Evaluating Plant Disease Detection Mobile Applications: Quality and Limitations

In this technologically advanced era, with the proliferation of artificial intelligence, many mobile apps are available for plant disease detection, diagnosis, and treatment, each with a variety of features. These apps need to be categorized and reviewed following a proper framework that ensures their quality. This study aims to present an approach to evaluating plant disease detection mobile apps, this includes providing ratings of distinct features of the apps and insights into the exploitation of artificial intelligence used in plant disease detection. For this purpose, plant disease detection apps were searched in three prominent app stores using a set of keywords. A total of 606 apps were found and from them 17 relevant apps were identified based on inclusion and exclusion criteria. The selected apps were reviewed by three raters using our devised app rating scale. User comments from the app stores are collected and analyzed to understand their expectations and views. Following the rating procedure, most apps earned acceptable ratings in software quality characteristics such as aesthetics, usability, and performance, but gained poor ratings in AI-based advanced functionality, which is the key aspect of this study. However, most of the apps cannot be used as a complete solution to plant disease detection, diagnosis, and treatment. Only one app, Plantix - your crop doctor, could successfully identify plants from images, detect diseases, maintain a rich plant database, and suggest potential treatments for the disease presented. It also provides a community where plant lovers can communicate with each other to gain additional benefits. In general, all existing apps need to improve functionalities, user experience, and software quality. Therefore, a set of design considerations has been proposed for future app improvements.