Abstract:The segmentation of intensity inhomogeneity images is always a challenging problem. There are two kinds of intensity inhomogeneities, one associated with the imaging devices and illumination variations, and the other associated with the essential characteristics of the intensities in objects and backgrounds. We name the second kind of intensity inhomogeneity as intensity complexity. In this paper, we focus on the segmentation of the images with intensity complexity. Our main argument is to quantify the complex intensities and convert them into useful features to improve segmentation accuracy. Two new quantities called the region intensity complexity index (RIC-Index) and factor (RIC-Factor) are introduced to quantify the intensity complexity. Then the quantified intensity complexity is incorporated into a variational level set framework. The total energy functional of the proposed framework consists of the following three items: a region intensity complexity term, a local region fitting energy term, and an edge-based energy term. The first term is defined by exploiting the region intensity complexity factor of the images. Mean and variance are utilized in the local region fitting energy to describe the image texture information. The last term of the energy functional, which is also derived from the region intensity complexity factor, incorporates the significant edge information. By integrating these three terms, the proposed model can handle intensity complexity images, especially two kinds of images: one with complex intensities in the objects, and the other with complex intensities in the backgrounds. The experimental results on 40 intensity complexity images and 1000 natural images from the Extended Complex Scene Saliency Dataset have indicated that our proposed algorithm can produce satisfactory segmentation results in comparison with five state-of-the-art methods and a deep learning approach.

Region-based object and background extraction via active contours

Human Contour Extraction Using Level Set

Neighbor-constrained Active Contours Without Edges.

Visual Contour Tracking Based on Inner-Contour Model Particle Filter under Complex Background

A convex neighbor-constrained active contour model for image segmentation

Neighborhood Aided Implicit Active Contours

A Hybride Active Contour Model Driven by Global and Local Image Information

Salient region detection using background contrast

Region Intensity Complexity Active Contours.

Upper Body Human Detection and Segmentation in Low Contrast Video

Active Contour Model Based Oil Distribution Matching and Its Image Segmentation Algorithm

Salient Object Extraction Based on Region Saliency Ratio

Active Contours Based on Image Laplacian Fitting Energy

Motion Objects Segmentation Using a New Level Set Based Method

Automatic Object Cosegmentation In Sparse Multiview Images

Active Contour Model Based on Non-Parametric 2D Entropy

Background Modeling Based On Region Segmentation

Region-edge-based active contours driven by hybrid and local fuzzy region-based energy for image segmentation

Dynamic Background Subtraction Using Spatial-Color Binary Patterns

Efficient Perceptual Region Detector Based on Object Boundary.

Region-Based Correlation Method of Kernel Background Modeling