Image segmentation and bias correction via an improved level set method

Intensity inhomogeneity causes considerable difficulty in the quantitative analysis of magnetic resonance (MR) images. Thus bias field estimation is a necessary pre-processing step before quantitative analysis of MR data. This paper presents a variational level set approach for bias correction and segmentation for images with intensity inhomogeneities. Our method is based on the observation that local intensity variations in relatively smaller regions are separable, despite the inseparability of the whole image. In the beginning we define a function for clustering the image pixels in a smaller neighborhood. The cluster centers in this objective function have a multiplicative factor that estimates the bias within the neighborhood. Generally the local intensity variations are described by the Gaussian distributions with different means and variances. In this work the objective functions are integrated over the entire domain with local Gaussian distribution of fitting energy, ultimately analyzing the data with a level set framework. Our method is able to capture bias of quite general profiles. Moreover, our model can also distinguish regions with similar intensity distribution with different variances. The proposed method has been rigorously validated with images acquired on variety of imaging modalities with promising results.