研究目的
To develop an automatic method for BVS malapposition analysis in intravascular optical coherence tomography images, enabling accurate and efficient detection and segmentation of BVS struts.
研究成果
The proposed method is accurate and robust for BVS struts detection and segmentation under complex background, and enables automatic malapposition analysis. It is of potential value for clinical research as well as medical care.
研究不足
The performance on sets with severe blood artifacts was slightly inferior, indicating potential areas for optimization in handling complex backgrounds.
1：Experimental Design and Method Selection:
The method involves struts detection using a deep learning-based detector (R-FCN), struts segmentation using dynamic programming, and malapposition analysis based on the segmentation.
2：Sample Selection and Data Sources:
17 pullbacks of IVOCT images were obtained from the FD-OCT system for training and testing.
3：List of Experimental Equipment and Materials:
IVOCT images from the C7-XR system (St. Jude, St. Paul, Minnesota).
4：Experimental Procedures and Operational Workflow:
Struts are detected by a trained R-FCN detector, segmented using dynamic programming, and then malapposition is analyzed.
5：Data Analysis Methods:
Quantitative evaluation includes True positive rate (TPR), false positive rate (FPR), center position error (CPE), and Dice coefficient.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
