EUVIMEDEUVIMED entdecken
Crossref

An Interpretable Deep Learning Framework for Distinguishing Esophagitis from Barrett’s Esophagus Using Endoscopic Images

Atheequllah Khan M., Krishna J.

Journal of Innovative Image Processing · 21.07.2026 · 8 (3) · 1170-1195

DOI: 10.36548/jiip.2026.3.020 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Innovative Image Processing, online erschienen am 21.07.2026. Verfasst von Atheequllah Khan M. und 1 weiteren Autor:innen. Die Arbeit nennt 24 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The accurate distinction between esophagitis and Barrett's Esophagus is important for prompt diagnosis and clinical management, since Barrett's Esophagus is known to be a precursor of esophageal adenocarcinoma. However, the endoscopic appearance is difficult to interpret, as the overall attachment and appearance of the mucosa are similar and can be interpreted differently among clinicians. The aim of this work is to present an interpretable deep learning framework to automatically classify endoscopic images of esophagitis or Barrett's Esophagus into either binary class. The framework was validated on a balanced set, which has been built from the publicly available Kvasir and HyperKvasir databases. To gain insight into the model transparency, Gradient-weighted Class Activation Mapping (Grad-CAM) was used to display the regions of the image that were more relevant for the classification decision. Three-fold stratified cross-validation (CV) was used to evaluate the experimental results and the fine-tuned CNNs consistently outperformed the feature extraction based methods. The most recent model that was evaluated was fine-tuned InceptionV3 with an accuracy of 99.80%, a precision of 99.60%, a recall of 100.00%, an F1-score of 99.80%, and an area under the receiver operating characteristic curve (AUC) of 1.000. The model was found to be more interpretable, as shown by Grad-CAM, which indicated that the model learned about clinically relevant mucosal abnormalities. The results of this study support the potential of the proposed framework as an accurate and interpretable computer-aided diagnostic tool for assessment of esophageal diseases in endoscopy.

Zitieren

APA

M., A. K., & J., K. (2026). An Interpretable Deep Learning Framework for Distinguishing Esophagitis from Barrett’s Esophagus Using Endoscopic Images. Journal of Innovative Image Processing, 8(3), 1170-1195. https://doi.org/10.36548/jiip.2026.3.020

Vancouver

M. AK, J. K. An Interpretable Deep Learning Framework for Distinguishing Esophagitis from Barrett’s Esophagus Using Endoscopic Images. Journal of Innovative Image Processing. 2026;8(3):1170-1195. doi:10.36548/jiip.2026.3.020

BibTeX
@article{M2026,
  author = {M., Atheequllah Khan and J., Krishna},
  title = {An Interpretable Deep Learning Framework for Distinguishing Esophagitis from Barrett’s Esophagus Using Endoscopic Images},
  journal = {Journal of Innovative Image Processing},
  year = {2026},
  volume = {8},
  number = {3},
  pages = {1170--1195},
  doi = {10.36548/jiip.2026.3.020}
}
Kostenlos starten →