Content based facial image search

В работе рассмотрена проблема распознавания лиц на изображениях с низким разрешением. Предложен новый подход для решения задач детекции и идентификации лиц. Представлены результаты тестирования алгоритмов детекции и идентификации лиц на распространенных базах изображений.The task of searching facial images consists of two problems: face detection and face recognition. This paper describes face detection and recognition methods based on special image representation - matrix of brightness variation formed from derivative signs of image intensity function. The partial invariance of algorithms to illumination conditions have shown. All proposed algorithms were tested on the proper facial image bases. Experimental results confirm efficiency of proposed methods