With the increase of expectation for higher quality of life, consumers have higher demands for quality food. Food authentication is the technical means of ensuring food quality, which is intended to confirm "food is what it says on the tin". A popular approach to food authentication is based on spectroscopy analysis which has been widely used for identifying and quantifying the chemical compositions of an object. Such approach is nondestructive and effective but expensive. This paper presents an image-based approach to food authentication using image processing and pattern recognition techniques. In this approach, flashlight is used to illuminate apples and images of the illuminated apples are captured by a smartphone. These images are represented by LBP (local binary pattern) image descriptors. Data pre-processing algorithms are used to prepare the image representations, and pattern recognition algorithms including k-nearest neighbors and support vector machine are used for classification. This approach is evaluated in a food differentiation (to separate organic apples from non-organic ones) experiment using a reasonable collection of apple samples, resulting in the highest classification accuracy of 86.7%. It is shown that this low-cost approach has potential to lead to a viable solution to empower consumers in food authentication.
展开▼
