Paper Image

Learning image-text relations

Published on:

9 November 2023

Primary Category:

Computer Vision and Pattern Recognition

Paper Authors:

Yongfeng Chena,

Jin Liua,

Zhijing Yang,

Ruihan Chena,

Junpeng Tan

Bullets

Key Details

Proposes active learning method to handle challenging negative pairs

Introduces new loss function based on a caption evaluation metric

Improves model's ability to distinguish positive/negative pairs

Increases generalization ability on large datasets

Outperforms previous methods on two benchmarks

AI generated summary

Learning image-text relations

This paper proposes a new model to improve image-text matching, which is the task of retrieving corresponding images or text captions. The model introduces active learning techniques and a new loss function to better handle challenging samples and improve generalization.

Answers from this paper

Comments

No comments yet, be the first to start the conversation...

Sign up to comment on this paper

Sign Up