Towards food safety: Identifying fake rice using SVM and contour analysis

In India, rice is the most widely consumed staple food compared to wheat and other millets. However, the emergence of fake or plastic rice, visually indistinguishable from real rice, has raised serious food safety concerns. Consumption of such adulterated rice, made from toxic materials like polystyrene, poses significant health risks. Hence, differentiating counterfeit rice from genuine varieties has become essential. The growing prevalence of synthetic rice poses significant threats to food safety and public health, necessitating the development of reliable and scalable detection systems. This study presents a machine learning-based approach for the identification of fake rice using Support Vector Machine (SVM) and contour analysis. The proposed system leverages image-based features, including morphological, color, and contour-based characteristics, to distinguish real rice grains from spurious ones. In its initial version, referred to as Prediction 1.0, the model employs traditional image processing techniques for feature extraction and uses SVM-support vector machine classifier to perform binary classification. Preprocessing steps such as resizing, orientation correction, and data augmentation enhanced the model’s ability to generalize across different rice varieties and imaging conditions. The system achieved a high classification confidence of 97.3%, demonstrating its robustness and reliability in detecting fake rice. Visual results from the model underscore the effectiveness of contour-based features in discriminating between genuine and counterfeit grains. The findings highlight the strong potential of combining contour analysis with machine learning techniques to build cost-effective, scalable, and automated solutions for quality control in the rice industry.