Interactive Image Segmentation Using Adaptive Constraint Propagation

ABSTRACT:

In this paper, we propose interactive image segmentation using adaptive constraint propagation (ACP), called ACP Cut. In interactive image segmentation, the interactive inputs provided by users play an important role in guiding image segmentation. However, these simple inputs often cause bias that leads to failure in preserving object boundaries. To effectively use this limited interactive information, we employ ACP for semi supervised kernel matrix learning which adaptively propagates the interactive information into the whole image, while successfully keeping the original data coherence. Moreover, ACP Cut adopts seed propagation to achieve discriminative structure learning and reduce the computational complexity. Experimental results demonstrate that the ACP Cut extracts foreground objects successfully from the background and outperforms the state-of-the-art methods for interactive image segmentation in terms of both effectiveness and efficiency.

PROJECT OUTPUT VIDEO:

EXISTING SYSTEM:

• In, KP Cut performed interactive image segmentation over a data structure learned from kernel propagation which propagated the user’s interactive information through the whole image. Although the constraint propagation methods have produced a good global discriminative structure, there is still room for improvement in effectiveness and efficiency when applying them to inter-active image segmentation.

• For instance, the hard constraints forced samples, i.e. pixels or superpixels in image segmentation, with must-link constraints to be overlapped and with cannot-link constraints to be orthogonal after kernel mapping, which have largely distorted the data discriminative structure in the learned kernel.

DISADVANTAGES OF EXISTING SYSTEM:

• Complexity is high.

• They lacked the ability to provide all objects in an image, especially when the desired objects contain multiple basic features. Thus, it is hard for automatic image segmentation to fully meet the user’s requirements and preferences.

PROPOSED SYSTEM:

• In this paper, we propose ACP Cut to propagate characteristics of the user’s interactive information into the whole image successfully while maintaining global data coherence, as well as learn a global image discriminative structure for interactive image segmentation. ACP Cut adopts adaptive constraints instead of traditional hard constraints to learn a global discriminative structure.

• In our previous work, we have provided the original ACP formulation for semi supervised kernel matrix learning (SS-KML). However, it has high computational complexity because its computational cost rises very rapidly as the number of samples is increased. Thus, it is not suitable for practical applications which need fast processing.

• To deal with this problem, we employ seed propagation in the ACP learning procedure to remarkably improve the computational complexity of ACP. We apply ACP with seed propagation to interactive image segmentation, called ACP Cut, and verify its effectiveness and efficiency in image segmentation.

• First, we extract features from superpixels obtained by mean-shift segmentation in an image. Then, we generate pairwise constraints from the user’s interactive information. Next, we perform ACP with seed propagation on both features and pairwise constraints to learn a global discriminative structure in an image.

• Finally, we assign a label of foreground/background to each superpixel based on the learned discriminative structure, thus segmenting fore-ground objects from background.

ADVANTAGES OF PROPOSED SYSTEM:

• Seed propagation in ACP Cut recursively learns the discriminative structure and greatly reduces the computational complexity. Experimental results and their corresponding evaluations indicate that ACP Cutout-performs the other methods for interactive image segmentation in terms of both effectiveness and efficiency.

• Consequently, ACP Cut enhances the global discrimination of foreground and background by learning with adaptive constraints and seed propagation and achieves good segmentation results.

SYSTEM REQUIREMENTS:

HARDWARE REQUIREMENTS:

• System           : Pentium Dual Core.

• Hard Disk      : 120 GB.

• Monitor          : 15’’ LED

• Input Devices : Keyboard, Mouse

• Ram                : 1GB.

SOFTWARE REQUIREMENTS:

• Operating system : Windows 7.

• Coding Language : MATLAB

• Tool                      : MATLAB R2013A

REFERENCE:

Meng Jian and Cheolkon Jung, Member, IEEE, “Interactive Image Segmentation Using Adaptive Constraint Propagation”, IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 25, NO. 3, MARCH 2016.