3D Anthropometric Algorithm from A Single Viewpoint RGB-D Camera

The technology of human body parameter measurement makes the intelligent life such as 3D virtual fitting room and automatic customized service become a reality which have high requirements for the simplicity and flexibility. In this paper, a method of human body parameter measurement based on RGB-D camera and single viewpoint is proposed, to construct a system that is both convenient and flexible. Through a direct analysis of the data on the human body surface of front, back and side, the parameters can be measured including waist circumference and hip circumference etc., which avoids the measurement error caused by the shaking of human body. Based on the fusion of 2D data and 3D data, an efficient human profile extraction algorithm is designed for the measurement to be performed under general circumstances. The crucial sparse points of human body parts for modeling each part are extracted to improve the measurement accuracy. Compared to other works based on a single viewpoint, the method proposed by us achieves similar measurement accuracy, while significantly reducing the complexity of the equipment and measurement process.