{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/vitgaze-gaze-following-with-interaction","title":"ViTGaze: Gaze Following with Interaction Features in Vision Transformers","arxiv_id":"2403.12778","date":"2024-03-19","proceeding":null,"authors":["Yuehao Song","Xinggang Wang","Jingfeng Yao","Wenyu Liu","Jinglin Zhang","Xiangmin Xu"],"abstract":"Gaze following aims to interpret human-scene interactions by predicting the person's focal point of gaze. Prevailing approaches often adopt a two-stage framework, whereby multi-modality information is extracted in the initial stage for gaze target prediction. Consequently, the efficacy of these methods highly depends on the precision of the preceding modality extraction. Others use a single-modality approach with complex decoders, increasing network computational load. Inspired by the remarkable success of pre-trained plain vision transformers (ViTs), we introduce a novel single-modality gaze following framework called ViTGaze. In contrast to previous methods, it creates a novel gaze following framework based mainly on powerful encoders (relative decoder parameters less than 1%). Our principal insight is that the inter-token interactions within self-attention can be transferred to interactions between humans and scenes. Leveraging this presumption, we formulate a framework consisting of a 4D interaction encoder and a 2D spatial guidance module to extract human-scene interaction information from self-attention maps. Furthermore, our investigation reveals that ViT with self-supervised pre-training has an enhanced ability to extract correlation information. Many experiments have been conducted to demonstrate the performance of the proposed method. Our method achieves state-of-the-art (SOTA) performance among all single-modality methods (3.4% improvement in the area under curve (AUC) score, 5.1% improvement in the average precision (AP)) and very comparable performance against multi-modality methods with 59% number of parameters less.","url_abs":"https://arxiv.org/abs/2403.12778v2","url_pdf":"https://arxiv.org/pdf/2403.12778v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"vitgaze-gaze-following-with-interaction","repo_url":"https://github.com/hustvl/vitgaze","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"gaze-target-estimation","task_name":"Gaze Target Estimation"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/gaze-target-estimation-on-gazefollow","task":"Gaze Target Estimation","dataset":"GazeFollow","model":"ViTGaze","rank_in_archive_order":1,"of":1,"metrics":{"AUC":"0.949","Average Distance":"0.105"},"uses_additional_data":true},{"leaderboard":"/sota/gaze-target-estimation-on","task":"Gaze Target Estimation","dataset":"VideoAttentionTarget","model":"ViTGaze","rank_in_archive_order":1,"of":1,"metrics":{"AP":"0.905","AUC":"0.938","Average Distance":"0.102"},"uses_additional_data":true}],"syntology":{"syntology_url":"https://syntology.ai/paper/2403.12778","atlas_url":"https://app.syntology.ai/?focus=2403.12778","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}