TranSalNet: towards perceptually relevant visual saliency prediction.

Lou, Jianxun; Lin, Hanhe; Marshall, David; Saupe, Dietmar; Liu, Hantao

Abstract

Convolutional neural networks (CNNs) have significantly advanced computational modelling for saliency prediction. However, accurately simulating the mechanisms of visual attention in the human cortex remains an academic challenge. It is critical to integrate properties of human vision into the design of CNN architectures, leading to perceptually more relevant saliency prediction. Due to the inherent inductive biases of CNN architectures, there is a lack of sufficient long-range contextual encoding capacity. This hinders CNN-based saliency models from capturing properties that emulate viewing behaviour of humans. Transformers have shown great potential in encoding long-range information by leveraging the self-attention mechanism. In this paper, we propose a novel saliency model that integrates transformer components to CNNs to capture the long-range contextual visual information. Experimental results show that the transformers provide added value to saliency prediction, enhancing its perceptual relevance in the performance. Our proposed saliency model using transformers has achieved superior results on public benchmarks and competitions for saliency prediction models. The source code of our proposed saliency model TranSalNet is available at: https://github.com/LJOVO/TranSalNet.

Citation

LOU, J., LIN, H., MARSHALL, D., SAUPE, D. and LIU, H. 2022. TranSalNet: towards perceptually relevant visual saliency prediction. Neurocomputing [online], 494, pages 455-467. Available from: https://doi.org/10.1016/j.neucom.2022.04.080

• Journal Article Type: Article
• Acceptance Date: Apr 17, 2022
• Online Publication Date: Apr 21, 2022
• Publication Date: Jul 14, 2022
• Deposit Date: Apr 22, 2022
• Publicly Available Date: Jun 8, 2022
• Journal: Neurocomputing
• Print ISSN: 0925-2312
• Electronic ISSN: 1872-8286
• Publisher: Elsevier
• Peer Reviewed: Peer Reviewed
• Volume: 494
• Pages: 455-467
• DOI: https://doi.org/10.1016/j.neucom.2022.04.080
• Keywords: Saliency prediction; Deep learning; Transformer; Convolutional neural network
• Public URL: https://rgu-repository.worktribe.com/output/1646534
• Additional Information: The source code of the proposed saliency model TranSalNet is available at: https://github.com/LJOVO/TranSalNet.