Title :
Segmenting scenes with motion using Ambient Illumination Invariant Imaging
Author :
Rutgers, Andrew U. ; Lawrence, Peter D. ; Hall, Robert A.
Author_Institution :
Dept. of Electr. & Comput. Eng., Univ. of British Columbia, Vancouver, BC, Canada
Abstract :
Segmentation is a key step in many machine vision techniques, which is complicated by shadows or poor illumination. Linear Ratio Based Estimation is proposed as a technique using two cameras and a flash for Ambient Illumination Invariant Imaging to create images of a scene that appear to be taken with only flash illumination, even in the presence of strong ambient illumination. Segmentation was used as a demonstration application and showed a 72% improvement compared to previous techniques.
Keywords :
computer vision; image motion analysis; image segmentation; ambient illumination invariant imaging; flash illumination; linear ratio based estimation; machine vision techniques; scene segmentation; Accuracy; Ash; Cameras; Image segmentation; Lighting; Manuals; Object segmentation; illumination invariance; segmentation; shadow removal;
Conference_Titel :
Imaging Systems and Techniques (IST), 2011 IEEE International Conference on
Conference_Location :
Penang
Print_ISBN :
978-1-61284-894-5
DOI :
10.1109/IST.2011.5962207
