{"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/htr-vt-handwritten-text-recognition-with","title":"HTR-VT: Handwritten Text Recognition with Vision Transformer","arxiv_id":"2409.08573","date":"2024-09-13","proceeding":null,"authors":["Yuting Li","Dexiong Chen","Tinglong Tang","Xi Shen"],"abstract":"We explore the application of Vision Transformer (ViT) for handwritten text recognition. The limited availability of labeled data in this domain poses challenges for achieving high performance solely relying on ViT. Previous transformer-based models required external data or extensive pre-training on large datasets to excel. To address this limitation, we introduce a data-efficient ViT method that uses only the encoder of the standard transformer. We find that incorporating a Convolutional Neural Network (CNN) for feature extraction instead of the original patch embedding and employ Sharpness-Aware Minimization (SAM) optimizer to ensure that the model can converge towards flatter minima and yield notable enhancements. Furthermore, our introduction of the span mask technique, which masks interconnected features in the feature map, acts as an effective regularizer. Empirically, our approach competes favorably with traditional CNN-based models on small datasets like IAM and READ2016. Additionally, it establishes a new benchmark on the LAM dataset, currently the largest dataset with 19,830 training text lines. The code is publicly available at: https://github.com/YutingLi0606/HTR-VT.","url_abs":"https://arxiv.org/abs/2409.08573v1","url_pdf":"https://arxiv.org/pdf/2409.08573v1.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":"htr-vt-handwritten-text-recognition-with","repo_url":"https://github.com/yutingli0606/htr-vt","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":null},{"paper_slug":"htr-vt-handwritten-text-recognition-with","repo_url":"https://github.com/mortezagolzan/Hand_Writing_Text_Extarction","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"htr","task_name":"HTR"},{"task_slug":"handwritten-text-recognition","task_name":"Handwritten Text Recognition"}],"methods":[{"method_slug":"absolute-position-encodings","method_name":"Absolute Position Encodings"},{"method_slug":"adam","method_name":"Adam"},{"method_slug":"attention","method_name":"Attention"},{"method_slug":"bpe","method_name":"BPE"},{"method_slug":"dense-connections","method_name":"Dense Connections"},{"method_slug":"dropout","method_name":"Dropout"},{"method_slug":"label-smoothing","method_name":"Label Smoothing"},{"method_slug":"layer-normalization","method_name":"Layer Normalization"},{"method_slug":"linear-layer","method_name":"Linear Layer"},{"method_slug":"multi-head-attention","method_name":"Multi-Head Attention"},{"method_slug":"position-wise-feed-forward-layer","method_name":"Position-Wise Feed-Forward Layer"},{"method_slug":"residual-connection","method_name":"Residual Connection"},{"method_slug":"sharpness-aware-minimization","method_name":"Sharpness-Aware Minimization"},{"method_slug":"softmax","method_name":"Softmax"},{"method_slug":"transformer","method_name":"Transformer"},{"method_slug":"vision-transformer","method_name":"Vision Transformer"}],"datasets_introduced":[],"methods_introduced":[],"results":[{"leaderboard":"/sota/handwritten-text-recognition-on-iam","task":"Handwritten Text Recognition","dataset":"IAM","model":"HTR-VT(line-level)","rank_in_archive_order":9,"of":17,"metrics":{"CER":"4.7","WER":"14.9"},"uses_additional_data":false},{"leaderboard":"/sota/handwritten-text-recognition-on-iam-line","task":"Handwritten Text Recognition","dataset":"IAM(line-level)","model":"HTR-VT","rank_in_archive_order":2,"of":5,"metrics":{"Test CER":"4.7","Test WER":"14.9"},"uses_additional_data":false},{"leaderboard":"/sota/handwritten-text-recognition-on-lam-line","task":"Handwritten Text Recognition","dataset":"LAM(line-level)","model":"HTR-VT","rank_in_archive_order":1,"of":6,"metrics":{"Test CER":"2.8","Test WER":"7.4"},"uses_additional_data":false},{"leaderboard":"/sota/handwritten-text-recognition-on-read-2016","task":"Handwritten Text Recognition","dataset":"READ 2016","model":"HTR-VT(line-level)","rank_in_archive_order":2,"of":2,"metrics":{"CER (%)":"3.9","WER (%)":"16.5"},"uses_additional_data":false},{"leaderboard":"/sota/handwritten-text-recognition-on-read2016-line","task":"Handwritten Text Recognition","dataset":"READ2016(line-level)","model":"HTR-VT","rank_in_archive_order":1,"of":5,"metrics":{"Test CER":"3.9","Test WER":"16.5"},"uses_additional_data":false}],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=2409.08573","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}