EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Application of improved deep learning for detection of gastric cancer

Yamuna Bee Jafarullah Khan, Balaji Subramanian

Journal of Biological Regulators and Homeostatic Agents · 29.05.2026 · 40 (2)

DOI: 10.65746/jbrha68 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Biological Regulators and Homeostatic Agents, online erschienen am 29.05.2026. Verfasst von Yamuna Bee Jafarullah Khan und 1 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Gastric Cancer (GC) remains a major global health concern, with early-stage detection being critical for improving survival rates. However, identifying GC in its early stages is challenging due to subtle and diverse clinical manifestations. This research aims to develop an optimized deep learning framework to enhance the accuracy, efficiency, and robustness of GC detection, particularly in early-stage cases. A novel hybrid optimization technique, the Aquila Inherited Dragonfly Optimizer (AI-DFO), is proposed to fine-tune a Convolutional Neural Network (CNN) for GC classification. The approach incorporates advanced image pre-processing using median filtering and CLAHE, precise lesion segmentation with Mask R-CNN, and AI-DFO-based optimization to enhance feature learning and model generalization. The proposed AI-DFO+CNN model achieved state-of-the-art performance, with 99% accuracy for endoscopic images and 98% accuracy for histopathological images. Stage-wise analysis confirmed superior sensitivity in detecting both early and advanced GC. Furthermore, the model demonstrated strong robustness to noise and image deformation. Computational efficiency was improved with reduced training time and lower energy consumption compared to baseline models. The results validate the effectiveness of the proposed framework for reliable, early GC detection. The generalizable design of AI-DFO and the segmentation-classification pipeline make this approach scalable to other medical imaging tasks. Future research will focus on large-scale multi-center validation, integration with real-time clinical workflows, and the application of the AI-DFO framework to other complex medical image analysis tasks beyond gastric cancer.

Zitieren

APA

Khan, Y. B. J., & Subramanian, B. (2026). Application of improved deep learning for detection of gastric cancer. Journal of Biological Regulators and Homeostatic Agents, 40(2). https://doi.org/10.65746/jbrha68

Vancouver

Khan YBJ, Subramanian B. Application of improved deep learning for detection of gastric cancer. Journal of Biological Regulators and Homeostatic Agents. 2026;40(2). doi:10.65746/jbrha68

BibTeX
@article{Khan2026,
  author = {Khan, Yamuna Bee Jafarullah and Subramanian, Balaji},
  title = {Application of improved deep learning for detection of gastric cancer},
  journal = {Journal of Biological Regulators and Homeostatic Agents},
  year = {2026},
  volume = {40},
  number = {2},
  doi = {10.65746/jbrha68}
}
Kostenlos starten →