The multimodal MRI brain tumor segmentation based on AD-Net. (February 2023)

• Record Type: Journal Article
• Authors: Peng, Yanjun; Sun, Jindong
• Is Part Of: Biomedical signal processing and control. Volume 80:Part 2(2023)
• Publication Date: 2023-02
• Subjects: Deep learning -- Image processing -- Biomedical imaging segmentation -- Artificial neural networks; Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2022.104336
• Languages: English
• ISSNs: 1746-8094

Abstract

Abstract: Multimodal glioma images provide different features of tumor boundaries based on magnetic resonance imaging (MRI), where multimodal features are often challenging to extract for deep learning segmentation methods. Disturbance between features of different modes is an important factor restricting multimodal learning. To efficiently extract multimodal features, we propose an automatic weighted dilated convolutional network (AD-Net) to learn multimodal brain tumor features through channel feature separation learning. Specifically, the auto-weight dilated convolutional unit (AD unit) utilizes dual-scale convolutional feature maps to acquire channel separation features. We employ two learnable parameters to fuse dual-scale convolutional feature maps in encoding layers, and the two learnable parameters are automatically adjusted with the back propagation of the gradient. We adopt the Jensen–Shannon divergence to constrain the distribution of its feature map, which in turn regularizes the weights of the entire down-sampling. In addition, we use the training technique of deep supervision to achieve fast fitting. Our proposed method got dice scores of 0.90, 0.80, and 0.76 for the whole tumor (WT), the tumor core (TC), and the enhancing tumor (ET) on the BraTS20 dataset. The experimental results showed good performance with the AD-Net network. Highlights: We apply the dilated convolution with a dual-sized kernel to expand the receptive field of convolution and fuse the features to build the auto-weight dilated network. We use densely stacked grouping convolution to reduce the number of model parameters and speed up training. We adopt Jensen–Shannon divergence to constrain the distribution of its feature map, which in turn regularizes the weights of the entire down-sampling. We use the training technique of deep supervision, add additional loss calculation, and update the weight of the encoding layer with an additional gradient, to achieve the purpose of fast fitting.