File Download
There are no files associated with this item.
Supplementary
-
Citations:
- Appears in Collections:
Conference Paper: Polygon-free: Unconstrained Scene Text Detection with Box Annotations
Title | Polygon-free: Unconstrained Scene Text Detection with Box Annotations |
---|---|
Authors | |
Issue Date | 2022 |
Publisher | IEEE. |
Citation | 29th IEEE International Conference on Image Processing (ICIP), Bordeaux, France, 16-19 October, 2022. In Proceedings of IEEE International Conference on Image Processing (ICIP) 2022 How to Cite? |
Abstract | Although a polygon is a more accurate representation than an upright bounding box for text detection, the annotations of polygons are extremely expensive and challenging. Unlike existing works that employ fully-supervised training with polygon annotations, this study proposes an unconstrained text detection system termed Polygon-free (PF), in which most existing polygon-based text detectors ( e.g., PSENet [33],DB [16]) are trained with only upright bounding box annotations. Our core idea is to transfer knowledge from synthetic data to real data to enhance the supervision information of upright bounding boxes. This is made pos-sible with a simple segmentation network, namely Skeleton Attention Segmentation Network (SASN), that includes three vital components ( i.e., channel attention, spatial attention and skeleton attention map) and one soft cross-entropy loss. Experiments demonstrate that the proposed Polygon-free system can combine general detectors ( e.g., EAST, PSENet, DB) to yield surprisingly high-quality pixel-level results with only upright bounding box annotations on a variety of datasets ( e.g., ICDAR2019-Art, TotalText, IC-DAR2015). For example, without using polygon annotations, PSENet achieves an 80.5% F-score on TotalText [3] (vs. 80.9% of fully supervised counterpart), 31.1% better than training directly with upright bounding box annotations, and saves 80%+ labeling costs. We hope that PF can provide a new perspective for text detection to reduce the labeling costs. |
Persistent Identifier | http://hdl.handle.net/10722/315807 |
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Wu, W | - |
dc.contributor.author | Xie, E | - |
dc.contributor.author | Zhang, R | - |
dc.contributor.author | Wang, W | - |
dc.contributor.author | Luo, P | - |
dc.contributor.author | Zhou, H | - |
dc.date.accessioned | 2022-08-19T09:04:48Z | - |
dc.date.available | 2022-08-19T09:04:48Z | - |
dc.date.issued | 2022 | - |
dc.identifier.citation | 29th IEEE International Conference on Image Processing (ICIP), Bordeaux, France, 16-19 October, 2022. In Proceedings of IEEE International Conference on Image Processing (ICIP) 2022 | - |
dc.identifier.uri | http://hdl.handle.net/10722/315807 | - |
dc.description.abstract | Although a polygon is a more accurate representation than an upright bounding box for text detection, the annotations of polygons are extremely expensive and challenging. Unlike existing works that employ fully-supervised training with polygon annotations, this study proposes an unconstrained text detection system termed Polygon-free (PF), in which most existing polygon-based text detectors ( e.g., PSENet [33],DB [16]) are trained with only upright bounding box annotations. Our core idea is to transfer knowledge from synthetic data to real data to enhance the supervision information of upright bounding boxes. This is made pos-sible with a simple segmentation network, namely Skeleton Attention Segmentation Network (SASN), that includes three vital components ( i.e., channel attention, spatial attention and skeleton attention map) and one soft cross-entropy loss. Experiments demonstrate that the proposed Polygon-free system can combine general detectors ( e.g., EAST, PSENet, DB) to yield surprisingly high-quality pixel-level results with only upright bounding box annotations on a variety of datasets ( e.g., ICDAR2019-Art, TotalText, IC-DAR2015). For example, without using polygon annotations, PSENet achieves an 80.5% F-score on TotalText [3] (vs. 80.9% of fully supervised counterpart), 31.1% better than training directly with upright bounding box annotations, and saves 80%+ labeling costs. We hope that PF can provide a new perspective for text detection to reduce the labeling costs. | - |
dc.language | eng | - |
dc.publisher | IEEE. | - |
dc.relation.ispartof | Proceedings of IEEE International Conference on Image Processing (ICIP) 2022 | - |
dc.rights | Proceedings of IEEE International Conference on Image Processing (ICIP). Copyright © IEEE. | - |
dc.title | Polygon-free: Unconstrained Scene Text Detection with Box Annotations | - |
dc.type | Conference_Paper | - |
dc.identifier.email | Luo, P: pluo@hku.hk | - |
dc.identifier.authority | Luo, P=rp02575 | - |
dc.identifier.doi | 10.48550/arXiv.2011.13307 | - |
dc.identifier.hkuros | 335610 | - |
dc.publisher.place | United States | - |