A review on Applications of Image Processing Methods on Food Product's Quality Control- Part B: Model-Based and Transform-Based Texture Processing Methods

Consumers increased expectations of high quality food products as well as stringent regulations has increased the need for an accurate and fast method for quality assessment and control of the products in food industries. Machine vision with the aid of various image processing methods has been introduced as an objective, automate, and non-destructive approach capable for food quality control. Texture as one of the most important image's features has been used extensively in food quality monitoring applications. Generally, quantitative texture assessment methods are divided into four groups: statistical, structural, model-based, and transform-based methods. In the first part of this research (part A), the principles of the statistical image texture processing methods were reviewed. The subject of the second part (part B) is model-based and transform-based texture processing methods. Model-based methods quantify image texture by considering a mathematical equation explaining relation between pixel's intensities, while transform-based methods extract textural features of an image which is transformed by mathematical transforms. In the present paper, at first mechanisms of these groups of texture evaluation methods have been presented. Then, examples of recent studies related to employments of image texture in quality control of food products have been reviewed. The results of the previous studies indicate that after statistical methods, model-based and transform-based methods are the most accurate and popular texture evaluation methods in food industries.