Face detection based on open Cl design and image processing technology

摘要

Face detection is a system that automatically recognizes facial expressions. Extract our structure and describe its function of the contours of the eyebrows, eyes and mouth with elastic rectangles. This manual facial features are an improvement over the hybrid mask model method. For some have to reduce the recognition time and define the accuracy of the relevant cognition. Then, these vectors determine the network, which is being used to implement user facial expression and neural interactions. In short, FEID (Facial Expression Image Database) information is used in this study to determine face and face identification. 96.2% of test results and 92.8% of the FEID database show that personal facial recognition tests and official member face recognition can be approved. Face recognition 97.4% of FEID samples were widely recognized. Real-time face detection is a new method using a variety of computations. This algorithm uses the Local Binary Format (LBF) as a face detection feature vector. Code used using OpenCL units. Illumination is the use of irreversible gamma correction or Gaussian variations of different illumination conditions creating strong protocols. This implementation has proven to be faster compared to previous co-executions.