AIDoctor-Plant: Lightweight and Bilingual Plant Disease Diagnosis and Remedy Recommender System
Year: 2025
Type: article
Abstract: Diagnosing plant diseases and providing advisory for disease remediation are crucial for the agricultural sector’s growth, as it increases the crop yield and farmer’s profit. In literature, numerous works have been done to detect plant diseases and provide advisory for disease remediation. However, all existing works provide advisory without considering the severity of disease, which can significantly affect disease management practices. Moreover, some existing works follow client-server architecture and thus rely on an active Internet connection for plant disease diagnosis, which may not be available in remote locations. Hence, in order to mitigate these research gaps, a lightweight and few-shot framework named “AIDoctor-Plant” is proposed and developed as an Android mobile application for plant disease diagnosis and remedy recommendation in this paper. This application works in both English and Hindi, and it also considers disease severity to generate disease management advisory. The effectiveness of the AIDoctor-Plant application is evaluated with the help of Apple-Tree-Leaf-Disease-Segmentation dataset. Experimental results demonstrate that the developed application can detect diseases with 95.34% accuracy and segment diseased areas from leaf images with a mean intersection over union of 95.43%. Furthermore, Student's t-test has confirmed that the developed application precisely measures the plant disease severity with 99% confidence interval. This makes AIDoctor-Plant application suitable to be deployed in the real world for effectively and efficiently diagnosing plant diseases even in remote locations and for getting remedy measures to manage the identified disease at the estimated severity index.
Authors: Punam Bedi, Pushkar Gole, Sudeep Marwaha
Language: English
Topic: Smart Agriculture and AI
Subfield: Plant Science
Domain: Life Sciences
SDG: Zero hunger
Open Access status: hybrid
Estimated APC paid:
OpenAlex ID: W4411212916