Raman Spectroscopy and Improved Inception Network for Determination of FHB-Infected Wheat Kernels.

Abstract

Detection of infected kernels is important for Fusarium head blight (FHB) prevention and product quality assurance in wheat. In this study, Raman spectroscopy (RS) and deep learning networks were used for the determination of FHB-infected wheat kernels. First, the RS spectra of healthy, mild, and severe infection kernels were measured and spectral changes and band attribution were analyzed. Then, the Inception network was improved by residual and channel attention modules to develop the recognition models of FHB infection. The Inception-attention network produced the best determination with accuracies in training set, validation set, and prediction set of 97.13%, 91.49%, and 93.62%, among all models. The average feature map of the channel clarified the important information in feature extraction, itself required to clarify the decision-making strategy. Overall, RS and the Inception-attention network provide a noninvasive, rapid, and accurate determination of FHB-infected wheat kernels and are expected to be applied to other pathogens or diseases in various crops.

Plant phenotyping relevance

小麦種子のFHB感染状態という植物状態を、ラマン分光と改良深層学習モデルで非侵襲的に判定する手法を開発・評価しており、表現型取得が研究の中心です。

Code and data availability

The paper's supplementary materials (hosted publicly by MDPI) contain Figure S1, images of wheat kernels with varying degrees of FHB damage used in this study's phenotyping, plus parameter-setting tables for the classification models and networks. No separate spectral dataset or analysis code repository is stated; the