Image Segmentation by MAP-ML Estimations

Shifeng Chen, Liangliang Cao, Yueming Wang, Jianzhuang Liu, and Xiaoou Tang


Abstract

AbstractImage segmentation plays an important role in computer vision and image analysis. In this paper, image segmentation is formulated as a labeling problem under a probability maximization framework. To estimate the label configuration, an iterative optimization scheme is proposed to alternately carry out the maximum a posteriori (MAP) estimation and the maximum likelihood (ML) estimation. The MAP estimation problem is modeled with Markov random fields (MRFs) and a graph cut algorithm is used to find the solution to the MAP estimation. The ML estimation is achieved by computing the means of region features in a Gaussian model. Our algorithm can automatically segment an image into regions with relevant textures or colors without the need to know the number of regions in advance. Its results match image edges very well and are consistent with human perception. Comparing to six state-of-the-art algorithms, extensive experiments have shown that our algorithm performs the best.


 

 

AVERAGE VALUES OF PRI AND VOI FOR THE SEVEN ALGORITHMS ON THE IMAGES IN THE BERKELEY SEGMENTATION DATABASE.

 

Human

Our

NC

BW

MS

EG

LDC

SWC

PRI

VoI

0.8961

0.9219

0.7967

1.9307

0.7226

2.9247

0.7138

2.6295

0.7822

3.8152

0.7877

2.8350

0.7529

2.0288

0.7644

3.0266

 

 

 

 

visual1.png

Figure 1. Segmentation results on landscape images.

 

visual6.png

Figure 2. Segmentation results on felid images.

 

PRI_Final.png

Figure 3. PRI values achieved on individual images by different algorithms in Berkeley database. The values are plotted in increasing order.

 

VoI_Final.png

Figure 4. VoI values achieved on individual images by different algorithms in Berkeley database. The values are plotted in increasing order.

 

 

Ÿ  References:

Chen, L. Cao, Y. Wang, J. Liu, and X. Tang, Image Segmentation by MAP-ML Estimations, IEEE Trans. on Image Processing (TIP), vol. 19, no. 9, pp. 2254-2264, 2010. [pdf]

S. Chen, L. Cao, J. Liu, and X. Tang, Iterative MAP and ML Estimations for Image Segmentation, Proc. IEEE Conf. Computer Vision and Pattern Recognition (CVPR), 2007. [pdf]