Intellectual Property India Publishes Patent Application for A Multi-Cnn Based System And Method For Automated Kidney Disease Detection From Ct Images

MUMBAI, India, June 30 -- Intellectual Property India has published a patent application (202641076186 A) filed by Sr University on June 19, 2026, for A Multi-Cnn Based System And Method For Automated Kidney Disease Detection From Ct Images.

Inventors include Niharika Vemula; and Durgesh Nandan.

The application for the patent was published on June 26, 2026, under issue no. 26/2026.

Abstract: A MULTI-CNN BASED SYSTEM AND METHOD FOR AUTOMATED KIDNEY DISEASE DETECTION FROM CT IMAGES This study establishes an effective, stable, and interpretable deep learning approach for the automatic diagnosis and classification of renal illnesses based on CT images. Overcoming the major constraints of the existing art, the proposed system uses the latest preprocessing strategies to break an image down and standardize images that are beneficial to eliminate the existence of irrelevant structures and artifacts that may compromise the model accuracy. Data augmentation strategies improve the variety and decrease overfitting, resulting in good generalization in heterogeneous patient cohorts and imaging settings. The invention combines several optimized CNN architectures, such as VGG, ResNet, DenseNet, and MobileNet, to achieve high-fidelity extraction of spatially hierarchical features and still has low computational costs for use in everyday clinical practice. On balanced multi-class datasets, the system attains an accuracy of over 99 %, which shows that the system is reliable to classify cases with normal, cyst, stone, and tumor. In addition, an integrated XAI module produces explainable heatmap images that visualize which disease-related areas of the CT scans are predicted by the model, allowing clinicians to know, verify, and trust the predictions. In general, this study makes a considerable contribution to the computerization of the diagnosis of kidney disease in terms of increasing the accuracy of diagnosis and its efficiency and interpretability, as well as adherence to common clinical routines.

Disclaimer: Curated by HT Syndication.