSE measure versus classical similarity measures and under Gaussian noise.

18 938 Ahmed J. Ali, Zahir M. Hussain, and Mohammed S. Mechee References [] D. N. Parmar, B. B. Mehta, Face recognition methods and applications, International Journal of Computer Technology and Applications (IJCTA) 4 () (23). [2] L. G. Vu, A. Alsadoon, P. W. C. Prasad, A.Monem, A. Elchouemi, Face recognition template in photo indexing: A proposal of hybrid principal component analysis and triangular approach (pcaata), IEEE Southwest Symposium on Image Analysis and Interpretation (SSIAI) (26). [3] J. Nagi, S. K. Ahmed, F. Nagi, A matlab based face recognition system using image processing and neural networks, International Colloquium on Signal Processing and its Applications (28). [4] M. Alwakee, Z. Shaaban, Face recognition based on haar wavelet transform and principal component analysis via levenberg-marquardt back-propagation neural network, European Journal of Scientific Research 42 () (2). [5] R. Jafri, H. R. Arabnia, A survey of face recognition techniques, Journal of Information Processing Systems 5 (2) (29). [6] G. Sang, J. Li, Pose-invariant face recognition via rgb-d images, Computational Intelligence and Neuroscience (26). [7] S. D. Kakade, A review paper on face recognition techniques, International Journal for Research in Engineering Application & Management (IJREAM) (26). [8] C. H. The, R. T. Chin, On image analysis by the methods of moments, IEEE Transactions on Pattern Analysis and Machine Intelligence (4) (988). [9] A. Fathi, P. Alirezazadeh, A new global-gabor-zernike feature descriptor and its application to face recognition, J. Visual Communication and Image Representation (26). [] S. M. Imran, M. Rahman, D. Hatzinakos, Differential components of discriminative 2d gaussian-hermite moments for recognition of facial expressions, Pattern Recognition (26). [] S. M. M. Rahman, T. Howlader, D. Hatzinakos, On the selection of 2d krawtchouk moments for face recognition, Pattern Recognition (26). [2] F. Hajati, A. A. Raie, Y. Gao, 2.5d face recognition using patch geodesic moments, Pattern Recognition (22). [3] F. Zernike, Beugunstheorie des schneidenverfahrens und seiner verbesserten form, der phasenkonstrastmethode, Physica (934). [4] M. Teague, Image analysis via the general theory of moments, J. Opt. Soc. Amer. 7 (8) (98). [5] A. M. Hashim, Z. M. Hussian, Local and semi-global feature-correlative techniques for face recognition, Journal of Advanced Computer Science and Applications 5 (8) (24).

19 Snowflake Logarithmic Metrics for Face Recognition 939 [6] S. M. Lajevardi, Z. M. Hussain, Zernike moments for facial expression recognition, Proceedings of the 29 International Conference on Communication, Computer and Power (ICCCP 9) (29). [7] M. M. Deza, E. Deza, Encyclopaedia of distances, Springer (29). [8] K. Sharma, G. Joshi, M. Dutta, Analysis of shape and orientation recognition capability of complex zernike moments for signed gestures, IEEE International Conference on Signal Processing and Integrated Networks (SPIN) (25). [9] S. M. Lajevardi, Z. M. Hussian, Higher order orthogonal moments for facial expression recognition, Digital Signal Processing 2 (6) (2). [2] S. H. Cha, Comprehensive survey on distance/ similarity measures between probability density functions, International Journal of Mathematical Models and Methods in Applied Sciences (4) (27). [2] C.-C. Chen, H.-T. Chu, Similarity measurement between images, IEEE Proceedings of the 29th Annual International Computer Software and Applications Conference (COMPSAC 5) (25). [22] Z. Wang, A. Bovik, Mean squared error: Love it or leave it? a new look at signal fidelity measures, IEEE Signal Processing Magazine 26 () (29). [23] Z. Wang, A. Bovik, H. Sheikh, E. Simoncelli, Image quality assessment: From error visibility to structural similarity, IEEE Transactions on Image Processing 3 (4) (24). [24] D. Dosselmann, X. D. Yang, A comprehensive assessment of the structural similarity index 5 () (2). [25] L. Zhang, Zhang, L. X. Mou, D. Zhang, Fsim: A feature similarity index for image quality assessment, IEEE Transactions on Image Processing 2 (8) (2). [26] A. Li, X. She, Q. Sun, Color image quality assessment combining saliency and fsim, International Conference on Digital Image Processing (23). [27] AT & T Laboratories Cambridge, AT & T face database (formerly the ORL Database of Faces ), facedatabase.html (22). [28] L. Spacek, University of essex, department of computer science, essex.ac.uk/mv/allfaces/faces94.html. [29] Fei face database, image processing laboratory, department of electrical engineering, centro universitario da fei, sao bernardo do campo, sao paulo, brazil, cet/facedatabase.html. [3] C. Singh, N. Mittal, E. Walia, Face recognition using zernike and complex zernike moment features, Pattern Recognition and Image Analysis 2 () (2).

Performance of Image Similarity Measures under Burst Noise with Incomplete Reference

Performance of Image Similarity Measures under Burst Noise with Incomplete Reference Performance of Image Similarity Measures under Burst Noise with Incomplete Reference Nisreen R. Hamza 1, Hind R. M. Shaaban 1, Zahir M. Hussain 1,* and Katrina L. Neville 2 1 Faculty of Computer Science

More information

Face recognition using Zernike moments and Radon transform

Face recognition using Zernike moments and Radon transform Face recognition using Zernike moments and Radon transform Dr. Ali Mohammed Sahan 1, Kadhim Zghair Oudah 2, Adel Majed 3 1 IT Dept., Technical College of Management, Middle Technical University, Baghdad,

More information

Accurate Orthogonal Circular Moment Invariants of Gray-Level Images

Accurate Orthogonal Circular Moment Invariants of Gray-Level Images Journal of Computer Science 7 (5): 75-722, 20 ISSN 549-3636 20 Science Publications Accurate Orthogonal Circular Moment Invariants of Gray-Level Images Khalid Mohamed Hosny Department of Computer Science,

More information

FARSI CHARACTER RECOGNITION USING NEW HYBRID FEATURE EXTRACTION METHODS

FARSI CHARACTER RECOGNITION USING NEW HYBRID FEATURE EXTRACTION METHODS FARSI CHARACTER RECOGNITION USING NEW HYBRID FEATURE EXTRACTION METHODS Fataneh Alavipour 1 and Ali Broumandnia 2 1 Department of Electrical, Computer and Biomedical Engineering, Qazvin branch, Islamic

More information

Hu and Zernike Moments for Sign Language Recognition

Hu and Zernike Moments for Sign Language Recognition Hu and Zernike Moments for Sign Language Recognition K. C. Otiniano-Rodríguez 1, G. Cámara-Chávez 1, and D. Menotti 1 1 Computing Department, Federal University of Ouro Preto, Ouro Preto, MG, Brazil Abstract

More information

Invariant Pattern Recognition using Dual-tree Complex Wavelets and Fourier Features

Invariant Pattern Recognition using Dual-tree Complex Wavelets and Fourier Features Invariant Pattern Recognition using Dual-tree Complex Wavelets and Fourier Features G. Y. Chen and B. Kégl Department of Computer Science and Operations Research, University of Montreal, CP 6128 succ.

More information

Empirical Analysis of Invariance of Transform Coefficients under Rotation

Empirical Analysis of Invariance of Transform Coefficients under Rotation International Journal of Engineering Research and Development e-issn: 2278-67X, p-issn: 2278-8X, www.ijerd.com Volume, Issue 5 (May 25), PP.43-5 Empirical Analysis of Invariance of Transform Coefficients

More information

A Class of Image Metrics Based on the Structural Similarity Quality Index

A Class of Image Metrics Based on the Structural Similarity Quality Index A Class of Image Metrics Based on the Structural Similarity Quality Index Dominique Brunet 1, Edward R. Vrscay 1, and Zhou Wang 2 1 Department of Applied Mathematics, Faculty of Mathematics, University

More information

Multimedia Retrieval Distance. Egon L. van den Broek

Multimedia Retrieval Distance. Egon L. van den Broek Multimedia Retrieval 2018-1019 Distance Egon L. van den Broek 1 The project: Two perspectives Man Machine or? Objective Subjective 2 The default Default: distance = Euclidean distance This is how it is

More information

Aruna Bhat Research Scholar, Department of Electrical Engineering, IIT Delhi, India

Aruna Bhat Research Scholar, Department of Electrical Engineering, IIT Delhi, India International Journal of Scientific Research in Computer Science, Engineering and Information Technology 2017 IJSRCSEIT Volume 2 Issue 6 ISSN : 2456-3307 Robust Face Recognition System using Non Additive

More information

Keywords Eigenface, face recognition, kernel principal component analysis, machine learning. II. LITERATURE REVIEW & OVERVIEW OF PROPOSED METHODOLOGY

Keywords Eigenface, face recognition, kernel principal component analysis, machine learning. II. LITERATURE REVIEW & OVERVIEW OF PROPOSED METHODOLOGY Volume 6, Issue 3, March 2016 ISSN: 2277 128X International Journal of Advanced Research in Computer Science and Software Engineering Research Paper Available online at: www.ijarcsse.com Eigenface and

More information

On the reliability and stability of direct explicit Runge-Kutta integrators

On the reliability and stability of direct explicit Runge-Kutta integrators Global Journal of Pure and Applied Mathematics. ISSN 973-78 Volume, Number 4 (), pp. 3959-3975 Research India Publications http://www.ripublication.com/gjpam.htm On the reliability and stability of direct

More information

International Journal of Computer Engineering and Applications, Volume XII, Special Issue, August 18, ISSN

International Journal of Computer Engineering and Applications, Volume XII, Special Issue, August 18,   ISSN International Journal of Computer Engineering and Applications, Volume XII, Special Issue, August 18, www.ijcea.com ISSN 2321-3469 CONTENT-BASED IMAGE RETRIEVAL USING ZERNIKE MOMENTS AND SURF Priyanka

More information

Face recognition Computer Vision Spring 2018, Lecture 21

Face recognition Computer Vision Spring 2018, Lecture 21 Face recognition http://www.cs.cmu.edu/~16385/ 16-385 Computer Vision Spring 2018, Lecture 21 Course announcements Homework 6 has been posted and is due on April 27 th. - Any questions about the homework?

More information

Original Article Design Approach for Content-Based Image Retrieval Using Gabor-Zernike Features

Original Article Design Approach for Content-Based Image Retrieval Using Gabor-Zernike Features International Archive of Applied Sciences and Technology Volume 3 [2] June 2012: 42-46 ISSN: 0976-4828 Society of Education, India Website: www.soeagra.com/iaast/iaast.htm Original Article Design Approach

More information

Feature Extraction Using Zernike Moments

Feature Extraction Using Zernike Moments Feature Extraction Using Zernike Moments P. Bhaskara Rao Department of Electronics and Communication Engineering ST.Peter's Engineeing college,hyderabad,andhra Pradesh,India D.Vara Prasad Department of

More information

Shape of Gaussians as Feature Descriptors

Shape of Gaussians as Feature Descriptors Shape of Gaussians as Feature Descriptors Liyu Gong, Tianjiang Wang and Fang Liu Intelligent and Distributed Computing Lab, School of Computer Science and Technology Huazhong University of Science and

More information

Non-parametric Classification of Facial Features

Non-parametric Classification of Facial Features Non-parametric Classification of Facial Features Hyun Sung Chang Department of Electrical Engineering and Computer Science Massachusetts Institute of Technology Problem statement In this project, I attempted

More information

Comparing Robustness of Pairwise and Multiclass Neural-Network Systems for Face Recognition

Comparing Robustness of Pairwise and Multiclass Neural-Network Systems for Face Recognition Comparing Robustness of Pairwise and Multiclass Neural-Network Systems for Face Recognition J. Uglov, V. Schetinin, C. Maple Computing and Information System Department, University of Bedfordshire, Luton,

More information

Object Recognition Using Local Characterisation and Zernike Moments

Object Recognition Using Local Characterisation and Zernike Moments Object Recognition Using Local Characterisation and Zernike Moments A. Choksuriwong, H. Laurent, C. Rosenberger, and C. Maaoui Laboratoire Vision et Robotique - UPRES EA 2078, ENSI de Bourges - Université

More information

Science Insights: An International Journal

Science Insights: An International Journal Available online at http://www.urpjournals.com Science Insights: An International Journal Universal Research Publications. All rights reserved ISSN 2277 3835 Original Article Object Recognition using Zernike

More information

Feature Vector Similarity Based on Local Structure

Feature Vector Similarity Based on Local Structure Feature Vector Similarity Based on Local Structure Evgeniya Balmachnova, Luc Florack, and Bart ter Haar Romeny Eindhoven University of Technology, P.O. Box 53, 5600 MB Eindhoven, The Netherlands {E.Balmachnova,L.M.J.Florack,B.M.terHaarRomeny}@tue.nl

More information

An Efficient Algorithm for Fast Computation of Pseudo-Zernike Moments

An Efficient Algorithm for Fast Computation of Pseudo-Zernike Moments An Efficient Algorithm for Fast Computation of Pseudo-Zernike Moments Chong, Chee-Way Fac. of Engg & Tech., Multimedia University, Jalan Air Keroh Lama, 75450 Melaka, Malaysia. (Email:cwayc@pc.jaring.my)

More information

Robot Image Credit: Viktoriya Sukhanova 123RF.com. Dimensionality Reduction

Robot Image Credit: Viktoriya Sukhanova 123RF.com. Dimensionality Reduction Robot Image Credit: Viktoriya Sukhanova 13RF.com Dimensionality Reduction Feature Selection vs. Dimensionality Reduction Feature Selection (last time) Select a subset of features. When classifying novel

More information

Lecture: Face Recognition

Lecture: Face Recognition Lecture: Face Recognition Juan Carlos Niebles and Ranjay Krishna Stanford Vision and Learning Lab Lecture 12-1 What we will learn today Introduction to face recognition The Eigenfaces Algorithm Linear

More information

A New Efficient Method for Producing Global Affine Invariants

A New Efficient Method for Producing Global Affine Invariants A New Efficient Method for Producing Global Affine Invariants Esa Rahtu, Mikko Salo 2, and Janne Heikkilä Machine Vision Group, Department of Electrical and Information Engineering, P.O. Box 45, 94 University

More information

Image Similarity Test Using Eigenface Calculation

Image Similarity Test Using Eigenface Calculation 2017 IJSRST Volume 3 Issue 6 Print ISSN: 2395-6011 Online ISSN: 2395-602X Themed Section: Science and Technology Image Similarity Test Using Eigenface Calculation Nadya Andhika Putri 1, Andysah Putera

More information

A Generative Perspective on MRFs in Low-Level Vision Supplemental Material

A Generative Perspective on MRFs in Low-Level Vision Supplemental Material A Generative Perspective on MRFs in Low-Level Vision Supplemental Material Uwe Schmidt Qi Gao Stefan Roth Department of Computer Science, TU Darmstadt 1. Derivations 1.1. Sampling the Prior We first rewrite

More information

A Recognition System for 3D Embossed Digits on Non-Smooth Metallic Surface

A Recognition System for 3D Embossed Digits on Non-Smooth Metallic Surface 2011 International Conference on elecommunication echnology and Applications Proc.of CSI vol.5 (2011) (2011) IACSI Press, Singapore A Recognition System for 3D Embossed Digits on Non-Smooth Metallic Surface

More information

Eigenface-based facial recognition

Eigenface-based facial recognition Eigenface-based facial recognition Dimitri PISSARENKO December 1, 2002 1 General This document is based upon Turk and Pentland (1991b), Turk and Pentland (1991a) and Smith (2002). 2 How does it work? The

More information

Reconnaissance d objetsd et vision artificielle

Reconnaissance d objetsd et vision artificielle Reconnaissance d objetsd et vision artificielle http://www.di.ens.fr/willow/teaching/recvis09 Lecture 6 Face recognition Face detection Neural nets Attention! Troisième exercice de programmation du le

More information

Statistical Filters for Crowd Image Analysis

Statistical Filters for Crowd Image Analysis Statistical Filters for Crowd Image Analysis Ákos Utasi, Ákos Kiss and Tamás Szirányi Distributed Events Analysis Research Group, Computer and Automation Research Institute H-1111 Budapest, Kende utca

More information

Face detection and recognition. Detection Recognition Sally

Face detection and recognition. Detection Recognition Sally Face detection and recognition Detection Recognition Sally Face detection & recognition Viola & Jones detector Available in open CV Face recognition Eigenfaces for face recognition Metric learning identification

More information

Scale-Invariance of Support Vector Machines based on the Triangular Kernel. Abstract

Scale-Invariance of Support Vector Machines based on the Triangular Kernel. Abstract Scale-Invariance of Support Vector Machines based on the Triangular Kernel François Fleuret Hichem Sahbi IMEDIA Research Group INRIA Domaine de Voluceau 78150 Le Chesnay, France Abstract This paper focuses

More information

Real Time Face Detection and Recognition using Haar - Based Cascade Classifier and Principal Component Analysis

Real Time Face Detection and Recognition using Haar - Based Cascade Classifier and Principal Component Analysis Real Time Face Detection and Recognition using Haar - Based Cascade Classifier and Principal Component Analysis Sarala A. Dabhade PG student M. Tech (Computer Egg) BVDU s COE Pune Prof. Mrunal S. Bewoor

More information

A METHOD OF FINDING IMAGE SIMILAR PATCHES BASED ON GRADIENT-COVARIANCE SIMILARITY

A METHOD OF FINDING IMAGE SIMILAR PATCHES BASED ON GRADIENT-COVARIANCE SIMILARITY IJAMML 3:1 (015) 69-78 September 015 ISSN: 394-58 Available at http://scientificadvances.co.in DOI: http://dx.doi.org/10.1864/ijamml_710011547 A METHOD OF FINDING IMAGE SIMILAR PATCHES BASED ON GRADIENT-COVARIANCE

More information

Learning features by contrasting natural images with noise

Learning features by contrasting natural images with noise Learning features by contrasting natural images with noise Michael Gutmann 1 and Aapo Hyvärinen 12 1 Dept. of Computer Science and HIIT, University of Helsinki, P.O. Box 68, FIN-00014 University of Helsinki,

More information

A Novel Activity Detection Method

A Novel Activity Detection Method A Novel Activity Detection Method Gismy George P.G. Student, Department of ECE, Ilahia College of,muvattupuzha, Kerala, India ABSTRACT: This paper presents an approach for activity state recognition of

More information

MOMENT functions are used in several computer vision

MOMENT functions are used in several computer vision IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 13, NO. 8, AUGUST 2004 1055 Some Computational Aspects of Discrete Orthonormal Moments R. Mukundan, Senior Member, IEEE Abstract Discrete orthogonal moments

More information

Symmetric Two Dimensional Linear Discriminant Analysis (2DLDA)

Symmetric Two Dimensional Linear Discriminant Analysis (2DLDA) Symmetric Two Dimensional inear Discriminant Analysis (2DDA) Dijun uo, Chris Ding, Heng Huang University of Texas at Arlington 701 S. Nedderman Drive Arlington, TX 76013 dijun.luo@gmail.com, {chqding,

More information

Lecture 13 Visual recognition

Lecture 13 Visual recognition Lecture 13 Visual recognition Announcements Silvio Savarese Lecture 13-20-Feb-14 Lecture 13 Visual recognition Object classification bag of words models Discriminative methods Generative methods Object

More information

Neural networks (NN) 1

Neural networks (NN) 1 Neural networks (NN) 1 Hedibert F. Lopes Insper Institute of Education and Research São Paulo, Brazil 1 Slides based on Chapter 11 of Hastie, Tibshirani and Friedman s book The Elements of Statistical

More information

CITS 4402 Computer Vision

CITS 4402 Computer Vision CITS 4402 Computer Vision A/Prof Ajmal Mian Adj/A/Prof Mehdi Ravanbakhsh Lecture 06 Object Recognition Objectives To understand the concept of image based object recognition To learn how to match images

More information

Algorithms for Picture Analysis. Lecture 07: Metrics. Axioms of a Metric

Algorithms for Picture Analysis. Lecture 07: Metrics. Axioms of a Metric Axioms of a Metric Picture analysis always assumes that pictures are defined in coordinates, and we apply the Euclidean metric as the golden standard for distance (or derived, such as area) measurements.

More information

Study of Zernike moments using analytical Zernike polynomials

Study of Zernike moments using analytical Zernike polynomials Available online at www.pelagiaresearchlibrary.com Advances in Applied Science Research, 2012, 3 (1):583-590 ISSN: 0976-8610 CODEN (USA): AASRFC Study of Zernike moments using analytical Zernike polynomials

More information

Michal Kuneš

Michal Kuneš A (Zernike) moment-based nonlocal-means algorithm for image denoising Michal Kuneš xkunes@utia.cas.cz ZOI UTIA, ASCR, Friday seminar 13.03.015 Introduction - Uses nonlocal (NL) means filter - Introduce

More information

Old painting digital color restoration

Old painting digital color restoration Old painting digital color restoration Michail Pappas Ioannis Pitas Dept. of Informatics, Aristotle University of Thessaloniki GR-54643 Thessaloniki, Greece Abstract Many old paintings suffer from the

More information

A Novel Rejection Measurement in Handwritten Numeral Recognition Based on Linear Discriminant Analysis

A Novel Rejection Measurement in Handwritten Numeral Recognition Based on Linear Discriminant Analysis 009 0th International Conference on Document Analysis and Recognition A Novel Rejection easurement in Handwritten Numeral Recognition Based on Linear Discriminant Analysis Chun Lei He Louisa Lam Ching

More information

Face Recognition Using Global Gabor Filter in Small Sample Case *

Face Recognition Using Global Gabor Filter in Small Sample Case * ISSN 1673-9418 CODEN JKYTA8 E-mail: fcst@public2.bta.net.cn Journal of Frontiers of Computer Science and Technology http://www.ceaj.org 1673-9418/2010/04(05)-0420-06 Tel: +86-10-51616056 DOI: 10.3778/j.issn.1673-9418.2010.05.004

More information

Image Recognition Using Modified Zernike Moments

Image Recognition Using Modified Zernike Moments Sensors & Transducers 204 by IFSA Publishing S. L. http://www.sensorsportal.com Image Recognition Using Modified ernike Moments Min HUANG Yaqiong MA and Qiuping GONG School of Computer and Communication

More information

Iterative face image feature extraction with Generalized Hebbian Algorithm and a Sanger-like BCM rule

Iterative face image feature extraction with Generalized Hebbian Algorithm and a Sanger-like BCM rule Iterative face image feature extraction with Generalized Hebbian Algorithm and a Sanger-like BCM rule Clayton Aldern (Clayton_Aldern@brown.edu) Tyler Benster (Tyler_Benster@brown.edu) Carl Olsson (Carl_Olsson@brown.edu)

More information

Orientation Map Based Palmprint Recognition

Orientation Map Based Palmprint Recognition Orientation Map Based Palmprint Recognition (BM) 45 Orientation Map Based Palmprint Recognition B. H. Shekar, N. Harivinod bhshekar@gmail.com, harivinodn@gmail.com India, Mangalore University, Department

More information

Classification of Hand-Written Digits Using Scattering Convolutional Network

Classification of Hand-Written Digits Using Scattering Convolutional Network Mid-year Progress Report Classification of Hand-Written Digits Using Scattering Convolutional Network Dongmian Zou Advisor: Professor Radu Balan Co-Advisor: Dr. Maneesh Singh (SRI) Background Overview

More information

Multi-scale Geometric Summaries for Similarity-based Upstream S

Multi-scale Geometric Summaries for Similarity-based Upstream S Multi-scale Geometric Summaries for Similarity-based Upstream Sensor Fusion Duke University, ECE / Math 3/6/2019 Overall Goals / Design Choices Leverage multiple, heterogeneous modalities in identification

More information

On New Radon-Based Translation, Rotation, and Scaling Invariant Transform for Face Recognition

On New Radon-Based Translation, Rotation, and Scaling Invariant Transform for Face Recognition On New Radon-Based Translation, Rotation, and Scaling Invariant Transform for Face Recognition Tomasz Arodź 1,2 1 Institute of Computer Science, AGH, al. Mickiewicza 30, 30-059 Kraków, Poland 2 Academic

More information

COS 429: COMPUTER VISON Face Recognition

COS 429: COMPUTER VISON Face Recognition COS 429: COMPUTER VISON Face Recognition Intro to recognition PCA and Eigenfaces LDA and Fisherfaces Face detection: Viola & Jones (Optional) generic object models for faces: the Constellation Model Reading:

More information

Blur Image Edge to Enhance Zernike Moments for Object Recognition

Blur Image Edge to Enhance Zernike Moments for Object Recognition Journal of Computer and Communications, 2016, 4, 79-91 http://www.scirp.org/journal/jcc ISSN Online: 2327-5227 ISSN Print: 2327-5219 Blur Image Edge to Enhance Zernike Moments for Object Recognition Chihying

More information

Invariant Feature Extraction from Fingerprint Biometric Using Pseudo Zernike Moments

Invariant Feature Extraction from Fingerprint Biometric Using Pseudo Zernike Moments Invariant Feature Extraction from Biometric Using Pseudo Zernike Moments C.Lakshmi Deepika 1, Dr. A Kandaswamy 2, C. Vimal 3, and B. Sathish 4 1,2,3,4 Department of Biomedical Engineering, PSG College

More information

AN INVESTIGATION ON EFFICIENT FEATURE EXTRACTION APPROACHES FOR ARABIC LETTER RECOGNITION

AN INVESTIGATION ON EFFICIENT FEATURE EXTRACTION APPROACHES FOR ARABIC LETTER RECOGNITION AN INVESTIGATION ON EFFICIENT FEATURE EXTRACTION APPROACHES FOR ARABIC LETTER RECOGNITION H. Aboaisha 1, Zhijie Xu 1, I. El-Feghi 2 z.xu@hud.ac.uk Idrisel@ee.edu.ly 1 School of Computing and Engineering,

More information

Enhanced Fourier Shape Descriptor Using Zero-Padding

Enhanced Fourier Shape Descriptor Using Zero-Padding Enhanced ourier Shape Descriptor Using Zero-Padding Iivari Kunttu, Leena Lepistö, and Ari Visa Tampere University of Technology, Institute of Signal Processing, P.O. Box 553, I-330 Tampere inland {Iivari.Kunttu,

More information

ARestricted Boltzmann machine (RBM) [1] is a probabilistic

ARestricted Boltzmann machine (RBM) [1] is a probabilistic 1 Matrix Product Operator Restricted Boltzmann Machines Cong Chen, Kim Batselier, Ching-Yun Ko, and Ngai Wong chencong@eee.hku.hk, k.batselier@tudelft.nl, cyko@eee.hku.hk, nwong@eee.hku.hk arxiv:1811.04608v1

More information

APPLICATION OF RADIAL BASIS FUNCTION NEURAL NETWORK, TO ESTIMATE THE STATE OF HEALTH FOR LFP BATTERY

APPLICATION OF RADIAL BASIS FUNCTION NEURAL NETWORK, TO ESTIMATE THE STATE OF HEALTH FOR LFP BATTERY International Journal of Electrical and Electronics Engineering (IJEEE) ISSN(P): 2278-9944; ISSN(E): 2278-9952 Vol. 7, Issue 1, Dec - Jan 2018, 1-6 IASET APPLICATION OF RADIAL BASIS FUNCTION NEURAL NETWORK,

More information

ECE 661: Homework 10 Fall 2014

ECE 661: Homework 10 Fall 2014 ECE 661: Homework 10 Fall 2014 This homework consists of the following two parts: (1) Face recognition with PCA and LDA for dimensionality reduction and the nearest-neighborhood rule for classification;

More information

Keywords- Source coding, Huffman encoding, Artificial neural network, Multilayer perceptron, Backpropagation algorithm

Keywords- Source coding, Huffman encoding, Artificial neural network, Multilayer perceptron, Backpropagation algorithm Volume 4, Issue 5, May 2014 ISSN: 2277 128X International Journal of Advanced Research in Computer Science and Software Engineering Research Paper Available online at: www.ijarcsse.com Huffman Encoding

More information

Discriminant Uncorrelated Neighborhood Preserving Projections

Discriminant Uncorrelated Neighborhood Preserving Projections Journal of Information & Computational Science 8: 14 (2011) 3019 3026 Available at http://www.joics.com Discriminant Uncorrelated Neighborhood Preserving Projections Guoqiang WANG a,, Weijuan ZHANG a,

More information

Interpreting Deep Classifiers

Interpreting Deep Classifiers Ruprecht-Karls-University Heidelberg Faculty of Mathematics and Computer Science Seminar: Explainable Machine Learning Interpreting Deep Classifiers by Visual Distillation of Dark Knowledge Author: Daniela

More information

Learning to Disentangle Factors of Variation with Manifold Learning

Learning to Disentangle Factors of Variation with Manifold Learning Learning to Disentangle Factors of Variation with Manifold Learning Scott Reed Kihyuk Sohn Yuting Zhang Honglak Lee University of Michigan, Department of Electrical Engineering and Computer Science 08

More information

What Do Neural Networks Do? MLP Lecture 3 Multi-layer networks 1

What Do Neural Networks Do? MLP Lecture 3 Multi-layer networks 1 What Do Neural Networks Do? MLP Lecture 3 Multi-layer networks 1 Multi-layer networks Steve Renals Machine Learning Practical MLP Lecture 3 7 October 2015 MLP Lecture 3 Multi-layer networks 2 What Do Single

More information

DETECTING HUMAN ACTIVITIES IN THE ARCTIC OCEAN BY CONSTRUCTING AND ANALYZING SUPER-RESOLUTION IMAGES FROM MODIS DATA INTRODUCTION

DETECTING HUMAN ACTIVITIES IN THE ARCTIC OCEAN BY CONSTRUCTING AND ANALYZING SUPER-RESOLUTION IMAGES FROM MODIS DATA INTRODUCTION DETECTING HUMAN ACTIVITIES IN THE ARCTIC OCEAN BY CONSTRUCTING AND ANALYZING SUPER-RESOLUTION IMAGES FROM MODIS DATA Shizhi Chen and YingLi Tian Department of Electrical Engineering The City College of

More information

IEEE TRANSACTIONS ON SYSTEMS, MAN, AND CYBERNETICS PART C: APPLICATIONS AND REVIEWS, VOL. 30, NO. 1, FEBRUARY I = N

IEEE TRANSACTIONS ON SYSTEMS, MAN, AND CYBERNETICS PART C: APPLICATIONS AND REVIEWS, VOL. 30, NO. 1, FEBRUARY I = N IEEE TRANSACTIONS ON SYSTEMS, MAN, AND CYBERNETICS PART C: APPLICATIONS AND REVIEWS, VOL. 30, NO. 1, FEBRUARY 2000 125 Translation, Rotation, and Scale-Invariant Object Recognition L. A. Torres-Méndez,

More information

Object Recognition Using a Neural Network and Invariant Zernike Features

Object Recognition Using a Neural Network and Invariant Zernike Features Object Recognition Using a Neural Network and Invariant Zernike Features Abstract : In this paper, a neural network (NN) based approach for translation, scale, and rotation invariant recognition of objects

More information

Notion of Distance. Metric Distance Binary Vector Distances Tangent Distance

Notion of Distance. Metric Distance Binary Vector Distances Tangent Distance Notion of Distance Metric Distance Binary Vector Distances Tangent Distance Distance Measures Many pattern recognition/data mining techniques are based on similarity measures between objects e.g., nearest-neighbor

More information

Role of Assembling Invariant Moments and SVM in Fingerprint Recognition

Role of Assembling Invariant Moments and SVM in Fingerprint Recognition 56 Role of Assembling Invariant Moments SVM in Fingerprint Recognition 1 Supriya Wable, 2 Chaitali Laulkar 1, 2 Department of Computer Engineering, University of Pune Sinhgad College of Engineering, Pune-411

More information

Unsupervised Learning with Permuted Data

Unsupervised Learning with Permuted Data Unsupervised Learning with Permuted Data Sergey Kirshner skirshne@ics.uci.edu Sridevi Parise sparise@ics.uci.edu Padhraic Smyth smyth@ics.uci.edu School of Information and Computer Science, University

More information

Using Entropy and 2-D Correlation Coefficient as Measuring Indices for Impulsive Noise Reduction Techniques

Using Entropy and 2-D Correlation Coefficient as Measuring Indices for Impulsive Noise Reduction Techniques Using Entropy and 2-D Correlation Coefficient as Measuring Indices for Impulsive Noise Reduction Techniques Zayed M. Ramadan Department of Electronics and Communications Engineering, Faculty of Engineering,

More information

Efficient Computation of Zernike and Pseudo-Zernike Moments for Pattern Classification Applications 1

Efficient Computation of Zernike and Pseudo-Zernike Moments for Pattern Classification Applications 1 REPRESENTATION, PROCESSING, ANALYSIS, AND UNDERSTANDING OF IMAGES Efficient Computation of Zernike and Pseudo-Zernike Moments for Pattern Classification Applications 1 G. A. Papakostas a, Y. S. Boutalis

More information

GLR-Entropy Model for ECG Arrhythmia Detection

GLR-Entropy Model for ECG Arrhythmia Detection , pp.363-370 http://dx.doi.org/0.4257/ijca.204.7.2.32 GLR-Entropy Model for ECG Arrhythmia Detection M. Ali Majidi and H. SadoghiYazdi,2 Department of Computer Engineering, Ferdowsi University of Mashhad,

More information

Local and Semi-Global Feature-Correlative Techniques for Face Recognition

Local and Semi-Global Feature-Correlative Techniques for Face Recognition Local and Semi-Global Feature-Correlative Techniques for Face Recognition Asaad Noori Hashim Dept. of Computer Science University of Babylon P.O.Box 4, Babylon, Iraq Zahir M. Hussain Dept. of Computer

More information

A Robust Modular Wavelet Network Based Symbol Classifier

A Robust Modular Wavelet Network Based Symbol Classifier A Robust Modular Wavelet Network Based Symbol Classifier A.K. Mishra, P.W. Fieguth, D.A. Clausi University of Waterloo, Vision and Image Processing Group akmishra, pfieguth, dclausi@uwaterloo.ca Abstract.

More information

Multiscale Autoconvolution Histograms for Affine Invariant Pattern Recognition

Multiscale Autoconvolution Histograms for Affine Invariant Pattern Recognition Multiscale Autoconvolution Histograms for Affine Invariant Pattern Recognition Esa Rahtu Mikko Salo Janne Heikkilä Department of Electrical and Information Engineering P.O. Box 4500, 90014 University of

More information

Subspace Methods for Visual Learning and Recognition

Subspace Methods for Visual Learning and Recognition This is a shortened version of the tutorial given at the ECCV 2002, Copenhagen, and ICPR 2002, Quebec City. Copyright 2002 by Aleš Leonardis, University of Ljubljana, and Horst Bischof, Graz University

More information

Face Detection and Recognition

Face Detection and Recognition Face Detection and Recognition Face Recognition Problem Reading: Chapter 18.10 and, optionally, Face Recognition using Eigenfaces by M. Turk and A. Pentland Queryimage face query database Face Verification

More information

MYSTERY BEHIND SIMILARITY MEASURES MSE AND SSIM. Gintautas Palubinskas

MYSTERY BEHIND SIMILARITY MEASURES MSE AND SSIM. Gintautas Palubinskas MYSTEY BEHID SIMILAITY MEASUES MSE AD SSIM Gintautas Palubinskas German Aerospace Center DL, emote Sensing Technolog Institute Oberpfaffenhofen, 834 Wessling, German ABSTACT Similarit or distance measures

More information

Two-stage Pedestrian Detection Based on Multiple Features and Machine Learning

Two-stage Pedestrian Detection Based on Multiple Features and Machine Learning 38 3 Vol. 38, No. 3 2012 3 ACTA AUTOMATICA SINICA March, 2012 1 1 1, (Adaboost) (Support vector machine, SVM). (Four direction features, FDF) GAB (Gentle Adaboost) (Entropy-histograms of oriented gradients,

More information

Lecture 8: Interest Point Detection. Saad J Bedros

Lecture 8: Interest Point Detection. Saad J Bedros #1 Lecture 8: Interest Point Detection Saad J Bedros sbedros@umn.edu Last Lecture : Edge Detection Preprocessing of image is desired to eliminate or at least minimize noise effects There is always tradeoff

More information

PCA FACE RECOGNITION

PCA FACE RECOGNITION PCA FACE RECOGNITION The slides are from several sources through James Hays (Brown); Srinivasa Narasimhan (CMU); Silvio Savarese (U. of Michigan); Shree Nayar (Columbia) including their own slides. Goal

More information

Power Load Forecasting based on Multi-task Gaussian Process

Power Load Forecasting based on Multi-task Gaussian Process Preprints of the 19th World Congress The International Federation of Automatic Control Power Load Forecasting based on Multi-task Gaussian Process Yulai Zhang Guiming Luo Fuan Pu Tsinghua National Laboratory

More information

Region Covariance: A Fast Descriptor for Detection and Classification

Region Covariance: A Fast Descriptor for Detection and Classification MITSUBISHI ELECTRIC RESEARCH LABORATORIES http://www.merl.com Region Covariance: A Fast Descriptor for Detection and Classification Oncel Tuzel, Fatih Porikli, Peter Meer TR2005-111 May 2006 Abstract We

More information

Iterative Laplacian Score for Feature Selection

Iterative Laplacian Score for Feature Selection Iterative Laplacian Score for Feature Selection Linling Zhu, Linsong Miao, and Daoqiang Zhang College of Computer Science and echnology, Nanjing University of Aeronautics and Astronautics, Nanjing 2006,

More information

Palmprint Verification with Moments

Palmprint Verification with Moments Palmprint Verification with Moments Ying-Han Pang Andrew T.B.J David N.C.L Hiew Fu San Faculty of Information Science and Technology Multimedia University Jalan Ayer Keroh Lama 75450, Melaka, Malaysia.

More information

Deterministic sampling masks and compressed sensing: Compensating for partial image loss at the pixel level

Deterministic sampling masks and compressed sensing: Compensating for partial image loss at the pixel level Deterministic sampling masks and compressed sensing: Compensating for partial image loss at the pixel level Alfredo Nava-Tudela Institute for Physical Science and Technology and Norbert Wiener Center,

More information

Comparative study of global invariant. descriptors for object recognition

Comparative study of global invariant. descriptors for object recognition Author manuscript, published in "Journal of Electronic Imaging (2008) 1-35" Comparative study of global invariant descriptors for object recognition A. Choksuriwong, B. Emile, C. Rosenberger, H. Laurent

More information

Face Recognition. Face Recognition. Subspace-Based Face Recognition Algorithms. Application of Face Recognition

Face Recognition. Face Recognition. Subspace-Based Face Recognition Algorithms. Application of Face Recognition ace Recognition Identify person based on the appearance of face CSED441:Introduction to Computer Vision (2017) Lecture10: Subspace Methods and ace Recognition Bohyung Han CSE, POSTECH bhhan@postech.ac.kr

More information

New image-quality measure based on wavelets

New image-quality measure based on wavelets Journal of Electronic Imaging 19(1), 118 (Jan Mar 2) New image-quality measure based on wavelets Emil Dumic Sonja Grgic Mislav Grgic University of Zagreb Faculty of Electrical Engineering and Computing

More information

Multi-Layer Boosting for Pattern Recognition

Multi-Layer Boosting for Pattern Recognition Multi-Layer Boosting for Pattern Recognition François Fleuret IDIAP Research Institute, Centre du Parc, P.O. Box 592 1920 Martigny, Switzerland fleuret@idiap.ch Abstract We extend the standard boosting

More information

SIMPLE GABOR FEATURE SPACE FOR INVARIANT OBJECT RECOGNITION

SIMPLE GABOR FEATURE SPACE FOR INVARIANT OBJECT RECOGNITION SIMPLE GABOR FEATURE SPACE FOR INVARIANT OBJECT RECOGNITION Ville Kyrki Joni-Kristian Kamarainen Laboratory of Information Processing, Department of Information Technology, Lappeenranta University of Technology,

More information

Unsupervised Learning: K- Means & PCA

Unsupervised Learning: K- Means & PCA Unsupervised Learning: K- Means & PCA Unsupervised Learning Supervised learning used labeled data pairs (x, y) to learn a func>on f : X Y But, what if we don t have labels? No labels = unsupervised learning

More information

Feature Extraction and Image Processing

Feature Extraction and Image Processing Feature Extraction and Image Processing Second edition Mark S. Nixon Alberto S. Aguado :*авш JBK IIP AMSTERDAM BOSTON HEIDELBERG LONDON NEW YORK OXFORD PARIS SAN DIEGO SAN FRANCISCO SINGAPORE SYDNEY TOKYO

More information

Riemannian Metric Learning for Symmetric Positive Definite Matrices

Riemannian Metric Learning for Symmetric Positive Definite Matrices CMSC 88J: Linear Subspaces and Manifolds for Computer Vision and Machine Learning Riemannian Metric Learning for Symmetric Positive Definite Matrices Raviteja Vemulapalli Guide: Professor David W. Jacobs

More information

A Novel PCA-Based Bayes Classifier and Face Analysis

A Novel PCA-Based Bayes Classifier and Face Analysis A Novel PCA-Based Bayes Classifier and Face Analysis Zhong Jin 1,2, Franck Davoine 3,ZhenLou 2, and Jingyu Yang 2 1 Centre de Visió per Computador, Universitat Autònoma de Barcelona, Barcelona, Spain zhong.in@cvc.uab.es

More information