High Accuracy Mobile 3D Scanning Using Structured Laser Beam Patterning

Abstract

A person's diet affects their weight, lifespan, and chances of occurrence of such medical problems as diabetes, obesity, and cancer, to name a few. Enhanced dietary assessment techniques are critical for epidemiological studies that target diet-related problems. Currently, nutritional research is considerably hindered by the low accuracy in estimating individual dietary intake, and, more specifically, portion size. Dietary assessment plays an increasingly significant role in modern medical research. While inadequate diets can increase the risk of diabetes, obesity, and cancer. It is essential to design accurate, cost-effective tools that measure dietary data to advance nutritional research. This dissertation describes a low-cost and efficient method of calculating nutritional information by using 3D reconstruction and image processing. This system is called Dietary Data Recorder System (DDRS), which consists of a smartphone, a laser projector, and the main algorithm, which extracts the data from the DDRS for volume and nutritional calculations. The DDRS software consists of four main algorithms: Automatic Laser Detection algorithm, Segmentation algorithm, 3D Mapping algorithm, and Nutritional Estimation algorithm. In particular, this dissertation focuses on the first two functions: Automatic Laser Detection algorithm and Segmentation algorithm.