Computer Vision and Machine Learning Analysis of Commercial Rice Grains: A Potential Digital Approach for Consumer Perception Studies

Abstract

Rice quality assessment is essential for meeting high-quality standards and consumer demands. However, challenges remain in developing cost-effective and rapid techniques to assess commercial rice grain quality traits. This paper presents the application of computer vision (CV) and machine learning (ML) to classify commercial rice samples based on dimensionless morphometric parameters and color parameters extracted using CV algorithms from digital images obtained from a smartphone camera. The artificial neural network (ANN) model was developed using nine morpho-colorimetric parameters to classify rice samples into 15 commercial rice types. Furthermore, the ANN models were deployed and evaluated on a different imaging system to simulate their practical applications under different conditions. Results showed that the best classification accuracy was obtained using the Bayesian Regularization (BR) algorithm of the ANN with ten hidden neurons at 91.6% (MSE = <0.01) and 88.5% (MSE = 0.01) for the training and testing stages, respectively, with an overall accuracy of 90.7% (Model 2). Deployment also showed high accuracy (93.9%) in the classification of the rice samples. The adoption by the industry of rapid, reliable, and accurate methods, such as those presented here, may allow the incorporation of different morpho-colorimetric traits in rice with consumer perception studies.

Keywords: artificial neural networks; morpho-colorimetry; object of interest; photogrammetry; smartphone.

Figure 1

Experimental setup for image acquisition consisting of (1) a lightbox; (2) 70 pieces of two LED light strips; (3) a platform; (4) cardboard to place the grain for image acquisition; (5) rice samples; (6) top opening of the lightbox used to acquire the images using a smartphone. LED = light-emitting diode.

Figure 2

Examples of the 15 commercial rice sample images used in the study. The details of the rice samples based on the labeled class ID correspond to the list in Table 1. Rice sample images shown in the figure were cropped for presentation purposes only.

Figure 3

Flow diagram of the method used to extract the morpho-colorimetric features from the rice grain image.

Figure 4

Diagram of a neural network model (Model 1) of the Bayesian Regularization algorithm with seven hidden neurons and sigmoid function showing nine inputs of morpho-colorimetric parameters and 15 outputs of commercial rice grains. The abbreviations for the morpho-colorimetric parameters (inputs) and commercial rice grains (outputs) are shown in Table 1 and Table 2. w = weight; b = bias.

Figure 5

Diagram of a neural network model (Model 2) of the Bayesian Regularization algorithm with seven hidden neurons and sigmoid function showing nine inputs of morpho-colorimetric parameters and 15 outputs of commercial rice grains. The abbreviation for morpho-colorimetric parameters (inputs) and commercial rice grains (outputs) are shown in Table 1 and Table 2. w = weight; b = bias.

Figure 6

Flow diagram of the rice classification using computer vision and machine learning analysis. ANN = artificial neural network.

Figure 7

The multivariate data analysis for (a) principal component analysis (PCA) biplot for morpho-colorimetric parameters of 17 commercial rice types, where PC1 = principal component one, PC2 = principal component two; and (b) cluster analysis of the commercial rice samples based on morpho-colorimetric parameters. The abbreviations for morpho-colorimetric parameters and rice samples are shown in Table 1 and Table 2.

Figure 8

The receiver operating characteristic (ROC) curve of the artificial neural network models to classify 15 commercial rice samples using morpho-colorimetric parameters as inputs for (a) Model 1 and (b) Model 2. The abbreviations for rice samples are shown in Table 1.