Deep Learning in Leaf Disease Detection (2014-2024): A Visualization-Based Bibliometric Analysis

Abstract

The agriculture industry is critical to delivering high-quality food and contributes significantly to the growth of economies and people which can be affected by the plant disease. This article demonstrates a visualization based bibliometric analysis to depict research trends in deep learning-based leaf disease detection from 2014 to January 2024. The publications used in this study are collected from the Scopus database. The research distributions with respect to sources and country, research trends, and research limits for deep learning in leaf disease detection studies are presented using Biblioshiny and VOSViewer software and visualization technologies. From 2014 to January 2024, the literature on this field has grown at an average rate of 53.41%. 1307 peer-reviewed publications from 54 countries are identified that are published in 594 distinct sources. India is the most productive country, accounting for 36.6% of total publications and 23% of total citations. Chitkara University Institute of Engineering and Technology was the most productive research institute, with 66 publications and 291 citations, while Computers and Electronics in Agriculture journal has the most citations in deep learning-based leaf disease detection research. The findings, in particular, show that "Convolution Neural Network", "Transfer Learning", "Ensemble Learning", etc., are the most widely used research topics in this field from 2014 to January 2024, and the research interest engrossed on applications of deep learning standard architectures. This study gives an insight into deep learning in leaf disease detection's general research patterns, which may assist researchers better understand and forecast the field's dynamic paths.