发表机构
Sardar Vallabhbhai National Institute of Technology; Norwegian University of Science and Technology (NTNU)（萨达尔·瓦拉巴伊国家理工学院; 挪威科技大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本研究提出基于Transformer的无监督GAN框架UnCapsTSR，用于提升低分辨率胶囊内镜图像的空间分辨率，通过BTV损失保证空间连续性，引入EndoQM指标，在多数据集上实现指标提升40%-80%。
AI 中文摘要
无线胶囊内镜（WCE）在患者胃肠道（GI）道中移动时采集并传输视频，用于检查胃肠道异常。尽管WCE比传统内镜更具优势，但受胶囊尺寸和无线传输限制，其图像分辨率较低。本研究提出UnCapsTSR，一种基于Transformer的无监督生成对抗网络（GAN）框架，用于提升低分辨率（LR）WCE图像的空间分辨率。该方法无需对真实世界LR数据进行显式退化估计，也无需真实LR-HR配对数据，采用双边总变分（BTV）损失确保超分辨率（SR）图像的空间连续性。研究还提供了从Kvasir胶囊数据集整理的新数据集，用于训练WCE SR模型，在未参与训练的KID和GIANA数据集上验证了泛化能力，并引入了针对WCE数据的非参考定量指标内镜质量度量（EndoQM）。实验结果显示，与现有最优无监督SR方法相比，在NIQE、BRISQUE、PIQE和EndoQM指标上均实现稳定提升，统计评估表明，在所有评估数据集上，EndoQM指标从LR到SR提升了40%至80%。
英文摘要
Wireless Capsule Endoscopy (WCE) captures and streams video while passing through a patient's Gastrointestinal (GI) tract and is used to examine its irregularities. Although advantageous over conventional endoscopy, WCE suffers from limitations related to capsule size and wireless transmission, resulting in images with coarser resolution. This work presents UnCapsTSR, an unsupervised transformer-based Generative Adversarial Network (GAN) framework for improving the spatial resolution of Low-Resolution (LR) WCE images. The proposed method accomplishes SR without explicit degradation estimation of real-world LR data and eliminates the need for true LR-HR pairs. UnCapsTSR employs a Bilateral Total Variation (BTV) loss to ensure spatial continuity in SR images. A newly curated dataset from the Kvasir Capsule dataset is also presented for training WCE SR models. Generalizability is validated on KID and GIANA datasets that are not used during training. A new non-reference metric, Endoscopy Quality Metric (EndoQM), is introduced for quantitative evaluation of domain-specific WCE data. Experiments demonstrate consistent improvement over state-of-the-art unsupervised SR approaches using NIQE, BRISQUE, PIQE, and EndoQM. Statistical evaluation shows 40 to 80 percent improvement in EndoQM from LR to SR across the evaluated datasets.
CommentsAccepted manuscript of the article published in Neurocomputing, Volume 665, Article 132161, 2026
Journal refNeurocomputing, Volume 665, Article 132161, 2026
DOI：10.1016/j.neucom.2025.132161
