Local Feature Detectors, Descriptors, and Image Representations: A Survey
This is an incremental survey paper that provides a comprehensive overview for researchers in computer vision and image retrieval.
The paper surveys local feature detectors, descriptors, and image representations for image retrieval, grouping and summarizing numerous existing methods and briefly reviewing recent deep learning-based approaches.
With the advances in both stable interest region detectors and robust and distinctive descriptors, local feature-based image or object retrieval has become a popular research topic. %All of the local feature-based image retrieval system involves two important processes: local feature extraction and image representation. The other key technology for image retrieval systems is image representation such as the bag-of-visual words (BoVW), Fisher vector, or Vector of Locally Aggregated Descriptors (VLAD) framework. In this paper, we review local features and image representations for image retrieval. Because many and many methods are proposed in this area, these methods are grouped into several classes and summarized. In addition, recent deep learning-based approaches for image retrieval are briefly reviewed.