AUTOMATIC fingerprint recognition systems (AFRS) are

Size: px
Start display at page:

Download "AUTOMATIC fingerprint recognition systems (AFRS) are"

Transcription

1 1808 IEEE SIGNAL PROCESSING LETTERS, VOL. 24, NO. 12, DECEMBER 2017 DeepPore: Fingerprint Pore Extraction Using Deep Convolutional Neural Networks Han-Ul Jang, Student Member, IEEE, Dongkyu Kim, Student Member, IEEE, Seung-Min Mun, Sunghee Choi, and Heung-Kyu Lee Abstract As technological developments have enabled highquality fingerprint scanning, sweat pores, one of the Level 3 features of fingerprints, have been successfully used in automatic fingerprint recognition systems (AFRS). Since the pore extraction process is a critical step for AFRS, high accuracy is required. However, it is difficult to extract the pore correctly because the pore shape depends on the person, region, and pore type. To solve the problem, we have presented a pore extraction method using deep convolutional neural networks and pore intensity refinement. The deep networks are used to detect pores in detail using a large area of a fingerprint image. We then refine the pore information by finding local maxima to identify pores with different intensities in the fingerprint image. The experimental results show that our pore extraction method performs better than the state-of-the-art methods. Index Terms Biometrics, convolutional neural network (CNN), fingerprint, pore extraction. I. INTRODUCTION AUTOMATIC fingerprint recognition systems (AFRS) are currently one of the most widely used biometric systems. Due to the development of the fingerprint scanner, more detailed Level 3 features are observed compared to Level 1 (pattern) and Level 2 (minutiae points), which are features used in low-resolution fingerprint images [1]. Level 3 features can be extracted from high-resolution fingerprint images of dots per inch (DPI), including sweat pores, incipient ridges, and edge contours [2]. The coordinates of pores are highly discriminative information [3] and the fingerprint recognition technology using Level 3 features shows high accuracy. The pore information is also used for partial matching [4] and fingerprint liveness detection [5] with Level 2 minutiae features. Pore extraction is an essential process for the variety of biometric systems based on pores. Thus, it is important to develop a pore extraction technique that has high accuracy. Manuscript received July 3, 2017; revised September 13, 2017; accepted September 26, Date of publication October 9, 2017; date of current version October 25, This work was supported by the National Research Foundation of Korea Grant funded by the Korea government (MSIT) (NRF- 2016R1A2B ). The associate editor coordinating the review of this manuscript and approving it for publication was Dr. Sanghoon Lee. (Corresponding author: Heung-Kyu Lee.) H.-U. Jang, D. Kim, S.-M. Mun, and H.-K. Lee are with the Multimedia Computing Laboratory, KAIST, Daejeon 34141, South Korea ( hanulj@kaist.ac.kr; d.kim@kaist.ac.kr; smmun@mmc.kaist.ac.kr; heunglee@ kaist.ac.kr). S. Choi is with the Geometric Computing Laboratory, KAIST, Daejeon 34141, South Korea ( sunghee@kaist.edu). Color versions of one or more of the figures in this letter are available online at Digital Object Identifier /LSP The existing pore extraction techniques are feature-based approaches [1], [6], [7] or a learning-based approach [8]. The feature-based approaches use isotropic or anisotropic models to detect pores. Qijun et al. proposed an adaptive pore detection technique using an anisotropic model [7]. Although, it is more accurate than the previous isotropic model based methods [1], [6], there are many limitations to adapt various pore shapes of various people. Labati et al. suggested the learning-based technique using convolutional neural networks (CNN) [8]. Since the technique involves subsampling, the pore intensity map is blurred that makes it difficult to accurately detect the pore location. Also, using the CNN structure without nonlinearity may limit the detection of various types of pores. The learning-based method uses Otsu s algorithm [9], which is binary thresholding (BT), for postprocessing to refine a pore intensity map. However, it is difficult to accurately detect pores with BT over the entire pore intensity map because the pore intensity is different for each pore. As a result, the accuracy of the learning-based method is similar to the feature-based technique using anisotropic models. Adaptive pore extraction is a very difficult problem because the shape of the pores differs from person to person and there are differences in the shape of the pores in the fingerprint of the same person. To overcome this issue on pore extraction, we propose a Deep CNN based Pore extraction (DeepPore). DeepPore consists of pore intensity extraction using CNN, and postprocessing using pore intensity refinement (PIR). The main contributions of our approach are the deep CNN architecture for detailed pore extraction and the refinement for pore peak finding, and more specific descriptions are as follows. 1) We suggest deep CNN for detailed pore extraction. Since deep CNN with many layers has a large receptive field [10], it can use a lot of information, such as the pore surrounding shape and the ridge shape of the pore belongs to, in the fingerprint image to detect pores. 2) We propose a postprocessing technique that detects pore peaks adaptively in the pore intensity map. In order to detect pore peaks with different intensities for each fingerprint region, we use a refinement method to find local maxima instead of the BT [9]. II. THE PROPOSED METHOD DeepPore consists of pore intensity extraction using CNN and postprocessing using PIR as presented in Fig. 1. In this section, we describe the CNN architecture for pore intensity extraction IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission. See standards/publications/rights/index.html for more information.

2 JANG et al.: DEEPPORE: FINGERPRINT PORE EXTRACTION USING DEEP CONVOLUTIONAL NEURAL NETWORKS 1809 Fig. 1. Overview of DeepPore. Fig. 2. CNN architecture for pore detection. Each convolutional layer contains 64 filters and some sample feature maps are drawn for visualization. The color of the feature map is inverted so that it can be visually confirmed. and then explain how to train the CNN. Finally, we introduce PIR, the postprocessing process. A. Proposed Network We use a deep convolutional network inspired by CNN-visual geometry group (VGG) [10] and train the CNN of DeepPore (CNN DP ) in a supervised learning manner. The goal of training to estimate a pore intensity map where the pores are enhanced and other patterns are reduced. We design the ground truth of a pore intensity map using soft labels. That is, the closer the distance to the pore, the higher the value of the label. Let l i,j and d i,j np denote the label value at a pixel coordinate (i, j), and the Euclidean distance between (i, j) and its nearest pore from (i, j) in the fingerprint image. If d i,j np is less than d τ, l i,j =1 d i,j np/d τ where d τ is the pore distance threshold. Otherwise (i, j) is a nonpore pixel, so l i,j =0. The CNN architecture of DeepPore is illustrated in Fig. 2. We perform the pore extraction process using d layers. Each layer has the same type: 64 filters with a size of 3 3. Therefore, 64 feature maps are generated for each layer. The layers 1 through d 1 perform convolution, batch normalization, and rectified linear unit (ReLU). In the last layer d, only the convolution process proceeds. Thus, the number of learnable layers is d and the number of nonlinearity functions is d 1. Weusethe fingerprint image as the input of CNN DP, and the output is the pore intensity map corresponding to the fingerprint image. The pore intensity map has the same size as the fingerprint image. We pad zeros before the convolution process in each layer to produce output equal in size to the input. Subsampling, which is commonly used for feature compression in object recognition, is not used because it may cause distortion in accurate pore position detection. In this letter, we use deep CNN because deep networks have larger receptive field sizes. For CNN with depth D and the size of the filters is 3 3, the receptive field size has (2D +1) (2D +1). In the task of pore extraction, the shape of the pores depends on the shape of the ridge to which they belong, their types (open or closed pore), and so on. We, therefore, use deep networks with large receptive field sizes to analyze a wider range Fig. 3. Visualization of (a) an example of pore intensity map and (b) an exampleof pore map. Asubblock of size is set for better visualization. of fingerprint information from the image and more precisely detect the pores. B. Training We now describe the process of training CNN DP.Letx and y denote the fingerprint image and the label of its pore intensity map, respectively. Given a training dataset {x i, y i } N i=1, our goal is to train model f to predict ŷ=f(x), where ŷ is the estimate of the pore intensity map and N is the batch size. The loss function of CNN DP is defined as follows: L DP = 1 2N N y i ŷ i 2. (1) i=1 C. Pore Intensity Refinement (PIR) The pore pixels appear as peaks in the pore intensity map. The intensity of each pore peak varies depending on the thickness of the ridge to which it belongs and its type (open or closed pore). Therefore, it is difficult to accurately detect the pores with BT over the entire pore intensity map. As shown in Fig. 3(a), the pore pixels are local maxima in the pore intensity map. We estimate the pore map by finding local maxima (pore peaks) [11]. Algorithm 1 presents our postprocessing method for pore map estimation. In summary, we estimate local maxima of the pore intensity map, which are equal to or greater than P τ, as pore pixels. The

3 1810 IEEE SIGNAL PROCESSING LETTERS, VOL. 24, NO. 12, DECEMBER 2017 Algorithm 1: Finding local maxima of a pore intensity map. M I is the pore intensity map of size R C M P is the pore map of size R C, ( (i, j) M i,j P 0) W i,j is the window of size s s and the center coordinate is (i, j) M I (W i,j ) is the matrix of M I belonging to W i,j window P τ is the pore detection threshold for i =1to R do for j =1to C do if M i,j I >= P τ then if M i,j I == max(m I (W i,j )) then M i,j P 1 M I (W i,j ) 0 end if end if end for end for pore map is generated by marking the coordinates corresponding to the estimated pore pixel to one and the other coordinates to zero. The example of the pore map is illustrated in Fig. 3(b). III. EXPERIMENTS A. Implementation Details We used a high-resolution-fingerprint (HRF) database [7] with 30 labeled images for pore extraction. The HRF database has pores coordinates, and each image has pixels and 1200 DPI resolution. For training CNN DP steadily, we partitioned each fingerprint image into overlapped subblocks and performed data augmentation to obtain sufficient training datasets. The size of the subblock was 80 80, which can accommodate the receptive field size of CNN DP with up to 10 learnable layers. A step size of 10 was used to generate overlapped subblocks, and a total of 89,250 subblocks were used in the experiment. The data augmentation included three types of flips and three types of counterclockwise rotation (90, 180, and 270 ). For testing, we used full-sized fingerprint images. The number of learnable layers d was 10, and ReLU was used as the activation function for training the model. CNN DP was trained using the Caffe [12] framework, which provides very fast CPU and GPU implementations. The base learning rate was set to We used batches of size 5 and trained all experiments for 100 epochs. To optimize CNN DP learning, we used the adaptive moment estimation (ADAM) to compute adaptive learning rates [13]. This is because, the ADAM has a better loss reduction than the stochastic gradient descent used for general purposes. We also used adjustable gradient clipping because the vanishing/exploding gradients problem can occur if the adaptive learning rate computed by ADAM is too [ high ][14]. In the learning process, the gradient is clipped to θ γ, θ γ, where γ is the current learning rate and θ =50is the gradient clipping parameter. Pore distance threshold d τ was set to 5. The size of the sliding window W for PIR was set to 11 11, which can accommodate TABLE I AVERAGE PERFORMANCE METRICS IN PERCENTAGE AND STANDARD DEVIATION (IN PARENTHESIS) DeepPore Labati et al. [8] DAPM [7] Adapt. DoG [7] R T (4.63) (7.81) 84.8 (4.5) 80.8 (6.5) R F 8.64 (4.15) (6.20) 17.6 (6.3) 22.2 (9.0) the range of pore adjacent pixels in the pore intensity map. Pore detection threshold P τ was set to d τ, the size of W, and P τ were experimentally set to values having the best performance. B. Evaluation Procedure We tested DeepPore using a five-fold validation strategy [15]. Three folds were used for training, one fold for validation, and one fold for testing. The validation fold is used to find the model with the lowest loss and the testing fold is used for evaluation. In each evaluation, the training, validation, and testing folds were mutually exclusive. Average performance is calculated by averaging the results of five evaluations. The true detection rate (R T ) and the false detection rate (R F ) are used to measure the accuracy of the pore extraction. R T is the ratio of the number of detected true pores to the number of all true pores, and R F is the number of falsely detected pores to the total number of detected pores. A good pore extraction algorithm should have a high R T and a low R F. We consider that the true pore is correctly detected if the Euclidean distance between the coordinate of the true pore and the coordinate of the detected pore is less than d p pixels. d p is set to RW /2, where RW is the average ridge width in the images of the test dataset. This is the same test environment as the state-of-the-art method [8]. IV. RESULTS A. Quantitative Evaluation We compare the state-of-the-art methods and DeepPore performance, and the results are described in Table I. DeepPore outperforms the existing feature-based methods such as adaptive DoG and dynamic anisotropic pore model (DAPM) [7] as well as the learning-based method [8]. DeepPore achieves higher R T,lowerR F, and lower standard deviation of R F compared to the existing techniques. Fig. 5 shows the performance of the networks with various depths. We measured the performance of each network by varying the pore detection threshold P τ. The networks with ten layers show the best performance when R F is less than 0.1. As the depth of network increases, the performance improves and the size of the performance improvement decreases. The performance of networks with ten layers is very slightly improved than that of networks with eight layers. It is expected that there will be no significant performance improvement even if the networks with a depth of eleven or more are used. In Table II, we compare performance against changes in networks and postprocessing. The networks and the postprocessing technique used by Labati s method [8], which is the

4 JANG et al.: DEEPPORE: FINGERPRINT PORE EXTRACTION USING DEEP CONVOLUTIONAL NEURAL NETWORKS 1811 Fig. 4. (a) Original image. The pore detection results of (b) DeepPore (c) and Labati s method [8]. The green, yellow, and red dots in these images represent the true positive, false negative, and false positive, respectively. TABLE III AVERAGE TIME BY TEST PHASE PER IMAGE Technique Generating pore intensity map (T G ) Postprocessing (T P ) Testing per image (T G + T P ) DeepPore 20 ms 30 ms 50 ms Labati et al. [8] 2 ms 30 ms 32 ms detection results of DeepPore and the existing method. In order to visualize the experimental results, true positive, false negative, and false positive are represented by green, yellow, and red, respectively. These results appear to suggest that the performance of DeepPore is superior to the comparison method. Fig. 5. Average performance metrics for networks with varying depths. TABLE II AVERAGE PERFORMANCE METRICS ON CHANGES IN NETWORKS AND POSTPROCESSING IN PERCENTAGE AND STANDARD DEVIATION (IN PARENTHESIS) CNN Labati with PIR CNN DP with BT R T (6.43) (8.57) R F (6.79) (3.61) state-of-the-art technique, are set as comparative experiments. The combination of Labati s networks and PIR is superior to that of Labati s method, but it is not as good as DeepPore. The combination of CNN DP and BT also gives the same result. These results show that both CNN DP and PIR contribute to the performance improvement. B. Qualitative Evaluation We evaluate the qualitative performance in comparison with the state-of-the-art method [8]. Fig. 4 visually shows the pore C. Computational Time Analysis In real applications, a short amount of time is required for pore extraction. We analyze the testing time of DeepPore and compare with the existing method [8]. We executed the tests using a quad-core machine with 4.0 GHz CPU, RAM 32 GB, and Nvidia GTX GB GPU. The operating system was Ubuntu LTS 64-bit and all methods were implemented using MATLAB. Table III demonstrates the average times for generating pore intensity map (T G ), the average times for postprocessing (T P ), and the average testing times that totalizes T G and T P. Since DeepPore has more layers than the comparison method [8], it needs more time for pore intensity map generation. However, the total testing time of DeepPore is as fast as 50 ms, so the pore extraction is possible in real time. V. CONCLUSION Adaptive pore extraction is a difficult problem because the pore information depends on the person, region, and pore type. To solve the problem, we have presented a pore extraction method using deep CNN and pore intensity refinement. We have demonstrated that our DeepPore outperforms the state-ofthe-art methods by a large margin on the benchmark database. In the future, we will research various biometric systems based on pores.

5 1812 IEEE SIGNAL PROCESSING LETTERS, VOL. 24, NO. 12, DECEMBER 2017 REFERENCES [1] A. K. Jain, Y. Chen, and M. Demirkus, Pores and ridges: High-resolution fingerprint matching using level 3 features, IEEE Trans. Pattern Anal. Mach. Intell., vol. 29, no. 1, pp , Jan [2] Q. Zhao et al., High resolution partial fingerprint alignment using pore valley descriptors, Pattern Recognit., vol. 43, no. 3, pp , [3] D. R. Ashbaugh, Quantitative-Qualitative Friction Ridge Analysis: An Introduction to Basic and Advanced Ridgeology. Boca Raton, FL, USA: CRC press, [4] D. Zhang, F. Liu, Q. Zhao, G. Lu, and N. Luo, Selecting a reference high resolution for fingerprint recognition using minutiae and pores, IEEE Trans. Instrum. Meas., vol. 60, no. 3, pp , Mar [5] P. Johnson and S. Schuckers, Fingerprint pore characteristics for liveness detection, in Proc. Int. Conf. Biometrics Special Interest Group, 2014, pp [6] N. R. Parsons et al., Rotationally invariant statistics for examining the evidence from the pores in fingerprints, Law, Probab. Risk, vol. 7, no. 1, pp. 1 14, [7] Q. Zhao et al., Adaptive fingerprint pore modeling and extraction, Pattern Recognit., vol. 43, no. 8, pp , [8] R. D. Labati et al., A novel pore extraction method for heterogeneous fingerprint images using convolutional neural networks, Pattern Recognit. Lett., [9] N. Otsu, A threshold selection method from gray-level histograms, IEEE Trans. Syst., Man, Cybern., vol. TSMC-9, no. 1, pp , Jan [10] K. Simonyan and A. Zisserman, Very deep convolutional networks for large-scale image recognition, arxiv: , [11] T. H. Cormen, Introduction to Algorithms. MIT press, [12] Y. Jiaet al., Caffe: Convolutional architecture for fast feature embedding, in Proc. 22nd ACM Int. Conf. Multimedia, 2014, pp [13] D. Kingma and J. Ba, Adam: A method for stochastic optimization, arxiv: , [14] R. Pascanu, T. Mikolov, and Y. Bengio, On the difficulty of training recurrent neural networks, in Proc. Int. Conf. Mach. Learn., 2013, pp [15] R. Kohavi et al., A study of cross-validation and bootstrap for accuracy estimation and model selection, in Proc. Int. Joint Conf. Artif., Stanford, CA, USA, vol. 14, no , pp

Real-Time Depth Estimation from 2D Images

Real-Time Depth Estimation from 2D Images Real-Time Depth Estimation from 2D Images Jack Zhu Ralph Ma jackzhu@stanford.edu ralphma@stanford.edu. Abstract ages. We explore the differences in training on an untrained network, and on a network pre-trained

More information

Robust Face Recognition Based on Convolutional Neural Network

Robust Face Recognition Based on Convolutional Neural Network 2017 2nd International Conference on Manufacturing Science and Information Engineering (ICMSIE 2017) ISBN: 978-1-60595-516-2 Robust Face Recognition Based on Convolutional Neural Network Ying Xu, Hui Ma,

More information

Finger Print Enhancement Using Minutiae Based Algorithm

Finger Print Enhancement Using Minutiae Based Algorithm Available Online at www.ijcsmc.com International Journal of Computer Science and Mobile Computing A Monthly Journal of Computer Science and Information Technology IJCSMC, Vol. 3, Issue. 8, August 2014,

More information

Smart Content Recognition from Images Using a Mixture of Convolutional Neural Networks *

Smart Content Recognition from Images Using a Mixture of Convolutional Neural Networks * Smart Content Recognition from Images Using a Mixture of Convolutional Neural Networks * Tee Connie *, Mundher Al-Shabi *, and Michael Goh Faculty of Information Science and Technology, Multimedia University,

More information

Content-Based Image Recovery

Content-Based Image Recovery Content-Based Image Recovery Hong-Yu Zhou and Jianxin Wu National Key Laboratory for Novel Software Technology Nanjing University, China zhouhy@lamda.nju.edu.cn wujx2001@nju.edu.cn Abstract. We propose

More information

Logical Templates for Feature Extraction in Fingerprint Images

Logical Templates for Feature Extraction in Fingerprint Images Logical Templates for Feature Extraction in Fingerprint Images Bir Bhanu, Michael Boshra and Xuejun Tan Center for Research in Intelligent Systems University of Califomia, Riverside, CA 9252 1, USA Email:

More information

Fingerprint Mosaicking by Rolling with Sliding

Fingerprint Mosaicking by Rolling with Sliding Fingerprint Mosaicking by Rolling with Sliding Kyoungtaek Choi, Hunjae Park, Hee-seung Choi and Jaihie Kim Department of Electrical and Electronic Engineering,Yonsei University Biometrics Engineering Research

More information

Improving Latent Fingerprint Matching Performance by Orientation Field Estimation using Localized Dictionaries

Improving Latent Fingerprint Matching Performance by Orientation Field Estimation using Localized Dictionaries Available Online at www.ijcsmc.com International Journal of Computer Science and Mobile Computing A Monthly Journal of Computer Science and Information Technology IJCSMC, Vol. 3, Issue. 11, November 2014,

More information

Robot localization method based on visual features and their geometric relationship

Robot localization method based on visual features and their geometric relationship , pp.46-50 http://dx.doi.org/10.14257/astl.2015.85.11 Robot localization method based on visual features and their geometric relationship Sangyun Lee 1, Changkyung Eem 2, and Hyunki Hong 3 1 Department

More information

An Area-Efficient BIRA With 1-D Spare Segments

An Area-Efficient BIRA With 1-D Spare Segments 206 IEEE TRANSACTIONS ON VERY LARGE SCALE INTEGRATION (VLSI) SYSTEMS, VOL. 26, NO. 1, JANUARY 2018 An Area-Efficient BIRA With 1-D Spare Segments Donghyun Kim, Hayoung Lee, and Sungho Kang Abstract The

More information

Deep Face Recognition. Nathan Sun

Deep Face Recognition. Nathan Sun Deep Face Recognition Nathan Sun Why Facial Recognition? Picture ID or video tracking Higher Security for Facial Recognition Software Immensely useful to police in tracking suspects Your face will be an

More information

Proceedings of the International MultiConference of Engineers and Computer Scientists 2018 Vol I IMECS 2018, March 14-16, 2018, Hong Kong

Proceedings of the International MultiConference of Engineers and Computer Scientists 2018 Vol I IMECS 2018, March 14-16, 2018, Hong Kong , March 14-16, 2018, Hong Kong , March 14-16, 2018, Hong Kong , March 14-16, 2018, Hong Kong , March 14-16, 2018, Hong Kong TABLE I CLASSIFICATION ACCURACY OF DIFFERENT PRE-TRAINED MODELS ON THE TEST DATA

More information

ASCII Art Synthesis with Convolutional Networks

ASCII Art Synthesis with Convolutional Networks ASCII Art Synthesis with Convolutional Networks Osamu Akiyama Faculty of Medicine, Osaka University oakiyama1986@gmail.com 1 Introduction ASCII art is a type of graphic art that presents a picture with

More information

A NEW FEATURE BASED IMAGE REGISTRATION ALGORITHM INTRODUCTION

A NEW FEATURE BASED IMAGE REGISTRATION ALGORITHM INTRODUCTION A NEW FEATURE BASED IMAGE REGISTRATION ALGORITHM Karthik Krish Stuart Heinrich Wesley E. Snyder Halil Cakir Siamak Khorram North Carolina State University Raleigh, 27695 kkrish@ncsu.edu sbheinri@ncsu.edu

More information

Content Based Image Retrieval Using Color Quantizes, EDBTC and LBP Features

Content Based Image Retrieval Using Color Quantizes, EDBTC and LBP Features Content Based Image Retrieval Using Color Quantizes, EDBTC and LBP Features 1 Kum Sharanamma, 2 Krishnapriya Sharma 1,2 SIR MVIT Abstract- To describe the image features the Local binary pattern (LBP)

More information

Deep Tracking: Biologically Inspired Tracking with Deep Convolutional Networks

Deep Tracking: Biologically Inspired Tracking with Deep Convolutional Networks Deep Tracking: Biologically Inspired Tracking with Deep Convolutional Networks Si Chen The George Washington University sichen@gwmail.gwu.edu Meera Hahn Emory University mhahn7@emory.edu Mentor: Afshin

More information

Locating 1-D Bar Codes in DCT-Domain

Locating 1-D Bar Codes in DCT-Domain Edith Cowan University Research Online ECU Publications Pre. 2011 2006 Locating 1-D Bar Codes in DCT-Domain Alexander Tropf Edith Cowan University Douglas Chai Edith Cowan University 10.1109/ICASSP.2006.1660449

More information

Adaptive Fingerprint Pore Model for Fingerprint Pore Extraction

Adaptive Fingerprint Pore Model for Fingerprint Pore Extraction RESEARCH ARTICLE OPEN ACCESS Adaptive Fingerprint Pore Model for Fingerprint Pore Extraction Ritesh B.Siriya, Milind M.Mushrif Dept. of E&T, YCCE, Dept. of E&T, YCCE ritesh.siriya@gmail.com, milindmushrif@yahoo.com

More information

Object Recognition II

Object Recognition II Object Recognition II Linda Shapiro EE/CSE 576 with CNN slides from Ross Girshick 1 Outline Object detection the task, evaluation, datasets Convolutional Neural Networks (CNNs) overview and history Region-based

More information

Finding Tiny Faces Supplementary Materials

Finding Tiny Faces Supplementary Materials Finding Tiny Faces Supplementary Materials Peiyun Hu, Deva Ramanan Robotics Institute Carnegie Mellon University {peiyunh,deva}@cs.cmu.edu 1. Error analysis Quantitative analysis We plot the distribution

More information

RESTORATION OF DEGRADED DOCUMENTS USING IMAGE BINARIZATION TECHNIQUE

RESTORATION OF DEGRADED DOCUMENTS USING IMAGE BINARIZATION TECHNIQUE RESTORATION OF DEGRADED DOCUMENTS USING IMAGE BINARIZATION TECHNIQUE K. Kaviya Selvi 1 and R. S. Sabeenian 2 1 Department of Electronics and Communication Engineering, Communication Systems, Sona College

More information

Synopsis. An Efficient Approach for Partial Fingerprint Recognition Based on Pores and SIFT Features using Fusion Methods

Synopsis. An Efficient Approach for Partial Fingerprint Recognition Based on Pores and SIFT Features using Fusion Methods Synopsis An Efficient Approach for Partial Fingerprint Recognition Based on Pores and SIFT Features using Fusion Methods Submitted By Mrs.S.Malathi Supervisor Dr(Mrs.) C.Meena Submitted To Avinashilingam

More information

Where s Waldo? A Deep Learning approach to Template Matching

Where s Waldo? A Deep Learning approach to Template Matching Where s Waldo? A Deep Learning approach to Template Matching Thomas Hossler Department of Geological Sciences Stanford University thossler@stanford.edu Abstract We propose a new approach to Template Matching

More information

arxiv: v1 [cs.cv] 20 Dec 2016

arxiv: v1 [cs.cv] 20 Dec 2016 End-to-End Pedestrian Collision Warning System based on a Convolutional Neural Network with Semantic Segmentation arxiv:1612.06558v1 [cs.cv] 20 Dec 2016 Heechul Jung heechul@dgist.ac.kr Min-Kook Choi mkchoi@dgist.ac.kr

More information

Convolution Neural Network for Traditional Chinese Calligraphy Recognition

Convolution Neural Network for Traditional Chinese Calligraphy Recognition Convolution Neural Network for Traditional Chinese Calligraphy Recognition Boqi Li Mechanical Engineering Stanford University boqili@stanford.edu Abstract script. Fig. 1 shows examples of the same TCC

More information

III. VERVIEW OF THE METHODS

III. VERVIEW OF THE METHODS An Analytical Study of SIFT and SURF in Image Registration Vivek Kumar Gupta, Kanchan Cecil Department of Electronics & Telecommunication, Jabalpur engineering college, Jabalpur, India comparing the distance

More information

RSRN: Rich Side-output Residual Network for Medial Axis Detection

RSRN: Rich Side-output Residual Network for Medial Axis Detection RSRN: Rich Side-output Residual Network for Medial Axis Detection Chang Liu, Wei Ke, Jianbin Jiao, and Qixiang Ye University of Chinese Academy of Sciences, Beijing, China {liuchang615, kewei11}@mails.ucas.ac.cn,

More information

A Laplacian Based Novel Approach to Efficient Text Localization in Grayscale Images

A Laplacian Based Novel Approach to Efficient Text Localization in Grayscale Images A Laplacian Based Novel Approach to Efficient Text Localization in Grayscale Images Karthik Ram K.V & Mahantesh K Department of Electronics and Communication Engineering, SJB Institute of Technology, Bangalore,

More information

A performance comparison of Deep Learning frameworks on KNL

A performance comparison of Deep Learning frameworks on KNL A performance comparison of Deep Learning frameworks on KNL R. Zanella, G. Fiameni, M. Rorro Middleware, Data Management - SCAI - CINECA IXPUG Bologna, March 5, 2018 Table of Contents 1. Problem description

More information

Semi-supervised Data Representation via Affinity Graph Learning

Semi-supervised Data Representation via Affinity Graph Learning 1 Semi-supervised Data Representation via Affinity Graph Learning Weiya Ren 1 1 College of Information System and Management, National University of Defense Technology, Changsha, Hunan, P.R China, 410073

More information

Face Recognition Using Vector Quantization Histogram and Support Vector Machine Classifier Rong-sheng LI, Fei-fei LEE *, Yan YAN and Qiu CHEN

Face Recognition Using Vector Quantization Histogram and Support Vector Machine Classifier Rong-sheng LI, Fei-fei LEE *, Yan YAN and Qiu CHEN 2016 International Conference on Artificial Intelligence: Techniques and Applications (AITA 2016) ISBN: 978-1-60595-389-2 Face Recognition Using Vector Quantization Histogram and Support Vector Machine

More information

CORRELATION BASED CAR NUMBER PLATE EXTRACTION SYSTEM

CORRELATION BASED CAR NUMBER PLATE EXTRACTION SYSTEM CORRELATION BASED CAR NUMBER PLATE EXTRACTION SYSTEM 1 PHYO THET KHIN, 2 LAI LAI WIN KYI 1,2 Department of Information Technology, Mandalay Technological University The Republic of the Union of Myanmar

More information

Convolution Neural Networks for Chinese Handwriting Recognition

Convolution Neural Networks for Chinese Handwriting Recognition Convolution Neural Networks for Chinese Handwriting Recognition Xu Chen Stanford University 450 Serra Mall, Stanford, CA 94305 xchen91@stanford.edu Abstract Convolutional neural networks have been proven

More information

Edge Histogram Descriptor, Geometric Moment and Sobel Edge Detector Combined Features Based Object Recognition and Retrieval System

Edge Histogram Descriptor, Geometric Moment and Sobel Edge Detector Combined Features Based Object Recognition and Retrieval System Edge Histogram Descriptor, Geometric Moment and Sobel Edge Detector Combined Features Based Object Recognition and Retrieval System Neetesh Prajapati M. Tech Scholar VNS college,bhopal Amit Kumar Nandanwar

More information

Published by: PIONEER RESEARCH & DEVELOPMENT GROUP (www.prdg.org) 1

Published by: PIONEER RESEARCH & DEVELOPMENT GROUP (www.prdg.org) 1 Minutiae Points Extraction using Biometric Fingerprint- Enhancement Vishal Wagh 1, Shefali Sonavane 2 1 Computer Science and Engineering Department, Walchand College of Engineering, Sangli, Maharashtra-416415,

More information

IN THE LAST decade, the amount of video contents digitally

IN THE LAST decade, the amount of video contents digitally IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY, VOL. 18, NO. 7, JULY 2008 983 Robust Video Fingerprinting for Content-Based Video Identification Sunil Lee, Member, IEEE, and Chang D. Yoo,

More information

Machine Learning Methods for Ship Detection in Satellite Images

Machine Learning Methods for Ship Detection in Satellite Images Machine Learning Methods for Ship Detection in Satellite Images Yifan Li yil150@ucsd.edu Huadong Zhang huz095@ucsd.edu Qianfeng Guo qig020@ucsd.edu Xiaoshi Li xil758@ucsd.edu Abstract In this project,

More information

Fingerprint Recognition using Fuzzy based image Enhancement

Fingerprint Recognition using Fuzzy based image Enhancement Fingerprint Recognition using Fuzzy based image Enhancement BhartiYadav 1, Ram NivasGiri 2 P.G. Student, Department of Computer Engineering, Raipur Institute of technology, Raipur, Chhattisgarh, India

More information

CHAPTER 6 MODIFIED FUZZY TECHNIQUES BASED IMAGE SEGMENTATION

CHAPTER 6 MODIFIED FUZZY TECHNIQUES BASED IMAGE SEGMENTATION CHAPTER 6 MODIFIED FUZZY TECHNIQUES BASED IMAGE SEGMENTATION 6.1 INTRODUCTION Fuzzy logic based computational techniques are becoming increasingly important in the medical image analysis arena. The significant

More information

I. INTRODUCTION. Image Acquisition. Denoising in Wavelet Domain. Enhancement. Binarization. Thinning. Feature Extraction. Matching

I. INTRODUCTION. Image Acquisition. Denoising in Wavelet Domain. Enhancement. Binarization. Thinning. Feature Extraction. Matching A Comparative Analysis on Fingerprint Binarization Techniques K Sasirekha Department of Computer Science Periyar University Salem, Tamilnadu Ksasirekha7@gmail.com K Thangavel Department of Computer Science

More information

Multi prototype fuzzy pattern matching for handwritten character recognition

Multi prototype fuzzy pattern matching for handwritten character recognition Multi prototype fuzzy pattern matching for handwritten character recognition MILIND E. RANE, DHABE P. S AND J. B. PATIL Dept. of Electronics and Computer, R.C. Patel Institute of Technology, Shirpur, Dist.

More information

arxiv: v1 [cs.cv] 11 Apr 2018

arxiv: v1 [cs.cv] 11 Apr 2018 Unsupervised Segmentation of 3D Medical Images Based on Clustering and Deep Representation Learning Takayasu Moriya a, Holger R. Roth a, Shota Nakamura b, Hirohisa Oda c, Kai Nagara c, Masahiro Oda a,

More information

Appearance-Based Place Recognition Using Whole-Image BRISK for Collaborative MultiRobot Localization

Appearance-Based Place Recognition Using Whole-Image BRISK for Collaborative MultiRobot Localization Appearance-Based Place Recognition Using Whole-Image BRISK for Collaborative MultiRobot Localization Jung H. Oh, Gyuho Eoh, and Beom H. Lee Electrical and Computer Engineering, Seoul National University,

More information

Additive Manufacturing Defect Detection using Neural Networks

Additive Manufacturing Defect Detection using Neural Networks Additive Manufacturing Defect Detection using Neural Networks James Ferguson Department of Electrical Engineering and Computer Science University of Tennessee Knoxville Knoxville, Tennessee 37996 Jfergu35@vols.utk.edu

More information

Stochastic Function Norm Regularization of DNNs

Stochastic Function Norm Regularization of DNNs Stochastic Function Norm Regularization of DNNs Amal Rannen Triki Dept. of Computational Science and Engineering Yonsei University Seoul, South Korea amal.rannen@yonsei.ac.kr Matthew B. Blaschko Center

More information

Deep Learning with Tensorflow AlexNet

Deep Learning with Tensorflow   AlexNet Machine Learning and Computer Vision Group Deep Learning with Tensorflow http://cvml.ist.ac.at/courses/dlwt_w17/ AlexNet Krizhevsky, Alex, Ilya Sutskever, and Geoffrey E. Hinton, "Imagenet classification

More information

Fingerprint Ridge Orientation Estimation Using A Modified Canny Edge Detection Mask

Fingerprint Ridge Orientation Estimation Using A Modified Canny Edge Detection Mask Fingerprint Ridge Orientation Estimation Using A Modified Canny Edge Detection Mask Laurice Phillips PhD student laurice.phillips@utt.edu.tt Margaret Bernard Senior Lecturer and Head of Department Margaret.Bernard@sta.uwi.edu

More information

Convolutional Neural Network Layer Reordering for Acceleration

Convolutional Neural Network Layer Reordering for Acceleration R1-15 SASIMI 2016 Proceedings Convolutional Neural Network Layer Reordering for Acceleration Vijay Daultani Subhajit Chaudhury Kazuhisa Ishizaka System Platform Labs Value Co-creation Center System Platform

More information

Show, Discriminate, and Tell: A Discriminatory Image Captioning Model with Deep Neural Networks

Show, Discriminate, and Tell: A Discriminatory Image Captioning Model with Deep Neural Networks Show, Discriminate, and Tell: A Discriminatory Image Captioning Model with Deep Neural Networks Zelun Luo Department of Computer Science Stanford University zelunluo@stanford.edu Te-Lin Wu Department of

More information

Logo Matching and Recognition for Avoiding Duplicate Logos

Logo Matching and Recognition for Avoiding Duplicate Logos Logo Matching and Recognition for Avoiding Duplicate Logos Lalsawmliani Fanchun 1, Rishma Mary George 2 PG Student, Electronics & Ccommunication Department, Mangalore Institute of Technology and Engineering

More information

CPSC 695. Geometric Algorithms in Biometrics. Dr. Marina L. Gavrilova

CPSC 695. Geometric Algorithms in Biometrics. Dr. Marina L. Gavrilova CPSC 695 Geometric Algorithms in Biometrics Dr. Marina L. Gavrilova Biometric goals Verify users Identify users Synthesis - recently Biometric identifiers Courtesy of Bromba GmbH Classification of identifiers

More information

Fusion of Hand Geometry and Palmprint Biometrics

Fusion of Hand Geometry and Palmprint Biometrics (Working Paper, Dec. 2003) Fusion of Hand Geometry and Palmprint Biometrics D.C.M. Wong, C. Poon and H.C. Shen * Department of Computer Science, Hong Kong University of Science and Technology, Clear Water

More information

Dot Text Detection Based on FAST Points

Dot Text Detection Based on FAST Points Dot Text Detection Based on FAST Points Yuning Du, Haizhou Ai Computer Science & Technology Department Tsinghua University Beijing, China dyn10@mails.tsinghua.edu.cn, ahz@mail.tsinghua.edu.cn Shihong Lao

More information

Development of an Automated Fingerprint Verification System

Development of an Automated Fingerprint Verification System Development of an Automated Development of an Automated Fingerprint Verification System Fingerprint Verification System Martin Saveski 18 May 2010 Introduction Biometrics the use of distinctive anatomical

More information

4/13/ Introduction. 1. Introduction. 2. Formulation. 2. Formulation. 2. Formulation

4/13/ Introduction. 1. Introduction. 2. Formulation. 2. Formulation. 2. Formulation 1. Introduction Motivation: Beijing Jiaotong University 1 Lotus Hill Research Institute University of California, Los Angeles 3 CO 3 for Ultra-fast and Accurate Interactive Image Segmentation This paper

More information

Color Local Texture Features Based Face Recognition

Color Local Texture Features Based Face Recognition Color Local Texture Features Based Face Recognition Priyanka V. Bankar Department of Electronics and Communication Engineering SKN Sinhgad College of Engineering, Korti, Pandharpur, Maharashtra, India

More information

Fingerprint Matching Incorporating Ridge Features Using Contourlet Transforms

Fingerprint Matching Incorporating Ridge Features Using Contourlet Transforms Fingerprint Matching Incorporating Ridge Features Using Contourlet Transforms M.S. Keerthana 1 Student,Department of CSE, K.S.Rangasamy College Of Technology,Tiruchengode,TamilNadu, India 1 ABSTRACT: This

More information

A Comparison of SIFT, PCA-SIFT and SURF

A Comparison of SIFT, PCA-SIFT and SURF A Comparison of SIFT, PCA-SIFT and SURF Luo Juan Computer Graphics Lab, Chonbuk National University, Jeonju 561-756, South Korea qiuhehappy@hotmail.com Oubong Gwun Computer Graphics Lab, Chonbuk National

More information

Real-time Object Detection CS 229 Course Project

Real-time Object Detection CS 229 Course Project Real-time Object Detection CS 229 Course Project Zibo Gong 1, Tianchang He 1, and Ziyi Yang 1 1 Department of Electrical Engineering, Stanford University December 17, 2016 Abstract Objection detection

More information

Image Enhancement Techniques for Fingerprint Identification

Image Enhancement Techniques for Fingerprint Identification March 2013 1 Image Enhancement Techniques for Fingerprint Identification Pankaj Deshmukh, Siraj Pathan, Riyaz Pathan Abstract The aim of this paper is to propose a new method in fingerprint enhancement

More information

Kaggle Data Science Bowl 2017 Technical Report

Kaggle Data Science Bowl 2017 Technical Report Kaggle Data Science Bowl 2017 Technical Report qfpxfd Team May 11, 2017 1 Team Members Table 1: Team members Name E-Mail University Jia Ding dingjia@pku.edu.cn Peking University, Beijing, China Aoxue Li

More information

Inception and Residual Networks. Hantao Zhang. Deep Learning with Python.

Inception and Residual Networks. Hantao Zhang. Deep Learning with Python. Inception and Residual Networks Hantao Zhang Deep Learning with Python https://en.wikipedia.org/wiki/residual_neural_network Deep Neural Network Progress from Large Scale Visual Recognition Challenge (ILSVRC)

More information

Facial Key Points Detection using Deep Convolutional Neural Network - NaimishNet

Facial Key Points Detection using Deep Convolutional Neural Network - NaimishNet 1 Facial Key Points Detection using Deep Convolutional Neural Network - NaimishNet Naimish Agarwal, IIIT-Allahabad (irm2013013@iiita.ac.in) Artus Krohn-Grimberghe, University of Paderborn (artus@aisbi.de)

More information

A new predictive image compression scheme using histogram analysis and pattern matching

A new predictive image compression scheme using histogram analysis and pattern matching University of Wollongong Research Online University of Wollongong in Dubai - Papers University of Wollongong in Dubai 00 A new predictive image compression scheme using histogram analysis and pattern matching

More information

Detecting Salient Contours Using Orientation Energy Distribution. Part I: Thresholding Based on. Response Distribution

Detecting Salient Contours Using Orientation Energy Distribution. Part I: Thresholding Based on. Response Distribution Detecting Salient Contours Using Orientation Energy Distribution The Problem: How Does the Visual System Detect Salient Contours? CPSC 636 Slide12, Spring 212 Yoonsuck Choe Co-work with S. Sarma and H.-C.

More information

A Statistical approach to line segmentation in handwritten documents

A Statistical approach to line segmentation in handwritten documents A Statistical approach to line segmentation in handwritten documents Manivannan Arivazhagan, Harish Srinivasan and Sargur Srihari Center of Excellence for Document Analysis and Recognition (CEDAR) University

More information

PATCH BASED LATENT FINGERPRINT MATCHING USING DEEP LEARNING. Jude Ezeobiejesi and Bir Bhanu

PATCH BASED LATENT FINGERPRINT MATCHING USING DEEP LEARNING. Jude Ezeobiejesi and Bir Bhanu PATCH BASED LATENT FINGERPRINT MATCHING USING DEEP LEARNING Jude Ezeobiejesi and Bir Bhanu Center for Research in Intelligent Systems University of California at Riverside, Riverside, CA 92521, USA e-mail:

More information

Joint Object Detection and Viewpoint Estimation using CNN features

Joint Object Detection and Viewpoint Estimation using CNN features Joint Object Detection and Viewpoint Estimation using CNN features Carlos Guindel, David Martín and José M. Armingol cguindel@ing.uc3m.es Intelligent Systems Laboratory Universidad Carlos III de Madrid

More information

Biometrics Technology: Hand Geometry

Biometrics Technology: Hand Geometry Biometrics Technology: Hand Geometry References: [H1] Gonzeilez, S., Travieso, C.M., Alonso, J.B., and M.A. Ferrer, Automatic biometric identification system by hand geometry, Proceedings of IEEE the 37th

More information

Biometric Security System Using Palm print

Biometric Security System Using Palm print ISSN (Online) : 2319-8753 ISSN (Print) : 2347-6710 International Journal of Innovative Research in Science, Engineering and Technology Volume 3, Special Issue 3, March 2014 2014 International Conference

More information

Learning Social Graph Topologies using Generative Adversarial Neural Networks

Learning Social Graph Topologies using Generative Adversarial Neural Networks Learning Social Graph Topologies using Generative Adversarial Neural Networks Sahar Tavakoli 1, Alireza Hajibagheri 1, and Gita Sukthankar 1 1 University of Central Florida, Orlando, Florida sahar@knights.ucf.edu,alireza@eecs.ucf.edu,gitars@eecs.ucf.edu

More information

PAVEMENT distress detection is a significant step for

PAVEMENT distress detection is a significant step for 1 Automatic Pavement Crack Detection Based on Structured Prediction with the Convolutional Neural Network Zhun Fan, Senior Member, IEEE, Yuming Wu, Jiewei Lu, and Wenji Li arxiv:1802.02208v1 [cs.cv] 1

More information

A Fast Caption Detection Method for Low Quality Video Images

A Fast Caption Detection Method for Low Quality Video Images 2012 10th IAPR International Workshop on Document Analysis Systems A Fast Caption Detection Method for Low Quality Video Images Tianyi Gui, Jun Sun, Satoshi Naoi Fujitsu Research & Development Center CO.,

More information

A MORPHOLOGY-BASED FILTER STRUCTURE FOR EDGE-ENHANCING SMOOTHING

A MORPHOLOGY-BASED FILTER STRUCTURE FOR EDGE-ENHANCING SMOOTHING Proceedings of the 1994 IEEE International Conference on Image Processing (ICIP-94), pp. 530-534. (Austin, Texas, 13-16 November 1994.) A MORPHOLOGY-BASED FILTER STRUCTURE FOR EDGE-ENHANCING SMOOTHING

More information

Study of Residual Networks for Image Recognition

Study of Residual Networks for Image Recognition Study of Residual Networks for Image Recognition Mohammad Sadegh Ebrahimi Stanford University sadegh@stanford.edu Hossein Karkeh Abadi Stanford University hosseink@stanford.edu Abstract Deep neural networks

More information

Fast Motion Estimation for Shape Coding in MPEG-4

Fast Motion Estimation for Shape Coding in MPEG-4 358 IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY, VOL. 13, NO. 4, APRIL 2003 Fast Motion Estimation for Shape Coding in MPEG-4 Donghoon Yu, Sung Kyu Jang, and Jong Beom Ra Abstract Effective

More information

Quality Enhancement of Compressed Video via CNNs

Quality Enhancement of Compressed Video via CNNs Journal of Information Hiding and Multimedia Signal Processing c 2017 ISSN 2073-4212 Ubiquitous International Volume 8, Number 1, January 2017 Quality Enhancement of Compressed Video via CNNs Jingxuan

More information

SCALE INVARIANT TEMPLATE MATCHING

SCALE INVARIANT TEMPLATE MATCHING Volume 118 No. 5 2018, 499-505 ISSN: 1311-8080 (printed version); ISSN: 1314-3395 (on-line version) url: http://www.ijpam.eu ijpam.eu SCALE INVARIANT TEMPLATE MATCHING Badrinaathan.J Srm university Chennai,India

More information

Learning based face hallucination techniques: A survey

Learning based face hallucination techniques: A survey Vol. 3 (2014-15) pp. 37-45. : A survey Premitha Premnath K Department of Computer Science & Engineering Vidya Academy of Science & Technology Thrissur - 680501, Kerala, India (email: premithakpnath@gmail.com)

More information

An Adaptive Threshold LBP Algorithm for Face Recognition

An Adaptive Threshold LBP Algorithm for Face Recognition An Adaptive Threshold LBP Algorithm for Face Recognition Xiaoping Jiang 1, Chuyu Guo 1,*, Hua Zhang 1, and Chenghua Li 1 1 College of Electronics and Information Engineering, Hubei Key Laboratory of Intelligent

More information

Dynamic Routing Between Capsules

Dynamic Routing Between Capsules Report Explainable Machine Learning Dynamic Routing Between Capsules Author: Michael Dorkenwald Supervisor: Dr. Ullrich Köthe 28. Juni 2018 Inhaltsverzeichnis 1 Introduction 2 2 Motivation 2 3 CapusleNet

More information

Detecting Printed and Handwritten Partial Copies of Line Drawings Embedded in Complex Backgrounds

Detecting Printed and Handwritten Partial Copies of Line Drawings Embedded in Complex Backgrounds 9 1th International Conference on Document Analysis and Recognition Detecting Printed and Handwritten Partial Copies of Line Drawings Embedded in Complex Backgrounds Weihan Sun, Koichi Kise Graduate School

More information

Structured Prediction using Convolutional Neural Networks

Structured Prediction using Convolutional Neural Networks Overview Structured Prediction using Convolutional Neural Networks Bohyung Han bhhan@postech.ac.kr Computer Vision Lab. Convolutional Neural Networks (CNNs) Structured predictions for low level computer

More information

International Journal of Electrical, Electronics ISSN No. (Online): and Computer Engineering 3(2): 85-90(2014)

International Journal of Electrical, Electronics ISSN No. (Online): and Computer Engineering 3(2): 85-90(2014) I J E E E C International Journal of Electrical, Electronics ISSN No. (Online): 2277-2626 Computer Engineering 3(2): 85-90(2014) Robust Approach to Recognize Localize Text from Natural Scene Images Khushbu

More information

INTRODUCTION TO DEEP LEARNING

INTRODUCTION TO DEEP LEARNING INTRODUCTION TO DEEP LEARNING CONTENTS Introduction to deep learning Contents 1. Examples 2. Machine learning 3. Neural networks 4. Deep learning 5. Convolutional neural networks 6. Conclusion 7. Additional

More information

DeepBIBX: Deep Learning for Image Based Bibliographic Data Extraction

DeepBIBX: Deep Learning for Image Based Bibliographic Data Extraction DeepBIBX: Deep Learning for Image Based Bibliographic Data Extraction Akansha Bhardwaj 1,2, Dominik Mercier 1, Sheraz Ahmed 1, Andreas Dengel 1 1 Smart Data and Services, DFKI Kaiserslautern, Germany firstname.lastname@dfki.de

More information

Keywords:- Fingerprint Identification, Hong s Enhancement, Euclidian Distance, Artificial Neural Network, Segmentation, Enhancement.

Keywords:- Fingerprint Identification, Hong s Enhancement, Euclidian Distance, Artificial Neural Network, Segmentation, Enhancement. Volume 5, Issue 8, August 2015 ISSN: 2277 128X International Journal of Advanced Research in Computer Science and Software Engineering Research Paper Available online at: www.ijarcsse.com Embedded Algorithm

More information

Distorted Fingerprint Verification System

Distorted Fingerprint Verification System Informatica Economică vol. 15, no. 4/2011 13 Distorted Fingerprint Verification System Divya KARTHIKAESHWARAN 1, Jeyalatha SIVARAMAKRISHNAN 2 1 Department of Computer Science, Amrita University, Bangalore,

More information

/$ IEEE

/$ IEEE IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS II: EXPRESS BRIEFS, VOL. 56, NO. 1, JANUARY 2009 81 Bit-Level Extrinsic Information Exchange Method for Double-Binary Turbo Codes Ji-Hoon Kim, Student Member,

More information

Direct Multi-Scale Dual-Stream Network for Pedestrian Detection Sang-Il Jung and Ki-Sang Hong Image Information Processing Lab.

Direct Multi-Scale Dual-Stream Network for Pedestrian Detection Sang-Il Jung and Ki-Sang Hong Image Information Processing Lab. [ICIP 2017] Direct Multi-Scale Dual-Stream Network for Pedestrian Detection Sang-Il Jung and Ki-Sang Hong Image Information Processing Lab., POSTECH Pedestrian Detection Goal To draw bounding boxes that

More information

Study of Local Binary Pattern for Partial Fingerprint Identification

Study of Local Binary Pattern for Partial Fingerprint Identification International OPEN ACCESS Journal Of Modern Engineering Research (IJMER) Study of Local Binary Pattern for Partial Fingerprint Identification Miss Harsha V. Talele 1, Pratvina V. Talele 2, Saranga N Bhutada

More information

TRAFFIC SIGN RECOGNITION USING A MULTI-TASK CONVOLUTIONAL NEURAL NETWORK

TRAFFIC SIGN RECOGNITION USING A MULTI-TASK CONVOLUTIONAL NEURAL NETWORK TRAFFIC SIGN RECOGNITION USING A MULTI-TASK CONVOLUTIONAL NEURAL NETWORK Dr. S.V. Shinde Arshiya Sayyad Uzma Shaikh Department of IT Department of IT Department of IT Pimpri Chinchwad college of Engineering

More information

AN EFFICIENT VIDEO WATERMARKING USING COLOR HISTOGRAM ANALYSIS AND BITPLANE IMAGE ARRAYS

AN EFFICIENT VIDEO WATERMARKING USING COLOR HISTOGRAM ANALYSIS AND BITPLANE IMAGE ARRAYS AN EFFICIENT VIDEO WATERMARKING USING COLOR HISTOGRAM ANALYSIS AND BITPLANE IMAGE ARRAYS G Prakash 1,TVS Gowtham Prasad 2, T.Ravi Kumar Naidu 3 1MTech(DECS) student, Department of ECE, sree vidyanikethan

More information

RESEARCH ON OPTIMIZATION OF IMAGE USING SKELETONIZATION TECHNIQUE WITH ADVANCED ALGORITHM

RESEARCH ON OPTIMIZATION OF IMAGE USING SKELETONIZATION TECHNIQUE WITH ADVANCED ALGORITHM 881 RESEARCH ON OPTIMIZATION OF IMAGE USING SKELETONIZATION TECHNIQUE WITH ADVANCED ALGORITHM Sarita Jain 1 Sumit Rana 2 Department of CSE 1 Department of CSE 2 Geeta Engineering College 1, Panipat, India

More information

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

International Journal of Computer Engineering and Applications, Volume XII, Special Issue, September 18, REAL-TIME OBJECT DETECTION WITH CONVOLUTION NEURAL NETWORK USING KERAS Asmita Goswami [1], Lokesh Soni [2 ] Department of Information Technology [1] Jaipur Engineering College and Research Center Jaipur[2]

More information

AN EFFICIENT BINARIZATION TECHNIQUE FOR FINGERPRINT IMAGES S. B. SRIDEVI M.Tech., Department of ECE

AN EFFICIENT BINARIZATION TECHNIQUE FOR FINGERPRINT IMAGES S. B. SRIDEVI M.Tech., Department of ECE AN EFFICIENT BINARIZATION TECHNIQUE FOR FINGERPRINT IMAGES S. B. SRIDEVI M.Tech., Department of ECE sbsridevi89@gmail.com 287 ABSTRACT Fingerprint identification is the most prominent method of biometric

More information

arxiv: v1 [cs.cv] 14 Dec 2016

arxiv: v1 [cs.cv] 14 Dec 2016 Detect, Replace, Refine: Deep Structured Prediction For Pixel Wise Labeling arxiv:1612.04770v1 [cs.cv] 14 Dec 2016 Spyros Gidaris University Paris-Est, LIGM Ecole des Ponts ParisTech spyros.gidaris@imagine.enpc.fr

More information

Fingerprint Matching using Gabor Filters

Fingerprint Matching using Gabor Filters Fingerprint Matching using Gabor Filters Muhammad Umer Munir and Dr. Muhammad Younas Javed College of Electrical and Mechanical Engineering, National University of Sciences and Technology Rawalpindi, Pakistan.

More information

Fingerprint Image Enhancement Algorithm and Performance Evaluation

Fingerprint Image Enhancement Algorithm and Performance Evaluation Fingerprint Image Enhancement Algorithm and Performance Evaluation Naja M I, Rajesh R M Tech Student, College of Engineering, Perumon, Perinad, Kerala, India Project Manager, NEST GROUP, Techno Park, TVM,

More information

Face Detection Using Convolutional Neural Networks and Gabor Filters

Face Detection Using Convolutional Neural Networks and Gabor Filters Face Detection Using Convolutional Neural Networks and Gabor Filters Bogdan Kwolek Rzeszów University of Technology W. Pola 2, 35-959 Rzeszów, Poland bkwolek@prz.rzeszow.pl Abstract. This paper proposes

More information