CQformer: Learning Dynamics Across Slices in Medical Image Segmentation

Prevalent studies on deep learning-based 3D medical image segmentation capture the continuous variation across 2D slices mainly via convolution, Transformer, inter-slice interaction, and time series models. In this work, via modeling this variation by an ordinary differential equation (ODE), we prop...

Full description

Saved in:
Bibliographic Details
Published in:IEEE transactions on medical imaging Vol. 44; no. 2; pp. 1043 - 1057
Main Authors: Zhang, Shengjie, Shen, Xin, Chen, Xiang, Yu, Ziqi, Ren, Bohan, Yang, Haibo, Zhang, Xiao-Yong, Zhou, Yuan
Format: Journal Article
Language:English
Published: United States IEEE 01.02.2025
Subjects:
Algorithms
Biomedical imaging
Brain - diagnostic imaging
Convolution
cross instance query
Databases, Factual
Decoding
Deep Learning
Feature extraction
Head
Humans
Image segmentation
Imaging, Three-Dimensional - methods
Magnetic Resonance Imaging - methods
Medical image segmentation
ordinary differential equation
Shape
Three-dimensional displays
Time series analysis
Tomography, X-Ray Computed
transformer
Transformers
ISSN:0278-0062, 1558-254X, 1558-254X
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Prevalent studies on deep learning-based 3D medical image segmentation capture the continuous variation across 2D slices mainly via convolution, Transformer, inter-slice interaction, and time series models. In this work, via modeling this variation by an ordinary differential equation (ODE), we propose a cross instance query-guided Transformer architecture (CQformer) that leverages features from preceding slices to improve the segmentation performance of subsequent slices. Its key components include a cross-attention mechanism in an ODE formulation, which bridges the features of contiguous 2D slices of the 3D volumetric data. In addition, a regression head is employed to shorten the gap between the bottleneck and the prediction layer. Extensive experiments on 7 datasets with various modalities (CT, MRI) and tasks (organ, tissue, and lesion) demonstrate that CQformer outperforms previous state-of-the-art segmentation algorithms on 6 datasets by 0.44%-2.45%, and achieves the second highest performance of 88.30% on the BTCV dataset. The code is available at https://github.com/qbmizsj/CQformer .
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 23
ISSN:0278-0062
1558-254X
1558-254X
DOI:10.1109/TMI.2024.3477555