TzuChanChuang/Dictionary-Learning-for-Hierarchical-Structured-Image-Categories
In this paper, we proposed a novel sparse coding based image classification approach in a hierarchical structure. Hierarchical Structured Dictionary Learning (HSDL) aims to exploit the visual correlation between object categories which are visually similar by learning multiple class-specific dictio- naries and more than one corresponded shared dictionary. A Fisher discrimination criterion based discriminative term is adopted for both the level of class-specific dictionaries and the level of shared dictionaries level to enhance the discrimination of dictionaries. The performance of HSDL has been evaluated on benchmark image databases.
正在读取项目信息…