{"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/syncvsr-data-efficient-visual-speech","title":"SyncVSR: Data-Efficient Visual Speech Recognition with End-to-End Crossmodal Audio Token Synchronization","arxiv_id":"2406.12233","date":"2024-06-18","proceeding":null,"authors":["Young Jin Ahn","Jungwoo Park","Sangha Park","Jonghyun Choi","Kee-Eung Kim"],"abstract":"Visual Speech Recognition (VSR) stands at the intersection of computer vision and speech recognition, aiming to interpret spoken content from visual cues. A prominent challenge in VSR is the presence of homophenes-visually similar lip gestures that represent different phonemes. Prior approaches have sought to distinguish fine-grained visemes by aligning visual and auditory semantics, but often fell short of full synchronization. To address this, we present SyncVSR, an end-to-end learning framework that leverages quantized audio for frame-level crossmodal supervision. By integrating a projection layer that synchronizes visual representation with acoustic data, our encoder learns to generate discrete audio tokens from a video sequence in a non-autoregressive manner. SyncVSR shows versatility across tasks, languages, and modalities at the cost of a forward pass. Our empirical evaluations show that it not only achieves state-of-the-art results but also reduces data usage by up to ninefold.","url_abs":"https://arxiv.org/abs/2406.12233v1","url_pdf":"https://arxiv.org/pdf/2406.12233v1.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":"syncvsr-data-efficient-visual-speech","repo_url":"https://github.com/KAIST-AILab/SyncVSR","is_official":1,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"jax","reach":null}],"tasks":[{"task_slug":"landmark-based-lipreading","task_name":"Landmark-based Lipreading"},{"task_slug":"lipreading","task_name":"Lipreading"},{"task_slug":"speech-recognition","task_name":"Speech Recognition"},{"task_slug":"visual-speech-recognition","task_name":"Visual Speech Recognition"},{"task_slug":"speech-recognition-1","task_name":"speech-recognition"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/landmark-based-lipreading-on-lrs2","task":"Landmark-based Lipreading","dataset":"LRS2","model":"SyncVSR","rank_in_archive_order":1,"of":1,"metrics":{"Word Error Rate (WER)":"74.6"},"uses_additional_data":false},{"leaderboard":"/sota/landmark-based-lipreading-on-lrw","task":"Landmark-based Lipreading","dataset":"LRW","model":"SyncVSR (Word Boundary)","rank_in_archive_order":1,"of":5,"metrics":{"Top 1 Accuracy":"80.3"},"uses_additional_data":false},{"leaderboard":"/sota/landmark-based-lipreading-on-lrw","task":"Landmark-based Lipreading","dataset":"LRW","model":"SyncVSR","rank_in_archive_order":2,"of":5,"metrics":{"Top 1 Accuracy":"75.1"},"uses_additional_data":false},{"leaderboard":"/sota/lipreading-on-lrw-1000","task":"Lipreading","dataset":"CAS-VSR-W1k (LRW-1000)","model":"SyncVSR (Word Boundary)","rank_in_archive_order":1,"of":9,"metrics":{"Top-1 Accuracy":"58.2"},"uses_additional_data":false},{"leaderboard":"/sota/lipreading-on-lrs2","task":"Lipreading","dataset":"LRS2","model":"SyncVSR","rank_in_archive_order":3,"of":25,"metrics":{"Word Error Rate (WER)":"16.5"},"uses_additional_data":true},{"leaderboard":"/sota/lipreading-on-lrs2","task":"Lipreading","dataset":"LRS2","model":"SyncVSR","rank_in_archive_order":11,"of":25,"metrics":{"Word Error Rate (WER)":"28.9"},"uses_additional_data":false},{"leaderboard":"/sota/lipreading-on-lrs3-ted","task":"Lipreading","dataset":"LRS3-TED","model":"SyncVSR","rank_in_archive_order":3,"of":23,"metrics":{"Word Error Rate (WER)":"21.5"},"uses_additional_data":true},{"leaderboard":"/sota/lipreading-on-lrs3-ted","task":"Lipreading","dataset":"LRS3-TED","model":"SyncVSR","rank_in_archive_order":12,"of":23,"metrics":{"Word Error Rate (WER)":"31.2"},"uses_additional_data":false},{"leaderboard":"/sota/lipreading-on-lip-reading-in-the-wild","task":"Lipreading","dataset":"Lip Reading in the Wild","model":"SyncVSR (Word Boundary)","rank_in_archive_order":1,"of":22,"metrics":{"Top-1 Accuracy":"95.0"},"uses_additional_data":false},{"leaderboard":"/sota/lipreading-on-lip-reading-in-the-wild","task":"Lipreading","dataset":"Lip Reading in the Wild","model":"SyncVSR","rank_in_archive_order":3,"of":22,"metrics":{"Top-1 Accuracy":"93.2"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2406.12233","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}