A Study Of Face Tracking, Identity Verification And Age Estimation

Face expresses the important information related to human characteristics. Due to the important application of face recognition in entrance verification, human-computer interaction and so on, it has become a hot research topic recently. The main research include face detection,face tracking, face recognition, age estimation, gender classification. The main contributions of this paper include the following three aspects:1) Comfirming the location of face target is the basis of face research.Robust face tracking algorithm can contribute to detect face to get better results.This paper combine intensity information(cross-bin distribute field, CDF), texture information(enhance histograms of oriented gradients, EHOG) and color information(color name, CN) in a tracking-by-detection framework, in which a simple tracker called CSK is extended for multi-dimension and multi-cue fusion. The proposed approach achieving 74.35% is better than most recent state-of-the-art tracking algorithms.2) The paper use CNN to recognize face.This paper presents a improved CNN-PCA hybrid model,the extracted features by CNN are concatenated to form a high-dimensional feature vector, whose dimension is compressed by PCA,98.4% verification rate is achieved on the LFW database.Besides,another improved hybrid model CNN-SVM, CNN works as a trainable feature extractor, SVM performs as a recognizer,get 98.18% verification rate.At last, we also improved the CNN by using PRe LU active function replace MFM,discarding the pooling layer, using both face identification and verification signals as supervision,get 98.67%verification rate.3) The paper tackle the estimation of apparent age in still face images. Our CNN use the VGG-16 architecture. We pose the age regression problem as adeep classification problem followed by regression problem, the prediction rate outperforming the human reference.