Multi-Scale Adaptive Level Set Segmentation Method Based on Saliency

Zhang Dan,Yang He,Li Tieshan,Chen C L Philip

doi:10.1109/access.2019.2945112

Zhang Dan, Yang He + Show 2 more

Open Access

https://doi.org/10.1109/access.2019.2945112

Copy DOI

Abstract

Image segmentation is an important research of computer vision. Due to the effects of intensity inhomogeneity, target edge and background complex, it is still challenging to achieve effective segmentation of target adaptively. To solve these issues, an image segmentation method based on saliency and level set is proposed in this paper. First, adaptive initial contour of level set is got by wavelet-based feature probability evaluation (WFPE) model, the initial contour is closer to the target contour, which can reduce background interference and evolve faster. Second, in order to realize the best detection of intensity mutation and locate the target edge more accurately, an edge constraint energy term is introduced with multi-scale information obtained by wavelet transform. Finally, to improve segmentation adaptability and speed, the region information and edge constraint energy term are merged into the adaptive active contour model, the final evolution curve evolves in coarse scale, and then interpolates to get the final segmentation contour. Experimental results show that the proposed method achieves high efficiency in the following aspects: adaptability to images, speed of evolution, close to human visual perception.

Highlights

As an emerging topic, image segmentation technology has received extensive attention due to its wide application in medical images [1]–[3], aerotechnics [4], video processing [5], [6] and so on
The experimental results show the following: (1) The method proposed by us has the best effect on the segmentation accuracy of intensity inhomogeneity images
We introduced the multi-scale edge constraint energy term, fused saliency region and edge information. It overcomes the intensity inhomogeneity of images, and utilizes high frequency information to evolve on low frequency images

Summary

Introduction

Image segmentation technology has received extensive attention due to its wide application in medical images [1]–[3], aerotechnics [4], video processing [5], [6] and so on. The segmentation of an image refers to subdividing the image into its components or objects. Segmentation is one of significant subjects in image analysis. It has many different methods, including threshold value methods [7]–[10], boundary tracking methods [11]–[13], et al segmentation of images with noise, intensity inhomogeneity and complex edge is still a big challenge and a hotspot. There are a series of breakthrough successes having been made in recent years, especially the active contour model approach [14], [15]. The boundary of target which is segmented by active contour model is smooth and closed

Methods

Discussion

Conclusion