Virtual trying-on experience

摘要

Virtual trying on experience of user trying on item by combining 3D model of item to be tried on, 3D generic model of human, and images of user. Images extracted from plurality of video frames at different orientations are used to determine user movement. Model origins identified in common three-dimensional coordinate space and aligned with origin reference points in images. Images captured of user rotating their head may be combined with generic 3D model of head on to which 3D models of pair of spectacles are superimposed using identified attachment points (e.g. bridge of nose, which may be rotated or translated by user for re-alignment). Maximum rotation angles in first and second directions (left and right) may be determined during video capture and divided into number of set angles. Generic male and female 3D models may be user-selected. Only a subset of captured video frames may be processed. The method may be in smartphone app for use when optician shopping for new glasses; camera captures video of user rotating their head and images overlaid on 3D models of generic head and glasses from database of stored models. User can see what would look like wearing different types of spectacles.