Kinect cloud normals: towards surface orientation estimation

Since the early days of Computer Vision, we have explored what is possible in the realm of ‘Scene Understanding’. The advent of consumer-grade RGBD cameras has broadened the possibilities within this realm. The data they provide is able to serve as ground truth information or training data for a class of algorithms, which would otherwise be extremely difficult, if not impossible, to train. This thesis serves the purpose of gathering data from such a source, specifically, it demonstrates how to collect a dual pair of depth and RGB images of a multitude of scenes and an approach to determine surface normals from these images. The goal of this endeavor is to provide a dataset of RGB images and surface normal estimates for each image so that the latter may serve as the ground truth for both training and evaluation of algorithms estimating surface normals from the RGB image alone.

Name thesis:degree_name

M.S.

Level thesis:degree_level

Thesis

Discipline thesis:degree_discipline

Computer Science

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2011

Author dc:creator

• Rungta, Pratik S.

Contributors dc:contributor

• Hoiem, Derek W.

• microsoft kinect
• dataset
• point cloud
• normal estimation

Statement dc:rights

• Copyright 2011 Pratik S. Rungta

Language dc:language

en

Handle dc:identifier

http://hdl.handle.net/2142/26108

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/26108