Computer Science · Computer Vision and Pattern Recognition
Predicting Human Body Measurements from a Front Image
Tzvetomir Vassilev, Metodi Dimitrov
See on the mapAbstract
Quite a few scientists have recently performed research on modeling human body and shape. The current paper describes a solution to predict human body girth measurement and weight from a front image taken with a modern smartphone camera. In order to achieve this, a set of 3D body scans and manually taken body measurements have been used as an input to build several statistical PCA models. Seven linear measurements are extracted from the human front image. The body height is entered by the user and together with the seven measurements used as an input to the statistical models. These models are then applied to predict body measures and weight. The results of the experiments are promising but more tests with pilot users have to be performed to get a stable solution.