MUMBAI, India, June 30 -- Intellectual Property India has published a patent application (202641056371 A) filed by S Sanjaay; V S Kirithic Adhithya; and Dr. S. Iniyan on May 04, 2026, for Apple Leaf Disease Detection Using Machine Learning And Deep Learning Models.

Inventors include S Sanjaay; V S Kirithic Adhithya; and Dr. S. Iniyan.

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

Abstract: As Under the conditions of increasing disease pressure and environmental variability in apple farming, accurate and automated disease diagnosis is very necessary. Conventional visual inspection methods used by farmers and botanists are time-consuming and unreliable, particularly in early disease stages where symptoms of Apple Scab, Black Rot, and Cedar Apple Rust can appear visually similar. In our model we have integrated deep learning feature extraction with classical machine learning classification for detecting and classifying apple leaf diseases. For the project we have used the PlantVillage dataset consisting of high-resolution apple leaf images across four classes. Our datasets contain preprocessed images that have gone through background removal, leaf segmentation, morphological operations, and feature enhancement. We have used many models like MobileNetV2, ResNet50, DenseNetl2l, EfficientNetBO, Custom CNN, SVM, Random Forest, and KNN. Out of all these models, the SVM classifier with MobileNetV2-based deep feature extraction was having the highest performance with an accuracy of 99.1 %, so we deployed our system with that configuration. An adaptive ensemble learning mechanism further combines the predictions of all individual models using a weighted voting technique, achieving ensemble confidence scores exceeding 90%. We have also built a disease severity estimation module using HSV color space analysis that classifies the extent of infection into Mild, Moderate, Severe, and Critical levels. We have built a platform through a Streamlit web application where farmers can upload leaf images and get the predicted disease class, severity level, confidence score, and visual outputs in plain agricultural language without requiring specialized technical knowledge.

Disclaimer: Curated by HT Syndication.