A mobile structured light system for food volume estimation

Abstract

Dietary intake provides valuable information for understanding the relationship between diet and health. Valid measurement of food volume is one of the most challenging problems in nutrient studies. Computer vision techniques have been used for better evaluation of dietary intake. In this paper, we propose a mobile structured light system (SLS) to measure daily dietary intake. Our system consists of a mobile smartphone and a laser attachment. The laser attachment projects grids on the food and the grid video is acquired by the camera on the phone. We apply our algorithm to identify the laser grids in the video frames, localize the intersection points, and generate depth maps for the corresponding images. We applied our algorithm to simulated data as well as real video frames from planar surfaces and foods. The results from SLS will be further combined with results of video segmentation, motion estimation, and other 3D reconstruction methods.