Localizing object parts in 3D from a single image

Yin S, Zhou B, Yang M J, et al

Sci China Inf Sci, 2019, 62(7): 074101

Object localization in 3D from 2D images is an important computer vision problem that enables modern robotic vision systems to interact properly with the objects present in the real world. Owing to its significance, techniques for recovering the 6-DOF pose and dimensions of objects from images has received increasing attention in recent years. However, to the best of our knowledge, the existing approaches all focus on object-level inference instead of analyzing object parts. Most of the time we expect a robot to interact with a semantic part of an object (e.g., holding the bar of a cup), part-level 3D object localization is considerably preferred in practical applications.
PDF Slides

Demo