Domain-Incremental Learning for Saliency Prediction

What is it about?

Saliency prediction (SAP) plays a crucial role in simulating the visual perception function of human beings. In practical situations, humans can quickly grasp saliency extraction in new image domains. However, current SAP methods mainly concentrate on training models in single domains, which do not effectively handle diverse content and styles present in real-world images. As a result, it would be of great significance if SAP models could efficiently adjust to new image domains. To this end, this paper aims to design SAP models that can imitate the incremental learning ability of human beings on multiple image domains, and name domain-incremental saliency prediction (DISAP). To make a trade-off between preventing the forgetting of historical domains and achieving high performance on new domains, we propose a progressively updated domain incremental encoder. This encoder consists of a domain-sharing branch and a domain-specific branch. The domain-sharing branch includes a feature selection mechanism to preserve crucial parameters after fine-tuning the model on each current domain. The remaining parameters are reserved to absorb knowledge from future domains. Furthermore, to capture the unique characteristics of each domain with relatively low computational overhead, we introduce a lightweight design to construct the domain-specific branch, enabling effective adaptation to new domains.

Featured Image

Photo by Steve Johnson on Unsplash

Photo by Steve Johnson on Unsplash

Why is it important?

Saliency prediction (SAP) plays a crucial role in simulating the visual perception function of human beings. In practical situations, humans can quickly grasp saliency extraction in new image domains. As a result, it would be of great significance if SAP models could efficiently adjust to new image domains.

Perspectives