Fine-Grained Image Categorization with Fisher Vector

Abstract

Fine-grained image categorization is a categorization task, where classifying objects should be the same basic-level class and have similar shape or visual appearances. Generally, the bag-of-words (BoW) model is popular in image categorization. However, in BoW model, the feature quantization for image representation is also a lossy process, which severely limits the descriptive power of the image representation. Fisher vectors employ soft assignments and reduce information loss due to quantization by calculating the gradient for each parameter separately, which have been shown to outperform other global representations on most benchmark datasets. In this paper, the acquired template is represented by Fisher Vector (FV). Combing FV with improved spatial pyramid matching (SPM) respectively, we use an approach, i.e., FV+SPM, to obtain feature representation. Experimental results show that our method outperforms state-of-the-art categorization approaches on the Caltech-UCSD Birds dataset.