遇见数据集

PNGs and Annotations of Brain Extraction Results

收藏
Zenodo2026-01-21 更新2026-05-29 收录
官方服务:

资源简介:

These are a series of PNGs from the MISTIE (Minimally Invasive Surgery Plus Alteplase for Intracerebral Hemorrhage Evacuation) III trial that were used to evaluate 7 skull stripping methods. We excluded implementations using Docker only or MATLAB due to high-performance computing cluster restrictions. If a Docker solutionwas provided, we attempted to run the code outside the container. The methods were: CTBET from Muschelli et al 2015 Robust CTBET CTBET-UNET CTBET-Docker MindGrab/Brainchop SynthStrip HD-CTBET To evaluate the following methods, we qualitatively visually inspected patients in the MISTIE phase III clinical trial. For all methods other than CTBET-UNET, there should be 5062 scans from 499 patients in the MISTIE phase III clinical trial. The CTBET-UNET trained on images that were 512x512, without any preprocessing done to images, so scans that are do not conform to that fail. CTBET-UNET successfully generated brain masks for 2,425 scans, but failed on approximately 40% of scans, namely those with gantry tilt.

提供机构:
Zenodo
创建时间:
2026-01-21
二维码
社区交流群
二维码
科研交流群
商业服务