Cell Phone Image-Based Persian Rice Detection and Classification Using Deep Learning Techniques

Mahmood Saeedi kelishami,Amin Saeidi Kelishami,Sajjad Saeedi Kelishami

2024-04-21

Abstract:This study introduces an innovative approach to classifying various types of Persian rice using image-based deep learning techniques, highlighting the practical application of everyday technology in food categorization. Recognizing the diversity of Persian rice and its culinary significance, we leveraged the capabilities of convolutional neural networks (CNNs), specifically by fine-tuning a ResNet model for accurate identification of different rice varieties and employing a U-Net architecture for precise segmentation of rice grains in bulk images. This dual-methodology framework allows for both individual grain classification and comprehensive analysis of bulk rice samples, addressing two crucial aspects of rice quality assessment. Utilizing images captured with consumer-grade cell phones reflects a realistic scenario in which individuals can leverage this technology for assistance with grocery shopping and meal preparation. The dataset, comprising various rice types photographed under natural conditions without professional lighting or equipment, presents a challenging yet practical classification problem. Our findings demonstrate the feasibility of using non-professional images for food classification and the potential of deep learning models, like ResNet and U-Net, to adapt to the nuances of everyday objects and textures. This study contributes to the field by providing insights into the applicability of image-based deep learning in daily life, specifically for enhancing consumer experiences and knowledge in food selection. Furthermore, it opens avenues for extending this approach to other food categories and practical applications, emphasizing the role of accessible technology in bridging the gap between sophisticated computational methods and everyday tasks.

Computer Vision and Pattern Recognition,Artificial Intelligence

What problem does this paper attempt to address?

The main problem this paper attempts to address is the classification and identification of different varieties of Persian rice using mobile image-based deep learning techniques. Specifically, the study focuses on two key aspects:

1. **Single Grain Classification**: Accurately classifying individual rice grain images into seven different Persian rice varieties (Ali Kazemi, Anbar Boo (Ghareeb), Hashemi, Khazar, Sadree Dom Siahe, Sadree Dom Zard, and Shirodi). This task requires precise recognition of subtle differences between different rice grains, which are often difficult to distinguish with the naked eye.

2. **Batch Prediction**: Analyzing images containing mixed rice grains and estimating the percentage composition of each type of rice grain in the sample. This task not only requires identifying individual rice grains in the image but also segmenting and quantifying each type of rice grain to provide a comprehensive compositional analysis of the mixed sample.

Traditional classification methods rely on manual inspection, which suffers from slow speed, low accuracy, and poor reliability, especially in the absence of extensive experience. Therefore, this study aims to develop an efficient, accurate, and user-friendly automated system that utilizes deep learning techniques (such as Convolutional Neural Networks (CNN) and U-Net) and everyday mobile phone cameras to achieve the classification and quality assessment of Persian rice. This not only helps improve the accuracy of agricultural production and consumer choice but also promotes the enhancement of food quality and market transparency.