FACE RECOGNITION : A COMPARATIVE STUDY OF PRINCIPAL COMPONENT ANALYSIS AND DIFFERENCE COMPONENT ANALYSIS (original) (raw)
Related papers
Analyzing Face Recognition Using Pca and Comparison between Different Distance Classifier.
2013
Principal component analysis (PCA) is a widely used technique that is quite efficient and reliable when used for face recognition. Face is the most dominant feature that strongly communicates identity of the person. It proves to be very useful and secure if used for biometric identification. Principal component analysis uses Eigen Face approach for extracting effective features (Eigen vectors) from a given face database and these features corresponds to the dissimilarities among the faces. Every face in the database can be represented as a linear combination of these eigenvectors with appropriate weight assoc presents a methodology for face
Face recognition system based on principal components analysis and distance measures
International Journal of Engineering & Technology, 2018
Face recognition plays a vital role and has a huge scope in the field of biometrics, image processing, artificial intelligence, pattern recognition and computer vision. This paper presents an approach to perform face recognition using Principal Components Analysis (PCA) as feature extraction technique and different distance measures as matching techniques. The proposed method is developed after the deep study of a number of face recognition methods and their outcomes. In the proposed method, Principal Components Analysis is used for facial features extraction and data representation. It generates eigenvalues of the facial images, hence, reduces the dimensionality. The recognition is produced using three different matching techniques (Euclidean, Manhattan and Mahalanobis) and the results are` presented. Yale and Aberdeen Face Databases are used to test and analyze the results of the proposed method.
Principal Component Analysis for Face Recognition
The main objective of this paper is to implement face recognition system using Principle Component Analysis (PCA), where a model is trained for each user. This target can be mainly decomposed into image preprocessing, feature extraction and feature match. The Euclidean distance and Chessboard distance classifiers for recognition are used. Finally, the comparisons between two classifiers are reported.
Biometric Face Recognition using Principal Component Analysis
Biometric Face Recognition, a technology being used worldwide is of immense importance as it helps in ensuring public safety, preventing crimes and also improves customer experience in several fields. It is a software application which is capable of identification or distinct verification of an individual by analyzing and then comparing the pattern projected by the person's facial contours. This paper discusses the construction of a face recognition system using Principal component analysis (PCA) which is a statistical approach used to reduce the number of variables and complications which could occur during the process of face recognition while extracting the foremost relevant information from the image of the face being recognized. A training database of around 200-300 images belonging to different individuals is being used from an open source named face94 for carrying out the face recognition by comparing these images to the face to be recognized. This recognition is done by projecting a new test image onto the subspace which is then classified by measuring the minimum Manhattan distance.
COMPARATIVE ANALYSIS OF FACE RECOGNITION BASED ON SIGNIFICANT PRINCIPAL COMPONENTS OF PCA TECHNIQUE
IAEME, 2019
Face recognition systems have been emerging as acceptable approaches for human authorization. Face recognition help in searching and classifying a face database and at a higher level help in identification of possible threats to security. In face recognition problem, the objective is to search a face in the reference face database that matches a given subject. The task of face recognition involves the extraction of feature vectors of the human face from the face image for differentiating it from other persons [6]. In this work, the comparative analysis is done based on the varying number of highly significant principal components (Eigenvectors) of PCA for face recognition. Experimental results show a small number of principal components of PCA are required for matching. PCA technique is a statistical technique, it reduces the dimension of the search space that best describes the images.
An efficient face recognition approach using PCA and minimum distance classifier
2011 International Conference on Image Information Processing, 2011
Facial expressions convey non-verbal cues, which play an important role in interpersonal relations. Automatic recognition of human face based on facial expression can be an important component of natural human-machine interface. It may also be used in behavioral science. Although human being can recognize the face practically without any effort, but reliable face recognition by machine is a challenge. This paper presents a new approach for recognizing the face of a person considering the expression of the same human face at different instances of time. This methodology is developed by combining principle component analysis (PCA) for feature extraction and minimum distance classifier (MDC) for classification. Experiment is done on AT&T dataset and the recognition rate achieves to 96.7% for different facial expressions.
Face Recognition using PCA and LDA Comparative Study
Biometrics is a system in which we used to recognise human on the basis of its physical or behavioural characteristics. Today all over the world every country wants security of data, physical access, etc. Face recognition is widely accepted technique in human being same things are used in computer vision by using the image processing. In this paper we have used Principle Component Analysis (PCA) and Linear Discriminant Analysis (LDA) for extraction of features. Techniques has been applied for identification of a person on various databases such as ORL, Indian & KVKR and distance is calculated by using Euclidian Distance between training images and testing images. In this experiment total 1266 images used apart from that 25 subject from KVKR Face database this database is developed under UGC-SAP Phase I (our own major contribution) having 10 pose of each subject, 40 subjects from ORL having 10 images each, and from IIT Indian database 56 subjects, 11 images per subjects. After apply PCA on ORL Database the highest RR 97.50% for experiment 1:8 images, on KVKR-Face Database we got 92.00% RR for 1:9 and on IIT-Indian database 62.50%RR for 1:9. After applied LDA on ORL database we got the result 80.00% is highest RR, also applied on KVKR-Face database we got excellent result 100% RR and on IIT-Indian database 64.29% RR.
Biometric recognition techniques have emerged as most promising option and secured means of authentication capable of sustaining the emerging ubiquitous computing. Facial recognition stands a better chance for public authentication since the trait can be unconsciously captured. Lot of algorithms had been proposed for facial recognition system; with a lot of contradictory report found in literatures on comparing those algorithms. The aim of this paper is to present independent, comparative study of two most popular appearance-based face recognition methods – Principal Component Analysis and Independent Component Analysis. This paper was motivated by insufficient information and detailed of independent comparisons of all possible algorithms implemented in available literature. This work presents comparative study for performance evaluation of face recognition system based on recognition rate, recognition time and error rate. The training and testing samples for both Algorithms were taken from AT & T (ORL) face Database. For training sample, the researcher selected 100, 130 and 175 face images for Training set I, Training set II and Training set III that belong to10, 13 and 35 persons respectively. The results showed that Indepenedent Component Analysis (ICA) perform better in term of recognition rate and error rate. But the Principal Component Analysis (PCA) perform excellently when we consider time complexity of the recognition algorithm, therefore it can be used for real time recognition system.