MUMBAI, India, July 7 -- Intellectual Property India has published a patent application (202641079726 A) filed by Sona College Of Technology on June 29, 2026, for A Multi-Stream Cnn System And Method For Automated Banana Plant Disease Diagnosis And Treatment Recommendation.
Inventors include Karthick K; Selvanathan N; Rajkumar S; Gokul N; Edwine Innocent X; and Indra Prakash L.
The application for the patent was published on July 03, 2026, under issue no. 27/2026.
Abstract: ABSTRACT The present invention relates to a deep learning-based system and method for automated diagnosis of banana plant diseases using a multi-stream convolutional neural network (Multi-CNN) architecture. The system receives a banana leaf image and performs preprocessing including normalization, segmentation, and image enhancement. The preprocessed image is simultaneously processed through a first convolutional neural network stream configured to extract chromatic disease indicators, a second convolutional neural network stream configured to extract texture-based disease indicators, and a third convolutional neural network stream configured to extract structural and edge-based disease indicators. Feature vectors generated by the respective streams are fused to form a unified disease representation, which is processed by a classification engine to identify a disease category or a healthy condition. The system further retrieves a corresponding treatment protocol from a treatment knowledge repository and generates disease-specific recommendations. The invention improves disease classification accuracy by reducing feature interference and enables real-time deployment on mobile, edge-computing, and cloud-based platforms.
Disclaimer: Curated by HT Syndication.
