The interaction of Integral Geometry and Machine Learning
(Research Center for Healthcare Industry Innovation, National Taipei University of Nursing and Health Sciences)
By combining the classic mathematical equations and the modern machine learning techniques, we develop an automatic detection system of surface areas and volumes for geometric objects which are not necessarily symmetric and ellipsoidal. We were asked by a ham producer company to find the surface areas and volumes for different sizes of hams. Instead of multiple cameras, we use Mask R-CNN technique with only one individual camera to measure the geometric quantities. The experimental results demonstrate that the algorithm proposed is robust, and the estimation of surface areas and volumes reach up to 95% for a variety of hams.
日 期：108年9月25日(星期三) 16:10~17:00