🎉 One Paper Has Been Accepted by AAAI 2024

AAAI 2024 officially released the list of accepted papers. We are thrilled to announce that Dr. Zhong's paper from our team has been included!

Research Background

Saccadic scanpath, a data representation of human visual behavior, has received broad interest in multiple domains. Scanpath is a complex eye-tracking data modality that includes the sequences of fixation positions and fixation duration, coupled with image information.

However, previous methods usually face the spatial misalignment problem of fixation features and loss of critical temporal data (including temporal correlation and fixation duration).

Key Contributions

In this study, we propose a Transformer-based scanpath model, SpFormer, to alleviate these problems. The main contributions include:

1. Fixation-Centric Paradigm

We propose a fixation-centric paradigm to extract the aligned spatial fixation features and tokenize the scanpaths, effectively addressing the spatial misalignment problem.

2. Local Meta Attention

According to the visual working memory mechanism, we design a local meta attention to reduce the semantic redundancy of fixations and guide the model to focus on the meta scanpath.

3. Duration Information Integration

We progressively integrate the duration information and fuse it with the fixation features to solve the problem of ambiguous location with the Transformer block increasing.

Experimental Results

We conduct extensive experiments on four databases under three tasks. The SpFormer establishes new state-of-the-art results in distinct settings, verifying its flexibility and versatility in practical applications.

The experimental results demonstrate that our approach significantly outperforms existing methods across multiple evaluation metrics, showcasing the effectiveness of the proposed spatio-temporal modeling framework.

Conclusion

This acceptance at AAAI 2024 represents a significant milestone for our research team. We believe that SpFormer will contribute to advancing the understanding of human visual behavior and inspire future research in eye-tracking analysis and related fields.

Congratulations to Dr. Wenqi Zhong and all co-authors for this outstanding achievement! 🎊