CLCU-Net: Cross-level connected U-shaped network with selective feature aggregation attention module for brain tumor segmentation. (August 2021)

• Record Type: Journal Article
• Authors: Wang, Y.L.; Zhao, Z.J.; Hu, S.Y.; Chang, F.L.
• Is Part Of: Computer methods and programs in biomedicine. Volume 207(2021)
• Publication Date: 2021-08
• Subjects: Deep learning -- Brain tumor segmentation -- Multi-scale feature connection -- Segmented attention module -- Selective feature aggregation; Medicine -- Computer programs -- Periodicals; Biology -- Computer programs -- Periodicals; Computers -- Periodicals; Medicine -- Periodicals; Médecine -- Logiciels -- Périodiques; Biologie -- Logiciels -- Périodiques; Biology -- Computer programs; Medicine -- Computer programs; Periodicals; Electronic journals; 610.28
• DOI: 10.1016/j.cmpb.2021.106154
• Languages: English
• ISSNs: 0169-2607

Abstract

Highlights: We developed a new encoder-decoder network, which not only has the skip concatenations between the encoding path and decoding path but also achieves feature connections separately on each path, to make the best of multi-scale features. We present a novel attention module, named segmented attention module (SAM), which combines channel attention mechanism with spatial attention mechanism reasonably and skillfully for segmentation tasks, to extract useful information and discard redundant information in connected features, i.e., achieve selective feature aggregation. We prove that a comprehensive combination of deep supervision and spatial pooling pyramid synergistically improves the performance of brain tumor segmentation. The proposed method outperforms six state-of-the-art methods introduced in 2015-2020, namely FCN, UNet, UNet3+, 3D-Unet, DeepLab v3+, and AttentionUnet. Abstract: Background and Objective: Brain tumors are among the most deadly cancers worldwide. Due to the development of deep convolutional neural networks, many brain tumor segmentation methods help clinicians diagnose and operate. However, most of these methods insufficiently use multi-scale features, reducing their ability to extract brain tumors' features and details. To assist clinicians in the accurate automatic segmentation of brain tumors, we built a new deep learning network to make full use of multi-scale features for improving the performance of brain tumor segmentation. Methods: We propose a novel cross-level connected U-shaped network (CLCU-Net) to connect different scales' features for fully utilizing multi-scale features. Besides, we propose a generic attention module (Segmented Attention Module, SAM) on the connections of different scale features for selectively aggregating features, which provides a more efficient connection of different scale features. Moreover, we employ deep supervision and spatial pyramid pooling (SSP) to improve the method's performance further. Results: We evaluated our method on the BRATS 2018 dataset by five indexes and achieved excellent performance with a Dice Score of 88.5%, a Precision of 91.98%, a Recall of 85.62%, a Params of 36.34M and Inference Time of 8.89ms for the whole tumor, which outperformed six state-of-the-art methods. Moreover, the performed analysis of different attention modules' heatmaps proved that the attention module proposed in this study was more suitable for segmentation tasks than the other existing popular attention modules. Conclusion: Both the qualitative and quantitative experimental results indicate that our cross-level connected U-shaped network with selective feature aggregation attention module can achieve accurate brain tumor segmentation and is considered quite instrumental in clinical practice implementation.