| Abstrakt: |
In this chapter, we mainly focus on the discussion of unsupervised face recognition, which corresponds to the matching procedure and the classification-related sub-steps of the enrollment procedure, including same face decision, feature extraction and encoding. If each single step is running passively, the autonomous procedure of recognition is guaranteed. After the introduction section, feature extraction and encoding methods are briefly covered in section 2. The matching algorithms are described in detail in section 3, which includes an image-based classification method, an adaptive way of selecting the thresholds for the classifier, and a video-based temporal filtering method. In section 4, the combined same face decision algorithms are discussed. Finally is the summary of the automatic and intelligent face recognition procedure. [ABSTRACT FROM AUTHOR] |
