Vijaya Saraswathi, a teacher-researcher at VNR Vignana Jyothi Institute of Engineering and Technology in Hyderabad

How a computer networking doctoral student turned decades of watching her father struggle with crop losses from plant leaf disease into an AI breakthrough. Vijaya Saraswathi, a teacher-researcher at VNR Vignana Jyothi Institute of Engineering and Technology in Hyderabad, knew firsthand how diseases can cause losses to tomato, potato, and pepper farmers.

Those childhood memories laid the groundwork for a deeply personal mission – to find a solution to leaf disease and help farmers catch plant diseases early, preventing heavy harvests from going to waste.

She, along with a team of researchers, has amassed over 20,000 publicly available images and developed a solution to help farmers address the problem. The team secured a patent for the “Leaf Disease Detection System Using Convolutional Neural Networks”.

The team built their work on a Kaggle PlantVillage dataset that contained over 20,000 images. They trained a Convolutional Neural Network (CNN) model using standard image processing, feature extraction, and segmentation steps.

Leaf diseases, caused by pests, fungi, and bacteria, can pose a serious threat to food security by destroying crops from the top and bottom. Traditionally, spotting these diseases requires continuous, laborious monitoring. To prevent heavy losses and boost yields, early and automated detection has become vital.

“Early and accurate disease identification can help farmers make informed decisions, optimise pesticide usage, and enhance agricultural productivity,” she said.

“This solution reflects how AI and deep learning are being applied to solve real-world agricultural challenges,” she said.

“Our solution stands apart from existing systems by scoring a 96% accuracy rate. The work started in 2022, entered the patent process in 2023, and finally received approval in April 2026,” Saraswati told businessline.

Besides just detecting the disease, the system automatically suggests the right pesticide to resolve the issue. Currently, it is available on a web interface. “We are planning to build a mobile app to make it much more accessible for farmers,” she said.

Published on July 21, 2026