Model-based gradient field descriptor as a convenient tool for image recognition and analysis
Автор: Myasnikov Vladislav Valerievich
Журнал: Компьютерная оптика @computer-optics
Рубрика: Обработка изображений: Восстановление изображений, выявление признаков, распознавание образов
Статья в выпуске: 4 т.36, 2012 года.
Бесплатный доступ
In the paper we propose a new descriptor which is used to describe a digital image, specifically, a model-oriented descriptor of gradient field. Derived characteristics of the descriptor, that are considered as features of a digital image, allow to solve effectively the problems of image analysis, recognition and retrieval. The examples of such tasks solutions with the proposed descriptor are given.
Digital images, descriptors, features, analysis, recognition, image retrieval
Короткий адрес: https://sciup.org/14059126
IDR: 14059126