Skip to main navigation Skip to search Skip to main content

Attribute-Guided Zero-Shot CLIP in Image Classification

  • Guoxi Qiu
  • , Xiangyu Zhang
  • , Yong Xu
  • , Jinghua Wang*
  • *Corresponding author for this work
  • Harbin Institute of Technology

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Contrastive Language-Image Pre-training (CLIP) has shown strong performance in zero-shot image classification. However, it requires large datasets and high computational costs. In this paper, we propose a transferable module that improves CLIP's zero-shot classification accuracy by integrating image-specific attributes into prompts during inference. This method enhances CLIP's performance without additional training and extends its applicability to other vision-language models (VLMs), offering an efficient solution for image classification tasks.

Original languageEnglish
Title of host publication2025 IEEE International Conference on Multimedia and Expo
Subtitle of host publicationJourney to the Center of Machine Imagination, ICME 2025 - Conference Proceedings
PublisherIEEE Computer Society
ISBN (Electronic)9798331594954
DOIs
StatePublished - 2025
Externally publishedYes
Event2025 IEEE International Conference on Multimedia and Expo, ICME 2025 - Nantes, France
Duration: 30 Jun 20254 Jul 2025

Publication series

NameProceedings - IEEE International Conference on Multimedia and Expo
ISSN (Print)1945-7871
ISSN (Electronic)1945-788X

Conference

Conference2025 IEEE International Conference on Multimedia and Expo, ICME 2025
Country/TerritoryFrance
CityNantes
Period30/06/254/07/25

Keywords

  • Attribute
  • Contrastive Learning
  • image classification

Fingerprint

Dive into the research topics of 'Attribute-Guided Zero-Shot CLIP in Image Classification'. Together they form a unique fingerprint.

Cite this