Automatic Image Co-Segmentation: A Survey

TL;DR

This survey comprehensively reviews automatic image co-segmentation methods, analyzing visual cues, traditional and deep learning approaches, datasets, evaluation metrics, and future challenges to advance the field.

Contribution

It provides a detailed overview of existing co-segmentation algorithms, categorizes methods, compares performances, and discusses future research directions.

Findings

Deep learning methods outperform traditional approaches.

Standard datasets enable consistent evaluation.

Challenges include handling complex scenes and diverse objects.

Abstract

Image co-segmentation is important for its advantage of alleviating the ill-pose nature of image segmentation through exploring the correlation between related images. Many automatic image co-segmentation algorithms have been developed in the last decade, which are investigated comprehensively in this paper. We firstly analyze visual/semantic cues for guiding image co-segmentation, including object cues and correlation cues. Then we describe the traditional methods in three categories of object elements based, object regions/contours based, common object model based. In the next part, deep learning based methods are reviewed. Furthermore, widely used test datasets and evaluation criteria are introduced and the reported performances of the surveyed algorithms are compared with each other. Finally, we discuss the current challenges and possible future directions and conclude the paper. Hopefully, this comprehensive investigation will be helpful for the development of image co-segmentation technique.