Comparative Analysis of CNN Based Deep Learning Models for Automated Rice Varity Classification

Rice variety classification is crucial for quality control, yield estimation, andmarket value assessment. With advancements in deep learning, automated classification usingconvolutional models has gained importance. This study compares various deep learningmodels for rice variety classification, evaluating accuracy, efficiency, and robustness. A datasetof rice grain images is used to train and test models, analyzing key performance indicators likeaccuracy, precision, recall, and F1-score. The impact of data augmentation and transferlearning is also explored. Computational efficiency and training time are assessed for practicalapplicability. Results show deeper architectures perform better than shallower ones. The studyhighlights the potential of deep learning in automating classification for agriculture. Futurework will focus on hybrid models and advanced feature extraction.