Abstract:
Intensity inhomogeneity is one of the major artifacts in magnetic resonance imaging (MRI). Bias field present in MRI images alters true pixel value and produces spurious varying pixel intensities. This artifact affects the diagnosis by radiologists in a detrimental manner and also degrades the performance of computer-aided diagnosis algorithms such as segmentation. The present work proposes a novel network called InhomoNet for intensity inhomogeneity correction of MRI image. The generator architecture of InhomoNet consists of a new multi-scale local information module at each encoder block that helps to capture features at multiple scales. The horizontal and vertical kernels help to reduce the problems like loss of neighborhood information, gridding issues caused due to large dilated convolution operations. The attention-driven skip connections in the generator network are utilized to transfer optimal semantic and spatial localization information from the encoder to decoder blocks. Further, the present work proposes two new losses functions, i.e. histogram correlation and 3D pixel loss. These losses help to realize pixel consistency across different regions of brain MRI. The inculcation of the L1 loss provides guidance to the upsampling process as it compares the prediction from each decoder block with the ground truth. The proposed method is evaluated on simulated and real MRI data. The comparative analysis with popular state-of-the-art methods depicts the ability of the proposed method to perform intensity inhomogeneity correction accurately. © 2020 Elsevier Ltd
