Golden snub-nosed monkey face detection and tracking algorithm based on S-TLD

• Main Authors: Zheng Xin, Xu Pengfei, He Gang, Ye Guixin, Guo Songtao, Li Baoguo, Chen Xiaojiang, Fang Dingyi, Chang Hongli
• Format: Patent
• Language: Chinese; English
• Published: 03.08.2016
• Subjects: CALCULATING; COMPUTING; COUNTING; HANDLING RECORD CARRIERS; PHYSICS; PRESENTATION OF DATA; RECOGNITION OF DATA; RECORD CARRIERS

The invention discloses a golden snub-nosed monkey face detection and tracking algorithm based on S-TLD, and belongs to the field of image processing. The golden snub-nosed monkey face detection and tracking algorithm based on the S-TLD comprises the steps that skin color and hair color images of a golden snub-nosed monkey are acquired, and a facial feature library of the golden snub-nosed monkey is determined; the features in the facial feature library are optimized by using an SVM classifier so that an optimized SVM network model is obtained; and a video image to be identified is acquired, the golden snub-nosed monkey is located in the video image through combination of the SVM model, and the located golden snub-nosed monkey is tracked based on the TLD theory. The tracking algorithm is enabled to have high tracking precision under the condition that the background is complex, the movement amplitude of the target is large or the target is back to the video by using a standby classifier SVM. Meanwhile, the al