Detection of ring cell cancer in histopathological images with region of interest determined by SLIC superpixels method

Budak C., Mençik V.

Neural Computing and Applications, vol.34, no.16, pp.13499-13512, 2022 (SCI-Expanded) identifier

  • Publication Type: Article / Article
  • Volume: 34 Issue: 16
  • Publication Date: 2022
  • Doi Number: 10.1007/s00521-022-07183-8
  • Journal Name: Neural Computing and Applications
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Academic Search Premier, PASCAL, Applied Science & Technology Source, Biotechnology Research Abstracts, Compendex, Computer & Applied Sciences, Index Islamicus, INSPEC, zbMATH
  • Page Numbers: pp.13499-13512
  • Keywords: Deep learning, Diagnosing cancer, Gastric cancer, Histopathological images, Region of Interest (RoI)
  • Abdullah Gül University Affiliated: No


Gastric cancer is the sixth most common cancer and the fourth leading cause of cancer deaths worldwide. Gastric cancer presents with a more insidious onset and is most frequently discovered at an advanced stage. Early diagnosis is critical since the stage of the disease is determinant in the severity, treatment, and survival rate of cancer. In the study, the Region of Interest (RoI) was determined in histopathological images using image preprocessing techniques and signet ring cell carcinoma (SRCC) was detected with popular deep learning models VGG16, VGG19, and InceptionV3. The fine-tuning strategy was applied by customizing the last five layers of deep network models based on the target data. The parameters of accuracy, precision, recall, and F1-score were used to evaluate the model performance. Signet ring cell dataset taken from the competition “Digestive System Pathological Detection, and Segmentation Challenge 2019” was employed. When compared to results of the DigestPath2019 Grand challenge ring cell gastric cancer competition, higher accuracy rates were obtained using deep learning models with the accurate defined RoI images. VGG16 model exhibited a higher performance with accuracy of 95% and a F1-score of 95% among the models. The results obtained by the algorithm were analyzed and confirmed by the experienced pathologist.