addyDoctor images and to reproduce the results reported in the manuscript. All code used to implement the MO–BBBC framework, multitask classifier, uncertainty-aware curricula, calibration routines, and evaluation scripts (including leakage audits, OOD evaluation, and generation of all tables and figures) is freely available at: https://github.com/manhas82/MO-BBBC-Rice.git . Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Ethics statement This study uses only publicly available plant imagery and does not involve human participants, animal
Open resource ↗https://github.com/manhas82/MO-BBBC-Rice.git · lines:361-386Unverified paper record
Multi-objective Big Bang Big Crunch framework for reliable rice disease and variety classification with conditional calibration.
PloS one · 20 Mar 2026 · 10.1371/journal.pone.0340807
Abstract
Deploying rice disease detectors in the field remains challenging because models that are accurate in the lab are often poorly calibrated and provide limited uncertainty estimates, raising the risk of costly misclassification. This paper proposes a multi-objective Big-Bang Big-Crunch (MO-BBBC) framework that jointly performs disease detection and variety classification while optimizing six deployment-oriented criteria: classification error, calibration quality, uncertainty estimation, model size, inference latency, and energy consumption. The proposed framework presents conditional temperature scaling, an adaptive scheme that mitigates over-calibration and preserves reliability. The framework is implemented in Python on a lightweight two-headed classifier and evaluated on the Paddy Doctor dataset, MO-BBBC base framework achieves 90.6% disease accuracy and 97.9% variety accuracy; improves calibration to [Formula: see text] ([Formula: see text]% better than strong post-hoc baselines); achieves micro-AUC of 0.994/0.999 and micro-AP of 0.961/0.994 (disease/variety); delivers robust OOD detection (AUROC = 0.887/0.886); and supports real-time inference at [Formula: see text] ms and [Formula: see text] ms per 64-sample batch on CPU/GPU with Monte Carlo Dropout uncertainty. The resulting Pareto set enables practitioners to trade accuracy for efficiency and reliability, narrowing the gap between prototype validation and field deployment in precision agriculture.
Plant phenotyping relevance
植物病害状態を画像から検出する分類・校正・不確実性推定フレームワークが研究の中心であり、植物の病害表現型を対象とした手法開発と評価に該当する。
abstractThis paper proposes a multi-objective Big-Bang Big-Crunch (MO-BBBC) framework that jointly performs disease detection and variety classification while optimizing six deployment-oriented criteria
abstractThe proposed framework presents conditional temperature scaling, an adaptive scheme that mitigates over-calibration and preserves reliability.
abstractThe framework is implemented in Python on a lightweight two-headed classifier and evaluated on the Paddy Doctor dataset
Code and data availability
The paper publicly releases its authors' analysis code (MO-BBBC framework, calibration, leakage audits, evaluation scripts) on GitHub, and a Zenodo deposit containing the paper-specific split indices, metadata, and reproducibility notebook. The underlying PaddyDoctor plant image dataset used for all phenotyping/class-
All data underlying the findings in this study are available without restriction. The minimal dataset underlying the reported analyses (including the exact group-aware train/validation/test split indices, supporting metadata, and a complete reproducibility notebook) is publicly available on Zenodo at: https://doi.org/10.5281/zenodo.18471419 . The underlying images and labels used in this work come from the public PaddyDoctor image dataset [ 18 ], which can be accessed from the official project page https://paddydoc.github.io/dataset/ and via its IEEE DataPort record https://ieee-dataport.org/documents/paddy-doctor-visual-image-dataset-automated-paddy-disease-class
Open resource ↗Zenodo · 10.5281/zenodo.18471419 · lines:361-386validation/test split indices, supporting metadata, and a complete reproducibility notebook) is publicly available on Zenodo at: https://doi.org/10.5281/zenodo.18471419 . The underlying images and labels used in this work come from the public PaddyDoctor image dataset [ 18 ], which can be accessed from the official project page https://paddydoc.github.io/dataset/ and via its IEEE DataPort record https://ieee-dataport.org/documents/paddy-doctor-visual-image-dataset-automated-paddy-disease-classification-and-benchmarking . The Zenodo record contains the files needed to reconstruct our exact experimental partitions from the original PaddyDoctor images and to reproduce the results reported
Open resource ↗lines:361-386producibility notebook) is publicly available on Zenodo at: https://doi.org/10.5281/zenodo.18471419 . The underlying images and labels used in this work come from the public PaddyDoctor image dataset [ 18 ], which can be accessed from the official project page https://paddydoc.github.io/dataset/ and via its IEEE DataPort record https://ieee-dataport.org/documents/paddy-doctor-visual-image-dataset-automated-paddy-disease-classification-and-benchmarking . The Zenodo record contains the files needed to reconstruct our exact experimental partitions from the original PaddyDoctor images and to reproduce the results reported in the manuscript. All code used to implement the MO–BBBC framework, multi
Open resource ↗IEEE DataPort · lines:411-413This is an automatically classified, unverified record. Curator approval is required before any resource enters the Catalog.