Related Experiment Video
Updated: Sep 24, 2025
08:40
Quantitation of Protein Expression and Co-localization Using Multiplexed Immuno-histochemical Staining and Multispectral Imaging
Published on: April 8, 2016
12.9K
A novel registration-based algorithm for prostate segmentation via the combination of SSM and CNN
Chunxia Qin1,2, Puxun Tu1, Xiaojun Chen1
1School of Mechanical Engineering, Shanghai Jiao Tong University, Shanghai, China.
Medical Physics
|May 4, 2022
Summary
This study introduces a novel algorithm for precise prostate segmentation, combining a convolutional neural network (CNN) with a statistical shape model (SSM). The method accurately delineates prostate regions, improving accuracy for interventions like biopsies and targeted therapy.
Area of Science:
- Medical imaging and computational anatomy.
- Development of advanced image segmentation algorithms.
Background:
- Accurate prostate segmentation is crucial for interventions like biopsies and targeted therapy.
- Challenges in prostate delineation arise from tissue ambiguity and unclear anatomical boundaries.
Purpose of the Study:
- To propose a novel supervised registration-based algorithm for precise prostate segmentation.
- To combine a convolutional neural network (CNN) with a statistical shape model (SSM) for improved delineation.
Main Methods:
- A two-branch network integrating an SSM-Net for boundary prediction and a ResU-Net for probability map generation.
- SSM-Net predicts shape parameters and a normalized distance map.
- The final segmentation combines the distance map and probability map outputs.
Main Results:
- The algorithm achieved a Dice score of 0.907 and an average surface distance of 1.85 mm on public datasets (PROMISE12, NCI-ISBI 2013).
- Demonstrated superior accuracy and efficiency compared to existing methods.
- Model elasticity augmentation and fine-tuning improved delineation accuracy by 10% and 7% respectively.
Conclusions:
- The proposed segmentation method is effective and robust for prostate delineation.
- It holds potential for enhancing precision in prostate interventions.
