Design and Evaluation of a Real-Time Face Recognition System using Convolutional Neural Networks

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.

An efficient face recognition approach using PCA and minimum distance classifier

We compare four commonly used eigenvector methods, namely cyclic Jacobi's method of iteration, Wiedlandt's de∞ation, Hotelling's de∞ation, and MATLAB's own eigen- value method for the success of face recognition which is based on Principal Component Analysis (PCA). We report that the highest recognition rate is equally achieved by MATLAB's eigenvalue method and Hotelling's de∞ation. The former is observed to be the fastest for large numbers of dominant eigenfaces while scaling the best with the number of computational cores. On the other hand, the latter has a brief and open source code that can be easily modifled for a given purpose. We further investigate the impact of altering face images to improve the recognition rate. Difierent sets of images have been obtained from two well-known face databases, various efiects using imaging fllters have been applied to them, and the resulting sets have been used as both training and test sets. Recognition rates reveal that some of these flltered sets can be even better candidates for training and testing than the original sets.

A Comparison of Eigenvalue Methods for Principal Component Analysis

Face is the most easily identifiable characteristic of a person. Variations in facial expressions can be easily recognized by humans, while it is quite difficult for machines to recognize faces portraying varying facial expressions, pose, and illumination conditions efficiently. Face recognition works as a combination of feature extraction and classification. The selection of a combination of feature extraction technique and classifier to obtain maximum accuracy rate is a challenging task. This paper presents a unique combination of feature extraction technique and classifier that yields a satisfactory and more or less same accuracy rate when tested on more than one standard database. In this combination, features are extracted using principle coponent analysis (PCA). These extracted features are then fed to a minimum distance classification system. The proposed combination is tested on ORL and YALE datasets with an accuracy rate of 95.63% and 93.33%, respectively, considering variations in facial expressions, poses as well as illumination conditions.

Face Recognition Using PCA and Minimum Distance Classifier

Facial feature detection and recognition are widely used in current world scenarios and technologies. Almost every smartphone is embedding facial recognition & detection in security matters like phone unlocking. Another increasing use of this type of technology is being noticed in cameras and social media like snap chat. In this paper, we developed a system/platform in which facial features containing noise, eyes, mouth, and eyebrows can be detected from an image. The developed system can also recognize a human subject from two different input images. The proposed system works in two steps-in the first step, image segmentation has been employed on more than a few images to enhance the images by using various edge detection techniques and filtering out the noise. In the second part, features are extracted from an idea through histogram of gradient (HOG). The entire facial feature detection and recognition system was developed with a histogram of oriented gradients (HOG), support vector machine (SVM), and K nearest neighbor (KNN). The proposed system gives an accuracy of 96% in the detection and recognition of facial features and human subjects. The projected strategy of procedures offers improved truth once appeared otherwise in relevance the others mixture of ways in which. The assessment has been done in the Python platform on authentic images of the Indian face database.

Face Detection and Recognition System

A new approach is adopted in this paper based on Turk and Pentland’s eigenface method. It was found that the probability density function of the distance between the projection vector of the input face image and the average projection vector of the subject in the face database, follows Rayleigh distribution. In order to decrease the false acceptance rate and increase the recognition rate, the input face image has been recognized using two thresholds including the acceptance threshold and the rejection threshold. We also find out that the value of two thresholds will be close to each other as number of trials increases. During the training, in order to reduce the number of trials, the projection vectors for each subject has been averaged. The recognition experiments using the proposed algorithm show that the recognition rate achieves to 92.875% whilst the average number of judgment is only 2.56 times. Keywords—Eigenface, Face Recognition, Threshold, Rayleigh Distribution, Feature Extraction

ShunBao Li

Papers

The Modified Eigenface Method using Two Thresholds