SYNOPSIS FOR M.E. DISSERTATION M.E. Electronics-Part: II (SEMISTER-III) 1. NAME OF THE COLLEGE : - KARMAVEER BHAURAO PATIL COLLEGE OF ENGINEERING, SATARA. 2. NAME OF THE COURSE : - M.E. Electronics 3. NAME OF THE STUDENT (With PRN No.) : - MR.MULLA VASIM BASHIR (PRN NO - 1412703201) 4. DATE OF REGISTRATION : - August 2016 5. NAME OF THE GUIDE : - Dr. Patil Vikram S. Department of Electronics, KBPCOE, Satara. (With PG Recognition No. And Date) 6. PROPOSED TITLE Student sign : - SU/PG/BUTR/RECOG/2254/15 /06/2012 : - Real time Access Control Using Face Recognition With MultipleVisitor Detection. Guide sign 7. INTRODUCTION - Face recognition is a form of biometric identification involving recognition of individuals based on the salient characteristics of their face images [1]. Be it for government use such as law enforcement, voter identification, surveillance and immigration, or for commercial use such as gaming industry, face tagging on internet, e-commerce, healthcare and banking, a large number of real world applications utilize face recognition. As a result, there has been enormous interest in this area of research. Face recognition system is a complex image-processing problem in real world applications with complex effects of illumination, occlusion, and imaging condition on the live images. It is a combination of face detection and recognition techniques in image analyzes. Detection application is used to find position of the faces in a given image. Recognition algorithm is used to classify given images with known structured properties, which are used commonly in most of the computer vision applications. These images have some known properties like; same resolution, including same facial feature components, and similar eye alignment. These images will be referred as “standard image” in the further sections. Recognition applications uses standard images and detection algorithms detect the faces and extract face images which include eyes, eyebrows, nose, and mouth. That makes the algorithm more complicated than single detection or recognition algorithm. The first step for face recognition system is to acquire an image from a camera. Second step is face detection from the acquired image. As a third step, face recognition that takes the face images from output of detection part. Final step is person identity as a result of recognition part. Fig.1. Steps of Face Recognition System Applications Student sign Guide sign Types of biometrics Biometrics can measure both physiological and behavioural characteristics. Fig 2. Classification of biometrics. Why we choose face recognition over other biometric? There are number of reasons to choose face recognition. This includes the following1. It requires no physical interaction on behalf of the user. 2. It is accurate and allows for high enrolment and verification rates. 3. It does not require an expert to interpret the comparison result. 4. It can use your existing hardware infrastructure; existing cameras and image capture devices will work with no problems. 5. It is the only biometric that allow you to perform passive identification in a one to many environment (e.g.: identifying a terrorist in a busy Airport terminal). 8. RELEVANCEFace recognition has been one of the most interesting and important research fields in the past two decades. The reasons come from the need of automatic recognitions and surveillance systems, the interest in human visual system on face recognition, and the design Student sign Guide sign of human-computer interface, etc. These researches involve know-ledge and researchers from disciplines such as neuroscience, psychology, computer vision, pattern recognition, image processing, and machine learning, etc. The initial stage in face recognition system is to detect the face from an input image, the main objective of this stage is to find all the faces that appear in the image irrespective of its pose, aging, expression, illumination and disguise. In face recognition with the purpose of localizing and extracting the face region from the background factors like pose, illumination, occlusion and the size of image makes difficult to detect or to recognize the face correctly. Yan, Kriegman and Ahuja presented a classifications that is well accepted [2].classified methods are categorized into four types; they are Knowledge-based methods, Feature invariant approaches, Template matching, Appearance-based methods, each of these methods have its own limitations. Methods Knowledge-Based Methods Drawbacks Feature-Invariant Methods Difficult to translate human knowledge into rules precisely. Difficult to extend this approach to detect faces in different poses. Difficult to locate facial features due to several corruptions (illumination, noise, and occlusion). Difficult to detect features in complex background. Template-Based Methods Templates need to be initialized near the face images. Difficult to enumerate templates for various types of poses. Appearance Based Methods Major problem with these methods is that they require a very long computation time in the training phase. Student sign Guide sign Face detection by skin colour thresholding enhances the input image, segments the skin regions in colour spaces RGB and YCbCr, and combines the edge image with the skin colour image to separate between the skin regions and the background. The advantage of this method is the detection of faces in different illumination conditions, different sizes, different poses, and different expressions. 9. LITERATURE REVIEWFace detection is the first step of face recognition system. Output of the detection can be location of face region as a whole, and location of face region with facial features (i.e. eyes, mouth, eyebrow, nose etc.). Detection methods in the literature are difficult to classify strictly, because most of the algorithms are combination of methods for detecting faces to increase the accuracy. Facial features are important information for human faces and standard images can be generated using these information. In literature, many detection algorithms based on facial features are available. Ramya Srinivasan et al. [1] detect faces and facial features by extraction of skin like region with YCbCr color space and edges are detected in the skin like region. Then, eyes are found with Principal Component Analysis (PCA) [3] on the edged region. Finally, Mouth is found based on geometrical information. Another approach extracts skin like region with Normalized RGB color space and face is verified by template matching. To find eyes, eyebrows and mouth, color snakes are applied to verified face image. Bouzas and Arvanitopoulos [4] faces are verified with Linear Support Vector Machine (SVM). For final verification of face, eyes and mouth are found with the information of Cb and Cr difference. For eye region Cb value is greater than Cr value and for mouth region Cr value is greater than Cb value. Another application segments skin like regions with statistical model. Dahmane and Meunier built up a prototype to model facial expression considering various approaches to facial characteristics [5]. Also image quality plays an important role in face recognition systems. Higher the clarity of the face in an image, more is the accuracy and vice versa. This parameter is well explained by Jiansheng Chen et.al [6]. Mostly the success factor of any face recognition system depends on the quality of image provided. Face recognition is considered one of the most important biometric methods displaying some advantages over other biometric approaches, for being natural and passive, not requiring the cooperation of individuals as in other techniques such as iris recognition. This characteristic of facial recognition makes it ideal for applications in the field of security, where it is necessary to perform recognition through security cameras Student sign Guide sign without the cooperation of the individual. Hence various tasks of face recognition have to be studied [7]. Wang and Ji suggested an method which will increase the performance of the system. There experimental results show that using the method of performance modeling the accuracy of the system can be improved[8]. The segmented parts are taken as candidate and verification is done by calculating the entropy of the candidate image and use thresholding to verify face candidate [9]. Qiang-rong and Hua-lan [10] applied white balance correction before detecting faces. The color value is important for segmentation, so while acquiring the image colors may reflect false color. To overcome this, white balance correction should be done as a first step. Also, skin color can be modeled in elliptical region in Cb (blue difference chroma component) and Cr (red difference chroma component) channel in YCbCr color space. Skin like region is segmented if the color value is inside elliptic region and candidate regions are verified using template matching [11]. Peer et al. [12] detect faces using only skin segmentation in YCbCr color space and researchers generate the skin color conditions in RGB color space as well. N. Sudha et.al [13] proposed an algorithm based on principal component analysis (PCA) which corresponds to eigenvectors of the data covariance matrix arranged in descending order of Eigen values. 10. PROPOSED WORK:A. Problem Statement:Lot of work has been performed in the field of face recognition. But this work is somewhat limited or mainly focused to single face recognition only. But in an image there can be more than one person. So there is a need of a system which will identify multiple faces in image. Also the current system has difficulties to locate facial features due to illumination, noise and occlusion. Hence we are trying to build a system which will overcome the difficulties mentioned and will identify an authorised person from the group of persons for door accessing. B. Scope:The proposed system mainly focuses on detecting and recognizing multiple persons from an image. Work has been performed in single face detection but our challenge lies in recognizing authorized person from a crowdy environment. After successfully identifying the Student sign Guide sign person, the further application is to provide access to the user into the house which is a generalized application. In case of an unauthorized person is detected an alarm is raised. C. Objective of Work:Our project mainly deals with face recognizing system. So the proposed objectives of this project are: i) To successfully detect multiple faces in an image. ii) To remove the difficulties in face detection and recognition due to illumination, noise and occlusion. iii) To provide access to the owner if the recognition is successfully authenticated. iv) To raise an alarm if an intruder tries to enter the house. D. MethodologyIn face recognition system the most important point to be considered, is the degradation of the image according to various filtering or conversion techniques which will help in recognizing the person with the images stored in database. A design flow for the degradation of image is shown in fig 4. The first step is Load/Capture image. In this step the image is captured and fed to the system. Then the next step is obtaining Blur image. In image terms blurring means that each pixel in the source image gets spread over and mixed into surrounding pixels. Steps to Blur image: – Traverse through entire input image array. – Read individual pixel colour value (24-bit). – Split the colour value into individual R, G and B 8-bit values. – Calculate the RGB average of surrounding pixels and assign this average value to it. – Repeat the above step for each pixel. – Store the new value at same location in output image. Then RGB Colour Model is formed. The RGB colour model approximates the way human vision encodes images by using three primary colour channels: red, green, and blue. The RGB colour model is additive, which means the red, green, and blue channels combine to create all the available colours in the system. Student sign Guide sign Then Skin Colour Thresholding is done. Thresholding is the simplest method of image segmentation. It is usually used for feature extraction where required features of image are converted to white and everything else to black (or vice-versa). Steps for Thresholding: Load/Store image Blur RGB to LAB/HSV Colour Model conversion Skin Colour Thresholding Blob Detection Face Localization Cropping Grayscale Image Edge Detection Registered Faces Feature Extraction Feature Comparison & Output Post Processing Database Fig 3. Design flow. Student sign Guide sign –Traverse through entire input image array. – Read individual pixel colour value (24-bit) and convert it into gray scale. – Calculate the binary output pixel value (black or white) based on current threshold. – Store the new value at same location in output image. Then in next step Blob Detection is done. In the field of computer vision, blob detection refers to mathematical methods that are aimed at detecting regions in a digital image that differ in properties, such as brightness or colour, compared to areas surrounding those regions. Then next step is Face Localization. Here the various faces within an image are localized or targeted for further identification purpose. Then with Cropping the various faces within an image that is been processed are cropped so as to avoid unnecessary part of the image. After Face Localization, Grayscale digital image is obtained. Grayscale digital image is an image in which the value of each pixel is a single sample, that is, it carries only intensity information. Now Edge detection is the next stage. It is a fundamental tool in image processing, machine vision and computer vision, particularly in the areas of feature detection and feature extraction, which aim at identifying points in a digital image at which the image brightness changes sharply or, more formally, has discontinuities. The result of applying an edge detector to an image may lead to a set of connected curves that indicate the boundaries of objects, the boundaries of surface markings as well as curves that correspond to discontinuities in surface orientation. The next step Feature extraction involves obtaining relevant facial features from the data. These features could be certain face regions, variations, angles or measures, which can be human relevant (e.g. eyes spacing) or not. This phase has other applications like facial feature tracking or emotion recognition. Then in Feature comparison and output stage the extracted features are matched with that saved in the database. If they are matched, recognition is successful else, fail. Then in Post processing if recognition is successfully the microcontroller opens the Door else alarm is turned ON as shown in fig. 4. Student sign Guide sign Image capturing camera Personal computer Alarm DC Motor Door Access Fig 4. System block diagram. 11. FACILITIES AVAILABLE AND REQUIREMENTS:Hardware & Software:1. Pc system with accessories. 2. MATLAB software with appropriate toolbox. 3. Camera. 4. Cable to transmit a signal. 12. PROJECT SCHEDULE:Month/Year Description Aug 2016-Nov 2016 Literature survey & submission of synopsis. Dec 2016- Jan 2017 To study of different face recognition and feature extraction algorithm. Feb- Mar2017 Implementation of face recognition and feature extraction algorithm. April-May 2017 Implementation of Embedded Hardware for access control. May-June 2017 Comparing simulation & result June-July 2017 Dissertation phase completion. 13. Expected Date of Completion:-30th July 2017. 14. Approximate Expenditures:-20,000 INR. Student sign Guide sign REFERNCES[1] Ramya Srinivasan, Abhishek Nagar, Anshuman Tewari, Donato Mitrani, Amit RoyChowdhury, 2014, “Face recognition based on sigma sets of image features”, Proc. 2014 IEEE International Conference on Acoustic, Speech and Signal Processing (ICASSP),, Samsung Research America, Dallas. [2] M.-H. Yang, D. Kriegman, and N. Ahuja. Detecting faces in images: A survey IEEE Transactions on Pattern Analysis and Machine Intel-ligence, 24(1):34–58, January 2002. [3] K. Seo, W. Kim, C. Oh and J. Lee, 2002, “Face Detection And Facial Feature Extraction Using Colour Snake”, Proc. ISIE 2002 - 2002 IEEE International Symposium on Industrial Electronics, pp.457-462, L 'Aquila, Italy. [4] Dimitrios Bouzas, Nikolaos Arvanitopoulos and Anastasios Tefas 2014, “Graph Embedded Nonparametric Mutual Information For Supervised Dimensionality Reduction”, Proc. 2014 IEEE transactions on neural networks and learning systems. [5] Mohamed Dahmane and Jean Meunier, 2014, “Prototype-Based Modeling for Facial Expression Analysis”,Proc. IEEE transactions on multimedia, VOL. 16, NO. 6, October 2014. [6] Jiansheng Chen, Yu Deng, Gaocheng Bai, and Guangda Su, 2015, “Face Image Quality Assessment Based on Learning to Rank”, Proc. IEEE signal processing letters, vol. 22, no. 1, January 2015. [7] Luis Fernando Martins Carlos Junior and Joao Luis Garcia Rosa, 2014, “Face Recognition through a Chaotic Neural Network Model”, Proc. International Joint Conference on Neural Networks (IJCNN), July 2014, Beijing, China. [8] Peng Wang and Qiang JiS. Kherchaoui and A. Houacine, 2006, “Performance Modeling and Prediction of Face Recognition Systems”, Proc. IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR’06). Student sign Guide sign D. Huang, T. Lin, C. Ho and W. Hu, 2010, “Face Detection Based On Feature [9] Analysis And Edge Detection Against Skin Colour-like Backgrounds”, Proc. 2010Fourth International Conference on Genetic and Evolutionary Computing, pp.687-690, Shenzen, China. [10] J. Qiang-rong and L. Hua-lan, 2010, “Robust Human Face Detection in Complicated Colour Images”, Proc. 2010 The 2nd IEEE International Conference on Information Management and Engineering (ICIME), pp.218 – 221, Chengdu, China. [11] C. Aiping, P. Lian, T. Yaobin and N. Ning, 2010, “Face Detection Technology Based On Skin Colour Segmentation And Template Matching”, Proc. 2010Second International Workshop on Education Technology and Computer Science, pp.708711, Wuhan, China. [12] P. Peer, J. Kovac and F. Solina, 2003, “Robust Human Face Detection in Complicated Colour Images”, Proc. 2010 The 2nd IEEE International Conference on Information Management and Engineering (ICIME), pp. 218 – 221, Chengdu, China. [13] N. Sudha, A. R. Mohan and Pramod K. Mehe, 2011, “A Self-Configurable Systolic Architecture for Face Recognition System Based on Principal Component Neural Network”, Proc. IEEE transactions on circuits and systems for video technology, vol. 21, no. 8, August 2011. Mr. Mulla Vasim Bashir Student Head Electronics Engineering Dept. Karmaveer Bhaurao Patil Dr. Patil Vikram S. Project Guide Principal Karmaveer Bhaurao Patil College of Engineering, Satara. College of Engineering, Satara. Student sign Guide sign