{"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/fru-adapter-frame-recalibration-unit-adapter","title":"FRU-Adapter: Frame Recalibration Unit Adapter for Dynamic Facial Expression Recognition","arxiv_id":null,"date":"2025-02-28","proceeding":"Electronics 2025 2","authors":["Myungbeom Her","Hamza Ghulam Nabi","and Ji-HyeongHan*"],"abstract":"Dynamic facial expression recognition (DFER) is one of the most important\r\n challenges in computer vision, as it plays a crucial role in human–computer interaction. Re\r\ncently, adapter-based approaches have been introduced into DFER, and they have achieved\r\n remarkable success. However, the adapters still suffer from the following problems: over\r\nlooking irrelevant frames and interference with pre-trained information. In this paper, we\r\n propose a frame recalibration unit adapter (FRU-Adapter) which combines the strengths of\r\n a frame recalibration unit (FRU) and temporal self-attention (T-SA) to address the afore\r\nmentioned issues. The FRU initially recalibrates the frames by emphasizing important\r\n frames and suppressing less relevant frames. The recalibrated frames are then fed into\r\n T-SA to capture the correlations between meaningful frames. As a result, the FRU-Adapter\r\n captures enhanced temporal dependencies by considering the irrelevant frames in a clip.\r\n Furthermore, we propose a method for attaching the FRU-Adapter to each encoder layer\r\n in parallel to reduce the loss of pre-trained information. Notably, the FRU-Adapter uses\r\n only 2% of the total training parameters per task while achieving an improved accuracy.\r\n Extended experiments on DFER tasks show that the proposed FRU-Adapter not only out\r\nperforms the state-of-the-art models but also exhibits parameter efficiency. The source code\r\n will be made publicly available.","url_abs":"https://www.mdpi.com/2079-9292/14/5/978","url_pdf":"https://www.mdpi.com/2079-9292/14/5/978","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":"fru-adapter-frame-recalibration-unit-adapter","repo_url":"https://github.com/SeoulTech-HCIRLab/FRU-Adapter","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"dynamic-facial-expression-recognition","task_name":"Dynamic Facial Expression Recognition"},{"task_slug":"facial-expression-recognition-1","task_name":"Facial Expression Recognition"}],"methods":[{"method_slug":"adapter","method_name":"Adapter"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}