Active contours textural and inhomogeneous object extraction. (July 2016)

• Record Type: Journal Article
• Authors: Mabood, Lutful; Ali, Haider; Badshah, Noor; Chen, Ke; Khan, Gulzar Ali
• Is Part Of: Pattern recognition. Volume 55(2016:Jul.)
• Page Start: 87
• Page End: 99
• Publication Date: 2016-07
• Subjects: Image selective segmentation -- Level set -- Functional minimization -- Numerical method; Pattern perception -- Periodicals; Perception des structures -- Périodiques; Patroonherkenning; 006.4
• DOI: 10.1016/j.patcog.2016.01.021
• Languages: English
• ISSNs: 0031-3203

Abstract

Abstract: A new selective segmentation active contour model is proposed in this paper that embeds an enhanced image information. By utilizing the average image of channels (AIC), which handles texture and noise, our model is capable to selectively segment and capture objects with nonuniform features. Moreover, the AIC is fitted with linear functions which are updated regularly to accurately guide the level set function to handle nonconstant intensities. Furthermore, we employ prior information in terms of geometrical constraints which work in alliance with image information to capture objects with intensity inhomogeneity. Experiments show that the proposed method achieves better results than the latest selective segmentation models. In addition, our approach maintains the performance on some hard real and synthetic color images. Abstract : Highlights: A new selective segmentation active contour model is proposed. The proposed model is based on the concept of average image of channels. The proposed model is capable to selectively segment noisy/textural objects of interest.