Computer Science ›› 2015, Vol. 42 ›› Issue (8): 52-55.

Previous Articles     Next Articles

Unsupervised Image Segmentation Based on Saliency Detection

ZHOU Jing-bo, REN Yong-feng and YAN Yun-yang   

  • Online:2018-11-14 Published:2018-11-14

Abstract: Interactive image segmentation needs the user interactions which increases the time complexity and the user’sburden.We proposed an unsupervised image segmentation algorithm based on visual saliency.First,mean shift (MS) algorithm is used to obtain initial segmentation without overlapping.The regions generated by MS are represented by a region adjacency graph (RAG) and an edge exists only if two regions are adjacent.Second,the color dissimilarity and texture consistency between the regions are computed,which are adjacent,as the weight of the edge in our RAG.Then,the proposed algorithm defines the saliency index (SI) according to the color and spatial information of each region gene-rated by MS algorithm.The region with maximal SI is defined as the seed of object,and the region with minimal SI in the boundary is defined as the seed of background.Finally,region merging is performed according to the strategy of maximize similarity around the seed of object and background.The results show that the proposed algorithm obtains better segment results without any interactive information and avoids oversegmentation compared with other unsupervised image segmentation.

Key words: Unsupervised image segmentation,Saliency detection,Mean shift

[1] Boykov Y,Kolmogorov V.Computing geodesics and minimalsurfaces via graph cuts[C]∥Proceedings of Seventh IEEE International Conference on Computer Vision (ICCV).Nice,France,2003,1:26-33
[2] Rother C,Kolmogorov V,Blake A.Grab cut interactive fore-ground extraction using iterated graph cuts[J].ACM Transactions on Graphics,2004,23(3):309-314
[3] Cheng Y Z.Mean shift,mode seeking,and clustering[J].IEEETransactions on Pattern Analysis and Machine Intelligence,1995,17(8):790-799
[4] Nock R,Nielsen F.Statistic region merging[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2004,26(11):1452-1458
[5] Calderero F,Marques F.Region merging techniques using information theory statistical measures[J].IEEE Transactions on Image Processing,2010,19(6):1567-1586
[6] Calderero F,Marques F.General region merging approachesbased on information theory statistical measures[C]∥Procee-dings of IEEE International Conference in Image Processing.San Diego,2008:3016-3019
[7] Wan T,Canagarajah N,Achim A.Statistical multiscale imagesegmentation via alpha-stable modeling[C]∥Proceedings of IEEE International Conference in Image Processing.Texas,2007:357-360
[8] Haris K,Estradiadis S N,Maglaveras N,et al.Hybrid imagesegmentation using watersheds and fast region merging[J].IEEE Transactions on Image Processing,1998,7(12):1684-1699
[9] Liu H,Guo Q,Xu M,et al.Fast image segmentation using region merging with a k-nearest neighbor graph[C]∥Proceedings of IEEE International Conference Cybernetics and Intelligent Systems.Chengdu,2008:179-184
[10] Shu Y,Bilodeau G A,Cheriet F.Segmentation of laparoscopic images:Integrating graph-based segmentation and multistage region merging[C]∥Proceedings of the 2nd Canadian Conference on Computer and Robot Vision.Regina,2005:429-436
[11] Moscheni F,Bhattacharjee S,Kunt M.Spatio-temporal segmentation based on region merging[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,1998,20(9):897-915
[12] Ning J,Zhang L,Zhang D,et al.Interactive image segmentation by maximal similarity based region merging[J].Pattern Recognition,2010,43(2):445-456
[13] Peng B,Zhang L,Zhang D.Automatic image segmentation by dynamic region merging[J].IEEE Transactions on Image Processing,2011,20(12):3592-3605
[14] Cheng M,et al.Global contrast based salient region detection[C]∥Proceedings of IEEE Conference on Computer Vision and Pattern Recognition (CVPR).Providence,2011:409-416
[15] Felzenszwalb P F,Huttenlocher D P.Efficient graph based image segmentation[J].International Journal of Computer Vision,2004,59(2):167-181
[16] Martin D,Fowlkes C,Tal D,et al.A database of human segmented natural images and its application to evaluating segmentation algorithms and measuring ecological statistics[C]∥Proceedings of IEEE Conference on Computer Vision and Pattern Recognition (CVPR).Kauai,2001:416-423

No related articles found!
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!