Deep Sliding Shapes for Amodal 3D Object Detection in RGB-D Images
Authors:
- Shuran SongSSShuran Song
- Jianxiong XiaoJXJianxiong Xiao
- 1 June 2016
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 808-816
Abstract
We focus on the task of amodal 3D object detection in RGB-D images, which aims to produce a 3D bounding box of an object in metric form at its full extent. We introduce Deep Sliding Shapes, a 3D ConvNet formulation that takes a 3D volumetric scene from a RGB-D image as input and outputs 3D object bounding boxes. In our approach, we propose the first 3D Region Proposal Network (RPN) to learn objectness from geometric shapes and the first joint Object Recognition Network (ORN) to extract geometric features in 3D and color features in 2D. In particular, we handle objects of various sizes by training an amodal RPN at two different scales and an ORN to regress 3D bounding boxes. Experiments show that our algorithm outperforms the state-of-the-art by 13.8 in mAP and is 200× faster than the original Sliding Shapes.Keywords
All Related Versions
This publication has 17 references indexed in Scilit:
- Structured Prediction of Unobserved Voxels from a Single Depth Image2016
- Amodal Completion and Size Constancy in Natural Scenes2015
- VoxNet: A 3D Convolutional Neural Network for real-time object recognition2015
- Analysis and synthesis of 3D shape families via deep‐learned generative models of surfacesComputer Graphics Forum, 2015
- Aligning 3D models to RGB-D images of cluttered scenes2015
- 3D deep shape descriptor2015
- Unsupervised feature learning for 3D scene labeling2014
- Learning hierarchical sparse features for RGB-(D) object recognitionThe International Journal of Robotics Research, 2014
- Robust real-time visual odometry for dense RGB-D mapping2013
- Selective Search for Object RecognitionInternational Journal of Computer Vision, 2013