Jointly Learning the Discriminative Dictionary and Projection for Face Recognition

Recently, dictionary learning has become an active topic. However, the majority of dictionary learning methods directly employs original or predefined handcrafted features to describe the data, which ignores the intrinsic relationship between the dictionary and features. In this study, we present a method called jointly learning the discriminative dictionary and projection (JLDDP) that can simultaneously learn the discriminative dictionary and projection for both image-based and video-based face recognition. The dictionary can realize a tight correspondence between atoms and class labels. Simultaneously, the projection matrix can extract discriminative information from the original samples. Through adopting the Fisher discrimination criterion, the proposed framework enables a better fit between the learned dictionary and projection. With the representation error and coding coefficients, the classification scheme further improves the discriminative ability of our method. An iterative optimization algorithm is proposed, and the convergence is proved mathematically. Extensive experimental results on seven image-based and video-based face databases demonstrate the validity of JLDDP.

Video based Face Recognition with limited Resources

Face recognition(FR) has multi-domain applications and video based FR is good area of research in terms of accuracy and performance. Recent research has proved that Convolutional Neural Network (CNN) is a one of best solution for object detection and recognition as it extracts features from on its own and performs classification as well. As we go for higher accuracy models, size of network increases and it requires more time to process a frame or video as it involves more computations. This paper aims at building a FR model which is smaller in network size, requires limited resources while building and still achieves good accuracy. The system uses combination of deep learning and machine learning based solution. FR system is built with CNN-Suport Vector Machine(SVM) model where CNN performs feature extraction and SVM performs classification task. Results shows that CNN-SVM model gives higher accuracy (94.05% validation and 90.17% testing) compared to conventional CNN-softmax model (93.37% validation and 88.77% testing) with a small network size and also requires less training time. Results can be improved by using cross validation techniques.