Bojan Pepik, Michael Stark, Peter Gehler, Tobias Ritschel, Bernt Schiele
TL;DR本文提出了一种 3D 物体识别方法,利用 2D-3D 物体提升技术,逐步将对象检测输出增加到对象视点,关键点和 3D 形状估计,并在 Pascal3D + 数据集上同时实现 2D 边界框和视点估计的最佳方法。
Abstract
object class detection has been a synonym for 2D bounding box localization
for the longest time, fueled by the success of powerful statistical learning
techniques, combined with robust image representations. Only recently, there
has been a growing interest in revisiting the promise of