An Active Contour Model for the Segmentation of Images with Intensity Inhomogeneities and Bias Field Estimation

Abstract

Intensity inhomogeneity causes many difficulties in image segmentation and the understanding of magnetic resonance (MR) images. Bias correction is an important method for addressing the intensity inhomogeneity of MR images before quantitative analysis. In this paper, a modified model is developed for segmenting images with intensity inhomogeneity and estimating the bias field simultaneously. In the modified model, a clustering criterion energy function is defined by considering the difference between the measured image and estimated image in local region. By using this difference in local region, the modified method can obtain accurate segmentation results and an accurate estimation of the bias field. The energy function is incorporated into a level set formulation with a level set regularization term, and the energy minimization is conducted by a level set evolution process. The proposed model first appeared as a two-phase model and then extended to a multi-phase one. The experimental results demonstrate the advantages of our model in terms of accuracy and insensitivity to the location of the initial contours. In particular, our method has been applied to various synthetic and real images with desirable results.

Weighted kernel mapping model with spring simulation based watershed transformation for level set image segmentation

“…The proposed WKM model is compared with the aforementioned four models: Local Image Fitting (LIF) model [17], Bias Correction based Local Binary Fitting (BCLBF) model [18] and its modified model (MBCLBF) [22], as well as Level set image segmentation with Kernel Induced Data Term (KM) model [16]. In each experiment, the settings of initial curves and the parameters of each model are based on the recommendations of the original papers [16][17][18]22]. Some parameters are slightly adjusted in order to get the better results.…”

“…Over the past few decades, the active contour models have shown promising results through using level set methods in image segmentation [5][6][7][8][9][10][11][12]. The active contour models by level set methods can be broadly categorised into two basic types: edge-based methods [5][6][7][8][9] and region-based methods [10][11][12][13][14][15][16][17][18][19][20][21][22]. The edge-based methods utilise image gradients to drive the level set evolution.…”

“…Among them, the Chan-Vese (CV) model is one of the most representative and popular region-based models, and has been extended in various ways [15][16][17][18]. Nowadays, the region-based models can be mainly classified into two groups: global region-based models [12][13][14]16] and local region-based models [15,17,18,22].…”

“…The Local Images Fitting (LIF) model [17] introduces the local fitted image energy functional to extract the local information, then uses the Gaussian filtering to regularise the level set function. Huang et al [22] proposed a modified BCLBF model (MBCLBF) to improve the segmentation results of the images with intensity inhomogeneity. Comparing with the global region-based models, the local region-based models are more complex and time-consuming [16].…”

“…Unlike the existing methods in the literature that the corrected image [18,22] can be obtained in the level set evolving process, the proposed approach utilises the EWT method to process the image with intensity inhomogeneity or noise before the level set evolution. The weighted information for performing the level set image segmentation will be prepared via the EWT process as well.…”