Deep learning approach using modified DarkNet-53 for renal cell carcinoma grading
by G. Sathish Kumar; G. Uma Maheshwari; C. Selvan; M. Nagasuresh; D. Rasi; P. Swathypriyadharsini; Sathish Kumar Danasegaran
International Journal of Bioinformatics Research and Applications (IJBRA), Vol. 21, No. 1, 2025

Abstract: Accurate and effective diagnostic procedures are required for appropriate treatment planning for renal cell carcinoma, the most frequent form of kidney cancer. Using fusion module, a network dubbed modified DarkNet (MDNet) was developed for image-based small-target detection. We built MDNet on top of a modified version of DarkNet-53, which itself a scale matching approach, to increase its speed and accuracy. By combining the results of several convolutional neural network (CNN) models, the ensemble structure improves classification accuracy. The effectiveness of a classification algorithm using kidney histopathology pictures dataset is measured in accuracy, precision, recall, sensitivity, specificity and F1-score. The results show that the ensemble deep learning method outperforms both standalone CNN models and more conventional machine learning techniques in RCC classification. Overall grade classification accuracy of 98.9%, a sensitivity of 98.2%, and a high classification specificity of 98.7%, in distinguishing tissues.

Online publication date: Tue, 21-Jan-2025

The full text of this article is only available to individual subscribers or to users at subscribing institutions.

 
Existing subscribers:
Go to Inderscience Online Journals to access the Full Text of this article.

Pay per view:
If you are not a subscriber and you just want to read the full contents of this article, buy online access here.

Complimentary Subscribers, Editors or Members of the Editorial Board of the International Journal of Bioinformatics Research and Applications (IJBRA):
Login with your Inderscience username and password:

Username:        Password:          Forgotten your password?

Want to subscribe?
A subscription gives you complete access to all articles in the current issue, as well as to all articles in the previous three years (where applicable). See our Orders page to subscribe.

If you still need assistance, please email subs@inderscience.com