Single-Character-Based Embedding Feature Aggregation Using Cross-Attention for Scene Text Super-Resolution

In textual vision scenarios, super-resolution aims to enhance textual quality and readability to facilitate downstream tasks. However, the ambiguity of character regions in complex backgrounds remains challenging to mitigate, particularly the interference between tightly connected characters. In thi...

Full description

Saved in:
Bibliographic Details
Main Authors: Meng Wang, Qianqian Li, Haipeng Liu
Format: Article
Language:English
Published: MDPI AG 2025-04-01
Series:Sensors
Subjects:
Online Access:https://www.mdpi.com/1424-8220/25/7/2228
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849769329748344832
author Meng Wang
Qianqian Li
Haipeng Liu
author_facet Meng Wang
Qianqian Li
Haipeng Liu
author_sort Meng Wang
collection DOAJ
description In textual vision scenarios, super-resolution aims to enhance textual quality and readability to facilitate downstream tasks. However, the ambiguity of character regions in complex backgrounds remains challenging to mitigate, particularly the interference between tightly connected characters. In this paper, we propose single-character-based embedding feature aggregation using cross-attention for scene text super-resolution (SCE-STISR) to solve this problem. Firstly, a dynamic feature extraction mechanism is employed to adaptively capture shallow features by dynamically adjusting multi-scale feature weights based on spatial representations. During text–image interactions, a dual-level cross-attention mechanism is introduced to comprehensively aggregate the cropped single-character features with textual prior, also aligning semantic sequences and visual features. Finally, an adaptive normalized color correction operation is applied to mitigate color distortion caused by background interference. In TextZoom benchmarking, the text recognition accuracies of different recognizers are 53.6%, 60.9%, and 64.5%, which are improved by 0.9–1.4% over the baseline TATT, achieving an optimal SSIM value of 0.7951 and a PSNR of 21.84. Additionally, our approach improves accuracy by 0.2–2.2% over existing baselines on five text recognition datasets, validating the effectiveness of the model.
format Article
id doaj-art-91a42cc04b7c46ffab9f03e56f68d0bf
institution DOAJ
issn 1424-8220
language English
publishDate 2025-04-01
publisher MDPI AG
record_format Article
series Sensors
spelling doaj-art-91a42cc04b7c46ffab9f03e56f68d0bf2025-08-20T03:03:27ZengMDPI AGSensors1424-82202025-04-01257222810.3390/s25072228Single-Character-Based Embedding Feature Aggregation Using Cross-Attention for Scene Text Super-ResolutionMeng Wang0Qianqian Li1Haipeng Liu2School of Information Engineering and Automation, Kunming University of Science and Technology, Kunming 650500, ChinaSchool of Information Engineering and Automation, Kunming University of Science and Technology, Kunming 650500, ChinaSchool of Information Engineering and Automation, Kunming University of Science and Technology, Kunming 650500, ChinaIn textual vision scenarios, super-resolution aims to enhance textual quality and readability to facilitate downstream tasks. However, the ambiguity of character regions in complex backgrounds remains challenging to mitigate, particularly the interference between tightly connected characters. In this paper, we propose single-character-based embedding feature aggregation using cross-attention for scene text super-resolution (SCE-STISR) to solve this problem. Firstly, a dynamic feature extraction mechanism is employed to adaptively capture shallow features by dynamically adjusting multi-scale feature weights based on spatial representations. During text–image interactions, a dual-level cross-attention mechanism is introduced to comprehensively aggregate the cropped single-character features with textual prior, also aligning semantic sequences and visual features. Finally, an adaptive normalized color correction operation is applied to mitigate color distortion caused by background interference. In TextZoom benchmarking, the text recognition accuracies of different recognizers are 53.6%, 60.9%, and 64.5%, which are improved by 0.9–1.4% over the baseline TATT, achieving an optimal SSIM value of 0.7951 and a PSNR of 21.84. Additionally, our approach improves accuracy by 0.2–2.2% over existing baselines on five text recognition datasets, validating the effectiveness of the model.https://www.mdpi.com/1424-8220/25/7/2228scene text image super-resolutioncross-attentioncross-fertilizationtext recognition
spellingShingle Meng Wang
Qianqian Li
Haipeng Liu
Single-Character-Based Embedding Feature Aggregation Using Cross-Attention for Scene Text Super-Resolution
Sensors
scene text image super-resolution
cross-attention
cross-fertilization
text recognition
title Single-Character-Based Embedding Feature Aggregation Using Cross-Attention for Scene Text Super-Resolution
title_full Single-Character-Based Embedding Feature Aggregation Using Cross-Attention for Scene Text Super-Resolution
title_fullStr Single-Character-Based Embedding Feature Aggregation Using Cross-Attention for Scene Text Super-Resolution
title_full_unstemmed Single-Character-Based Embedding Feature Aggregation Using Cross-Attention for Scene Text Super-Resolution
title_short Single-Character-Based Embedding Feature Aggregation Using Cross-Attention for Scene Text Super-Resolution
title_sort single character based embedding feature aggregation using cross attention for scene text super resolution
topic scene text image super-resolution
cross-attention
cross-fertilization
text recognition
url https://www.mdpi.com/1424-8220/25/7/2228
work_keys_str_mv AT mengwang singlecharacterbasedembeddingfeatureaggregationusingcrossattentionforscenetextsuperresolution
AT qianqianli singlecharacterbasedembeddingfeatureaggregationusingcrossattentionforscenetextsuperresolution
AT haipengliu singlecharacterbasedembeddingfeatureaggregationusingcrossattentionforscenetextsuperresolution