Research Progress of Fine-grained Visual Classification: Basic Framework, Challenges, and Future Development

Congying Ma, Yuhan Pu

Research output: Chapter or section in a book/report/conference proceedingChapter in a published conference proceeding

Abstract

With the progress of multimedia technology, fine-grained visual classification (FGVC) has gradually attracted extensive attention in the field of computer vision. Thanks to the success of deep neural networks, both the accuracy and speed have achieved an unprecedented breakthrough. Hundreds of fine-grained classification techniques have been developed from the early, fully supervised methods to the current weakly ones. To this end, in this article, we provide a comprehensive survey of the recent achievements of FGVC. Specifically, we describe the formulation and setting of the FGVC, including the background, basic framework, and challenges. Then, we introduce the widely used datasets and compare the performance of different models. Lastly, we discuss the potential future directions of FGVC.

Original languageEnglish
Title of host publication2021 3rd International Academic Exchange Conference on Science and Technology Innovation (IAECST)
Place of PublicationU. S. A.
PublisherIEEE
Pages413-419
Number of pages7
Volume2021
ISBN (Electronic)9781665402675
DOIs
Publication statusPublished - 7 Feb 2022
Externally publishedYes
Event3rd International Academic Exchange Conference on Science and Technology Innovation, IAECST 2021 - Guangzhou, China
Duration: 10 Dec 202112 Dec 2021

Publication series

Name2021 3rd International Academic Exchange Conference on Science and Technology Innovation, IAECST 2021

Conference

Conference3rd International Academic Exchange Conference on Science and Technology Innovation, IAECST 2021
Country/TerritoryChina
CityGuangzhou
Period10/12/2112/12/21

Keywords

  • deep learning
  • Fine-grained visual classification
  • weakly supervised learning
  • zero-shot learning

ASJC Scopus subject areas

  • Chemical Engineering (miscellaneous)
  • Computer Science Applications
  • Signal Processing
  • Information Systems and Management
  • Engineering (miscellaneous)
  • Ceramics and Composites
  • Electronic, Optical and Magnetic Materials
  • Metals and Alloys

Fingerprint

Dive into the research topics of 'Research Progress of Fine-grained Visual Classification: Basic Framework, Challenges, and Future Development'. Together they form a unique fingerprint.

Cite this