BI-GCN: Boundary-Aware Input-Dependent Graph Convolution Network for Biomedical Image Segmentation

Meng, Yanda; Zhang, Hongrun; Gao, Dongxu; Zhao, Yitian; Yang, Xiaoyun; Qian, Xuesheng; Huang, Xiaowei; Zheng, Yalin

doi:10.48550/arxiv.2110.14775

Cited by 4 publications

(4 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To address these limitations, Li et al (2020) introduced an attention diagonal matrix to learn a better distance metric, which discarded the projection and re-projection processes. Inspired by (Li et al 2020), subsequent works (Meng et al 2021(Meng et al , 2023 developed the adjacency matrix based on energy aggregation in both image space and channel space. Moreover, a novel graph propagation model was constructed, which incorporated the boundary awareness.…”

Section: Graph Convolutional Network In Image Taskmentioning

confidence: 99%

Automatic breast ultrasound (ABUS) tumor segmentation based on global and local feature fusion

Li,

Ren,

Cheng

et al. 2024

Phys. Med. Biol.

View full text Add to dashboard Cite

Accurate segmentation of tumor regions in automated breast ultrasound (ABUS) images is of paramount importance in computer-aided diagnosis (CAD) system. However, the inherent diversity of tumors and the imaging interference pose great challenges to ABUS tumor segmentation. In this paper, we propose a global and local feature interaction model combined with graph fusion (GLGM), for 3D ABUS tumor segmentation. In GLGM, we construct a dual branch encoder-decoder, where both local and global features can be extracted. Besides, a global and local feature fusion (GLFF) module is designed, which employs the deepest semantic interaction to facilitate information exchange between local and global features. Additionally, to improve the segmentation performance for small tumors, a graph convolution-based shallow feature fusion module (SFFGC) is designed. It exploits the shallow feature to enhance the feature expression of small tumors in both local and global domains. The proposed method is evaluated on a private ABUS dataset and a public ABUS dataset. For the private ABUS dataset, the small tumors (volume smaller than 1 cm³) account for over 50% of the entire dataset. Experimental results show that the proposed GLGM model outperforms several state-of-the-art segmentation models in 3D ABUS tumor segmentation, particularly in segmenting small tumors.

show abstract

Section: Graph Convolutional Network In Image Taskmentioning

confidence: 99%

Automatic breast ultrasound (ABUS) tumor segmentation based on global and local feature fusion

Li,

Ren,

Cheng

et al. 2024

Phys. Med. Biol.

View full text Add to dashboard Cite

show abstract

“…A comprehensive comparison between DELFormer and 16 other excellent models is conducted to assess the effectiveness: DLinkNet, 31 NLinkNet, 32 DBRANet, 33 RCFSNet, 34 PSPNet, 35 DeepLabV3+, 36 MGU-Net, 37 DGCNet, 38 BI-GCN, 39 SGCN, 16 Seg-Road, 40 TransRoadNet, 41 TransResUNet, 42 SegFormer-b3, 43 SwinUNet, 44 and MTU-Net. 45 The loss and mIoU curves are shown in Fig.…”

Section: Performance Comparisonmentioning

confidence: 99%

DELFormer: detail-enhanced lightweight transformer for road segmentation

Xin,

Fu,

et al. 2023

J. Appl. Rem. Sens.

View full text Add to dashboard Cite

.The road segmentation task has become increasingly important in fields such as urban planning, traffic management, and environmental monitoring. However, most existing deep learning-based methods suffer from issues such as poor temporal effectiveness and connectivity, making it a significant challenge to achieve high-precision and high-efficiency road segmentation. We propose a road segmentation model based on a detail-enhanced lightweight transformer. Through the connectivity enhancement module, the issue of spatial information loss is addressed, enhancing the modeling capability of the road network connectivity. The model incorporates a detail-enhancement strategy to capture the relationship between roads and the environment, enhancing the perception and expression of details while maintaining low computational complexity. Furthermore, the use of a lightweight multiple feature fusion module promotes information fusion from features at different scales while a maintaining lightweight design. Extensive experiments on two publicly available datasets demonstrate that our method achieves the best performance in terms of real-time effectiveness and accuracy.

show abstract

“…Binary graph convolutional networks (Bi-GCNs) [25] use network parameter in a binary format while input node characteristics, while Bi-Directional Graph Convolutional Networks (Bi-GCN) [26] explore both bottom-up and top-down features. Graph Convolution [27] proposes an improved Laplacian for different tasks. These methods aim to enhance the accuracy and efficiency of traffic prediction models by incorporating different types of graphs and convolutional operations.…”

Section: Related Workmentioning

confidence: 99%

Implicit Sensing Self-Supervised Learning Based on Graph Multi-pretext tasks for Traffic Flow Prediction

Sattarzadeh,

Pathirana,

Palaniswami

2023

Preprint

View full text Add to dashboard Cite

In recent years, spatial-temporal graph neural networks (GNNs) have significantly improved traffic prediction by modeling intricate spatiotemporal dependencies in irregular traffic networks. However, these approaches may underutilize the intrinsic properties of traffic data and suffer from overfitting due to their local nature. This paper introduces the Implicit Sensing Self-Supervised learning model (ISSS) to address these issues. ISSS leverages a multi-pretext task framework for traffic flow prediction, integrating multiple self-supervised tasks, including contrastive learning and spatial jigsaw puzzles, to learn both specific and general representations. ISSS emphasizes capturing the intrinsic spatial relationships among sensor locations, particularly through spatial and temporal jigsaw puzzles. These tasks explore the correlation between traffic patterns at sensor locations and critical temporal periods, recognizing the significance of temporal information in traffic data. By transforming data into an alternative feature space, ISSS facilitates the adoption of contrastive learning tasks, enhancing regularization, and promoting a deeper understanding of learned traffic features, resulting in more specific representations. Comparative experiments on six datasets demonstrate ISSS's effectiveness in learning general and discriminative features in both supervised and unsupervised modes. ISSS outperforms existing models, showcasing its potential for improving traffic flow prediction while addressing challenges associated with local operations and overfitting.

show abstract

BI-GCN: Boundary-Aware Input-Dependent Graph Convolution Network for Biomedical Image Segmentation

Cited by 4 publications

References 33 publications

Automatic breast ultrasound (ABUS) tumor segmentation based on global and local feature fusion

Automatic breast ultrasound (ABUS) tumor segmentation based on global and local feature fusion

DELFormer: detail-enhanced lightweight transformer for road segmentation

Implicit Sensing Self-Supervised Learning Based on Graph Multi-pretext tasks for Traffic Flow Prediction

Contact Info

Product

Resources

About