Guardado en:
Detalles Bibliográficos
Autores principales: Liu, Wenjun, Wu, Qian, Hu, Yifeng, Li, Yuke
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:https://arxiv.org/abs/2512.14050
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866908714525523968
author Liu, Wenjun
Wu, Qian
Hu, Yifeng
Li, Yuke
author_facet Liu, Wenjun
Wu, Qian
Hu, Yifeng
Li, Yuke
contents We introduce SELECT (Scene tExt Label Errors deteCTion), a novel approach that leverages multi-modal training to detect label errors in real-world scene text datasets. Utilizing an image-text encoder and a character-level tokenizer, SELECT addresses the issues of variable-length sequence labels, label sequence misalignment, and character-level errors, outperforming existing methods in accuracy and practical utility. In addition, we introduce Similarity-based Sequence Label Corruption (SSLC), a process that intentionally introduces errors into the training labels to mimic real-world error scenarios during training. SSLC not only can cause a change in the sequence length but also takes into account the visual similarity between characters during corruption. Our method is the first to detect label errors in real-world scene text datasets successfully accounting for variable-length labels. Experimental results demonstrate the effectiveness of SELECT in detecting label errors and improving STR accuracy on real-world text datasets, showcasing its practical utility.
format Preprint
id arxiv_https___arxiv_org_abs_2512_14050
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle SELECT: Detecting Label Errors in Real-world Scene Text Data
Liu, Wenjun
Wu, Qian
Hu, Yifeng
Li, Yuke
Computer Vision and Pattern Recognition
We introduce SELECT (Scene tExt Label Errors deteCTion), a novel approach that leverages multi-modal training to detect label errors in real-world scene text datasets. Utilizing an image-text encoder and a character-level tokenizer, SELECT addresses the issues of variable-length sequence labels, label sequence misalignment, and character-level errors, outperforming existing methods in accuracy and practical utility. In addition, we introduce Similarity-based Sequence Label Corruption (SSLC), a process that intentionally introduces errors into the training labels to mimic real-world error scenarios during training. SSLC not only can cause a change in the sequence length but also takes into account the visual similarity between characters during corruption. Our method is the first to detect label errors in real-world scene text datasets successfully accounting for variable-length labels. Experimental results demonstrate the effectiveness of SELECT in detecting label errors and improving STR accuracy on real-world text datasets, showcasing its practical utility.
title SELECT: Detecting Label Errors in Real-world Scene Text Data
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2512.14050