Camera2Video

This paper presents a low-cost sensor system for food authentication based on computer vision and pattern recognition. The sensor system uses a smartphone to record a video of a sample under gradient coloured illumination and transforms the video into data vector for analysis. The sensor system is evaluated on the tasks of detecting olive oil adulteration and milk fat content, achieving 100% test accuracies. Since the sensor system is built in a smartphone without any additional hardware, it has the potential to serve as a simple and effective solution for food authentication.