Adaptive region growing based on boundary measures

Authors:
Yu Qiao;Jie Yang
Affiliations:
Institute of Image Processing and Pattern Recognition, Shanghai Jiao Tong University, Shanghai, P.R. China;Institute of Image Processing and Pattern Recognition, Shanghai Jiao Tong University, Shanghai, P.R. China
Venue:
ICONIP'11 Proceedings of the 18th international conference on Neural Information Processing - Volume Part I
Year:
2011

Citing 6
Cited 0

Minimum error thresholding

Pattern Recognition
Shape Modeling with Front Propagation: A Level Set Approach

IEEE Transactions on Pattern Analysis and Machine Intelligence
Thresholding based on variance and intensity contrast

Pattern Recognition
Arc-based evaluation and detection of ellipses

Pattern Recognition
Snakes, shapes, and gradient vector flow

IEEE Transactions on Image Processing
Region growing: a new approach

IEEE Transactions on Image Processing

Quantified Score

Hi-index	0.00

Visualization

Abstract

Weak boundary contrast, inhomogeneous background and overlapped intensity distributions of the object and background are main causes that may lead to failure of boundary detection for many image segmentation methods. An adaptive region growing method based on multiple boundary measures is presented. It consists of region expansion and boundary selection processes. During the region expansion process the region grows from a seed point. The background points adjacent to the current region are examined with local boundary measures. The region is expanded by iteratively growing the most qualified points. In the boundary selection process, the object boundary is determined with the global boundary measure that evaluates the boundary completeness. Experimental results demonstrate that our algorithm is robust against weak boundary contrast, inhomogeneous background and overlapped intensity distributions.