Citation :

M. Pradeep, M. Siddappa, "Enhanced Classification of Rice Varieties using Cross Modality Dynamic Bidirectional Knowledge Transfer Model," International Journal of Engineering Trends and Technology (IJETT), vol. 74, no. 4, pp. 299-312, 2026. Crossref, https://doi.org/10.14445/22315381/IJETT-V74I4P123

Abstract

In recent years, rice has been the staple food for nearly half of the global population, serving as a rich source of nutrients and contributing significantly to increased crop yields. Thus, accurate rice variety classification is essential, which corresponds to distinct spot shapes and sizes. However, existing Deep Learning (DL) models have several limitations that make it difficult to distinguish among various rice varieties because of their similar characteristics. To overcome this limitation, Centered Kernel Alignment-based Cross-Domain Knowledge Transfer (CKA-CDKT) is proposed to accurately identify and classify various types of rice grains. The EfficientNet-B7 model extracts the most relevant features that have significant information on rice varieties, which helps to efficiently enhance the classification of distinct rice varieties. Rice images are acquired from benchmark datasets and are preprocessed using the hybrid filtering technique of the modified median and Wiener filters, which effectively eliminates the noise present in the rice images. In addition, the proposed rice variety classification model employs the CKA similarity metric, which helps to transfer the relevant knowledge between domains, reducing the impact of misaligned features and resulting in a more accurate variety identification. The experimental results of the proposed CKA-CDKT model display an accuracy of 99.93%, which is greater than that of existing approaches such as Lightweight ConvNeXt.

Keywords

Bidirectional Knowledge Transfer, Cross-modality, EfficientNet-B7, Modified Median Filter, Rice Variety Classification, Wiener Filter.

References

[2] Thi-Thu-Hong Phan, and Le Huu Bao Nguyen, “Enhancing Rice Seed Purity Recognition Accuracy based on Optimal Feature Selection,” Ecological Informatics, vol. 86, pp. 1-17, 2025.
[CrossRef] [Google Scholar] [Publisher Link]

[3] Md. Masudul Islam et al., “Rice Variety Classification using Next Generation Convolutional Networks,” The Journal of Engineering, vol. 2025, no. 1, pp. 1-11, 2025.
[CrossRef] [Google Scholar] [Publisher Link]

[4] A. Aishwarya Hegde, Pruthviraj Umesh, and Mohit P. Tahiliani, “Comparison of Neural Networks for Binary Spatial Classification of Rice Field by Studying Temporal Pattern using Dual Polarimetric SAR Measurements,” Journal of the Indian Society of Remote Sensing, vol. 52, no. 12, pp. 2867-2885, 2024.
[CrossRef] [Google Scholar] [Publisher Link]

[5] Zhiliang Kan et al., “The Rapid Non-Destructive Differentiation of Different Varieties of Rice by Fluorescence Hyperspectral Technology Combined with Machine Learning,” Molecules, vol. 29, no. 3, pp. 1-23, 2024.
[CrossRef] [Google Scholar] [Publisher Link]

[6] AmirMasoud Taheri, Hossein Ebrahimnezhad, and Mohammadhossein Sedaaghi, “Ensemble Deep Learning for High-Precision Classification of 90 Rice Seed Varieties from Hyperspectral Images,” Journal of Ambient Intelligence and Humanized Computing, vol. 15, no. 6, pp. 2883-2899, 2024.
[CrossRef] [Google Scholar] [Publisher Link]

[7] Xiao Chen et al., “Infrared Microspectroscopy and Machine Learning: A Novel Approach to Determine the Origin and Variety of Individual Rice Grains,” Agriculture Communications, vol. 2, no. 2, pp. 1-9, 2024.
[CrossRef] [Google Scholar] [Publisher Link]

[8] Komal Sharma, Ganesh Kumar Sethi, and Rajesh Kumar Bawa, “A Comparative Analysis of Deep Learning and Deep Transfer Learning Approaches for Identification of Rice Varieties,” Multimedia Tools and Applications, vol. 84, no. 10, pp. 6825-6842, 2024.
[CrossRef] [Google Scholar] [Publisher Link]

[9] K. Mahadevan, A. Punitha, and Jaganathan Suresh, “Automatic Recognition of Rice Plant Leaf Diseases Detection using Deep Neural Network with Improved Threshold Neural Network,” e-Prime-Advances in Electrical Engineering, Electronics and Energy, vol. 8, pp. 1-13, 2024.
[CrossRef] [Google Scholar] [Publisher Link]

[10] Lin Shaodan et al., “Application of UAV-based Imaging and Deep Learning in Assessment of Rice Blast Resistance,” Rice Science, vol. 30, no. 6, pp. 652-660, 2023.
[CrossRef] [Google Scholar] [Publisher Link]

[11] Chinmay Kurade et al., “An Automated Image Processing Module for Quality Evaluation of Milled Rice,” Foods, vol. 12, no. 6, pp. 1-14, 2023.
[CrossRef] [Google Scholar] [Publisher Link]

[12] Xianyu Guo et al., “Fine Classification of Rice Paddy using Multitemporal Compact Polarimetric SAR C Band Data based on Machine Learning Methods,” Frontiers of Earth Science, vol. 18, no. 1, pp. 30-43, 2024.
[CrossRef] [Google Scholar] [Publisher Link]

[13] Mengda Chai et al., “A Novel Method for Rice Identification: Coupling Raman Spectroscopy with Fourier Spectrum and Analyzing with Deep Learning,” Journal of Food Composition and Analysis, vol. 136, 2024.
[CrossRef] [Google Scholar] [Publisher Link]

[14] Amandeep Kaur, Kalpna Guleria, and Naresh Kumar Trivedi, “A Deep Learning-based Model for Biotic Rice Leaf Disease Detection,” Multimedia Tools and Applications, vol. 83, no. 36, pp. 83583-83609, 2024.
[CrossRef] [Google Scholar] [Publisher Link]

[15] Liu Yang et al., “Predictive Modeling of Rice Milling Degree for Three Typical Chinese Rice Varieties using Interpretative Machine Learning Methods,” Journal of Food Science, vol. 89, no. 10, pp. 6553-6574, 2024.
[CrossRef] [Google Scholar] [Publisher Link]

[16] Pengtao Lv et al., “An Improved Lightweight ConvNext for Rice Classification,” Alexandria Engineering Journal, vol. 112, pp. 84-97, 2025.
[CrossRef] [Google Scholar] [Publisher Link]

[17] Nabin Kumar Naik et al., “Machine Learning with Analysis-of-Variance-based Method for Identifying Rice Varieties,” Journal of Agriculture and Food Research, vol. 18, pp. 1-5, 2024.
[CrossRef] [Google Scholar] [Publisher Link]

[18] Ratnavel Rajalakshmi et al., “RiceSeedNet: Rice Seed Variety Identification using Deep Neural Network,” Journal of Agriculture and Food Research, vol. 16, pp. 1-12, 2024.
[CrossRef] [Google Scholar] [Publisher Link]

[19] Mustafa Tasci et al., “An Efficient Classification of Rice Variety with Quantized Neural Networks,” Electronics, vol. 12, no. 10, pp. 1-16, 2023.
[CrossRef] [Google Scholar] [Publisher Link]

[20] Mohammadreza Razavi et al., “Rice-ResNet: Rice Classification and Quality Detection by Transferred ResNet Deep Model,” Software Impacts, vol. 20, pp. 1-5, 2024.
[CrossRef] [Google Scholar] [Publisher Link]

[21] Haya Mesfer Alshahrani et al., “Quantum-Inspired Moth Flame Optimizer Enhanced Deep Learning for Automated Rice Variety Classification,” IEEE Access, vol. 11, pp. 125593-125600, 2023.
[CrossRef]  [Google Scholar] [Publisher Link]

[22] Xudong Li et al., “A Novel Residual Learning of Multi-Scale Feature Extraction Model for the Classification of Rice Grain Varieties,” Computers and Electronics in Agriculture, vol. 237, 2025.
[CrossRef]  [Google Scholar] [Publisher Link]

[23] Jintao Liu et al., “GMMNet: A Precise Classification Model for Rice Grains During Rice Processing,” Expert Systems with Applications, vol. 287, 2025.
[CrossRef]  [Google Scholar] [Publisher Link]

[24] Murat Koklu, Rice Image Dataset, 2025. [Online]. Available: https://www.kaggle.com/datasets/muratkokludataset/rice-image-dataset

[25] Lei Fan et al., “GrainSpace: A Large-Scale Dataset for Fine-Grained and Domain-Adaptive Recognition of Cereal Grains,” 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), New Orleans, LA, USA, pp. 21084-21093, 2022.
[CrossRef]  [Google Scholar] [Publisher Link]