Image Enhancement for Rotation Sushi Plates Detection in Smart Multi-Platform Billing System

Conveyor belt sushi restaurants have become worldwide popular. The customers are drawn to the low prices and accessibility of the rotation sushi to save their dining time. These small dishes are typically served on different kinds of plates depending on price. Counting the number of each type of plates can then be computed for bill payment. An AI-based approach for sushi plate detection to overcome the time-consuming plate count is introduced in this work. A smart billing system is developed on multi-platform mobile application. The user can take a photo of stacks of consumed dishes as the input image to the YOLOv4 object detector for classifying the different colored plates. This paper focuses on parts of image enhancement as the system can work on iOS and Android systems. The techniques of color transfer with LAB color space and CLAHE are applied for image quality improvement. The experimental results reported the achievement of higher confidence and the successful detection of the objects from out of color space images.