This repository contains the BiomedSegFM dataset, a crucial resource for the CVPR 2025 Competition: Foundation Models for 3D Biomedical Image Segmentation . Foundation Models for Interactive 3D Biomedical Image Segmentation (Homepage) Foundation Models for Text guided 3D Biomedical Image Segmentation (Homepage) CVPR 2025 Competition: Foundation Models for 3D Biomedical Image Segmentation Highly recommend watching the webinar recording to learn about the task settings and baseline methods. The dataset covers five commonly used 3D biomedical image modalities: CT, MR, PET, Ultrasound, and Microscopy. All the images are from public datasets with a License for redistribution. To reduce dataset size, all labeled slices are extracted and preprocessed into npz files. Each npz file contains imgs : image data; shape: (D,H,W); Intensity range: [0, 255] gts : ground truth; shape: (D,H,W); spacing Folder structure 3D train npz all: complete training set 3D train npz random 10percent 16G: randomly selected 10% cases from the above training set. In other words, these cases have been included in the complete training set. Participants are allowed to use other criteria to select 10% cases as the co…
Runs entirely in your browser via DuckDB-Wasm — this dataset's real data file is loaded once, then queried locally. Nothing is sent to a server.
We use cookies for essential functionality and analytics. You can accept or reject analytics cookies.Cookie policy