Application of constrained local neural fields in face recognition

The facial feature points localization is the core of face recognition, and its accuracy directly affects the accuracy of face recognition system. The accuracy of facial feature points is affected by light, noise, background, and face gestures. Considering the theoretical value and practical significance of facial feature points localization, this thesis goes into the most advanced algorithm of facial feature points localization-Constrained Local Neural Fields. Taking into account the reliability of each patch model (feature point detector), CLNF combines the Local Neural Field patch model. The analysis contrasts the advantages of Constrained Local Models.