A Survey of Recent Advances in CNN-Based Fine-Grained Visual Categorization

Abstract

Fine-grained visual classification (FGVC) is an important task in the field of computer vision (CV), which aims to classify sub-categories that are hard to distinguish (e.g., identifying different species of birds). It has become a challenging problem due to the larger inter-class similarity and intra-class variability in fine-grained datasets. In recent years, with the booming of deep learning, convolutional neural networks (CNN) have brought a new opportunity to FGVC, and a large number of CNN-based algorithms have been proposed, which have advanced the rapid development of FGVC. In this paper, we introduce the characteristics and challenges of FGVC tasks first, then analyze the latest methods from the strongly-supervised to and weakly-supervised ones, and compare the performances of these algorithms on commonly used FGVC datasets. Finally, we categorize the FGVC related applications.