Deep Polarization Cues for Transparent Object Segmentation
Authors:
- 1 June 2020
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 8599-8608
Abstract
Segmentation of transparent objects is a hard, open problem in computer vision. Transparent objects lack texture of their own, adopting instead the texture of scene background. This paper reframes the problem of transparent object segmentation into the realm of light polarization, i.e., the rotation of light waves. We use a polarization camera to capture multi-modal imagery and couple this with a unique deep learning backbone for processing polarization input data. Our method achieves instance segmentation on cluttered, transparent objects in various scene and background conditions, demonstrating an improvement over traditional image-based approaches. As an application we use this for robotic bin picking of transparent objects.Keywords
This publication has 36 references indexed in Scilit:
- Choosing smartly: Adaptive multimodal fusion for object detection in changing environments2016
- TransCut: Transparent Object Segmentation from a Light-Field Image2015
- Polarized 3D: High-Quality Depth Sensing with Polarization Cues2015
- Reconstruction of transparent objects in unstructured scenes with a depth camera2013
- Active Polarization DescatteringIEEE Transactions on Pattern Analysis and Machine Intelligence, 2008
- Polarization and Phase-Shifting for 3D Scanning of Translucent Objects2007
- Recovery of surface orientation from diffuse polarizationIEEE Transactions on Image Processing, 2006
- Polarization imaging applied to 3D reconstruction of specular metallic surfaces2005
- Transparent surface modeling from a pair of polarization imagesIEEE Transactions on Pattern Analysis and Machine Intelligence, 2004
- Measurement of surface orientations of transparent objects using polarization in highlightSystems and Computers in Japan, 2001