Your browser doesn't support javascript.
Graph-BAS(3)Net: Boundary-Aware Semi-Supervised Segmentation Network with Bilateral Graph Convolution
18th IEEE/CVF International Conference on Computer Vision (ICCV) ; : 7366-7375, 2021.
Article in English | Web of Science | ID: covidwho-1927512
ABSTRACT
Semi-supervised learning (SSL) algorithms have attracted much attentions in medical image segmentation by leveraging unlabeled data, which challenge in acquiring massive pixel-wise annotated samples. However, most of the existing SSLs neglected the geometric shape constraint in object, leading to unsatisfactory boundary and non-smooth of object. In this paper, we propose a novel boundary-aware semi-supervised medical image segmentation network, named Graph-BAS(3)Net, which incorporates the boundary information and learns duality constraints between semantics and geometrics in the graph domain. Specifically, the proposed method consists of two components a multi-task learning framework BAS(3)Net and a graph-based cross-task module BGCM. The BAS(3)Net improves the existing GAN-based SSL by adding a boundary detection task, which encodes richer features of object shape and surface. Moreover, the BGCM further explores the co-occurrence relations between the semantics segmentation and boundary detection task, so that the network learns stronger semantic and geometric correspondences from both labeled and unlabeled data. Experimental results on the LiTS dataset and COVID-19 dataset confirm that our proposed Graph-BAS(3) Net outperforms the state-of-the-art methods in semi-supervised segmentation task.

Full text: Available Collection: Databases of international organizations Database: Web of Science Language: English Journal: CVF International Conference on Computer Vision (ICCV) Year: 2021 Document Type: Article

Similar

MEDLINE

...
LILACS

LIS


Full text: Available Collection: Databases of international organizations Database: Web of Science Language: English Journal: CVF International Conference on Computer Vision (ICCV) Year: 2021 Document Type: Article