Title :
Data-driven nonlinear diffusion for object segmentation
Author :
Xu, Li-Qun ; Izquierdo, Ebroul
Author_Institution :
BT Adv. Commun. Res., Ipswich, UK
Abstract :
We propose a novel technique for effective object segmentation, which is based on the combination of image simplification via data-driven nonlinear diffusion and subsequent efficient segmentation of the simplified image. In particular, the data, taking the form of a disparity field from a stereo analysis, has been used to modulate the diffusion process. The strength of this strategy consists of an ability to smooth considerably the details of the imaging scene within the objects´ boundaries while inhibiting the diffusion across the boundaries, preserving and even enhancing the object borders. As such, from the simplified image, a simple but efficient histogram-based thresholding and labeling technique can be used to extract precisely an object boundary in its entirety
Keywords :
feature extraction; image segmentation; smoothing methods; stereo image processing; data-driven nonlinear diffusion; diffusion process modulation; disparity field; efficient histogram-based thresholding; efficient labeling technique; image segmentation; image simplification; imaging scene detail smoothing; object borders enhancement; object boundaries; object boundary extraction; object segmentation; stereo analysis; Diffusion processes; Filtering; Image color analysis; Image edge detection; Image motion analysis; Image segmentation; Image sequence analysis; Image texture analysis; Layout; Object segmentation;
Conference_Titel :
Image Processing, 2000. Proceedings. 2000 International Conference on
Conference_Location :
Vancouver, BC
Print_ISBN :
0-7803-6297-7
DOI :
10.1109/ICIP.2000.900959