Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{205403,
        author = {ANMOL BOSE and Anup Singh Kushwaha},
        title = {A Data-Centric Review of AI-Based Rice Disease Detection: Challenges in Generalization, Dataset Diversity, and Varietal Representation},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {13},
        number = {1},
        pages = {6648-6653},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=205403},
        abstract = {This paper presents an analysis of AI-based rice disease recognition from the perspective of data-related problems, such as issues associated with generalization and variety representation. Although deep learning architectures, especially CNNs, have demonstrated excellent classification results in controlled experiments, their ability to perform adequately in field settings is still questionable. A systematic narrative review covering peer-reviewed literature published from 2015 to 2025 was conducted. As a result, we observe that all existing models heavily depend on datasets containing samples collected under highly consistent environmental conditions and relatively few varietals. As a result, there is a risk of overfitting, and the models' generalization becomes problematic. At the same time, one important limitation identified during our research is the lack of diversity among rice varieties used in existing models, namely, black rice. The study interprets the data issue as a representational bias.},
        keywords = {Artificial intelligence, CNN, dataset diversity, plant disease detection, rice disease classification, varietal representation.},
        month = {June},
        }