Although progress in face recognition is encouraging, the accuracy rate of face recognition remains to be increased. Since the face image quality has a positive influence on face recognition accuracy, the image enhanc...Although progress in face recognition is encouraging, the accuracy rate of face recognition remains to be increased. Since the face image quality has a positive influence on face recognition accuracy, the image enhancement methods are popular in face recognition. Most current image enhancement methods aim at improving visual appearance, but cannot improve recognition accuracy remarkably. In this paper, a feature evaluation operator is designed to overcome this problem. The operator selects patches with the best quality, and then face image is reconstructed with the selected patches. The proposed algorithm is tested on two different face recognition applications. Accuracy is raised after enhancement, and the result proves that the proposed algorithm is effective.展开更多
基金the National Natural Science Foundation of China(Nos.61876112,61303104,61601311,61603022,61373090 and 61203238)the Natural Science Foundation of Beijing(Nos.4162017 and 4132014)+4 种基金the Support Project of High-Level Teachers in Beijing Municipal Universities in the Period of 13th Five-Year Plan(No.CIT&TCD20170322)the Project of Beijing Excellent Talents(No.2016000020124G088)the Beijing Municipal Education Research Plan Project(No.SQKM201810028018)the Capacity Building for Sci-Tech Innovation-Fundamental Scientific Research Funds(No.025185305000/134/187/188/189)the Youth Innovative Research Team of Capital Normal University
文摘Although progress in face recognition is encouraging, the accuracy rate of face recognition remains to be increased. Since the face image quality has a positive influence on face recognition accuracy, the image enhancement methods are popular in face recognition. Most current image enhancement methods aim at improving visual appearance, but cannot improve recognition accuracy remarkably. In this paper, a feature evaluation operator is designed to overcome this problem. The operator selects patches with the best quality, and then face image is reconstructed with the selected patches. The proposed algorithm is tested on two different face recognition applications. Accuracy is raised after enhancement, and the result proves that the proposed algorithm is effective.
