We have created an anatomically and mechanically realistic brain phantom from polyvinyl alcohol cryogel (PVA-C) for validation of image processing methods for segmentation, reconstruction, registration, and denoising. PVA-C is material widely used in medical imaging phantoms for its mechanical similarities to soft tissues. The phantom was cast in a mold designed using the left hemiphere of the Colin27 brain dataset and contains deep sulci, a complete insular region, and an anatomically accurate left ventricle. Marker spheres were also implanted to enable good registration. Inflatable catheters were used to simulate tissue deformations.
The phantom was designed for triple modality imaging, giving good contrast images in computed tomography, ultrasound, and magnetic resonance imaging. Imaging data from each modality, and from multiple deformation level, were acquired. This page contains images files of the PVA Brain phantom. For each deformation instant, you can download one pack that archives reconstructed and unreconstructed B-mode ultrasound (US) images at 5.2cm and 7.1cm depth, as well as reformatted MR and CT volumes. These images are made freely available to the image processing community.
- "MR-CT" contains the MR and CT images. The MR sequence contains a 3D T1, Flair, T2 and Diffusion weighted sequence.
- "US" contains the acquired ultrasound slices, with different
view. For each sequence, the raw data in stradx format
(http://mi.eng.cam.ac.uk/~rwp/stradx/stradx_files.html) are provided,
as well as the reconstructed volumes in nifty format using the distance
weighted interpolation. For each acquisition, a transformation matrix
(.mat) is also provided that matches the US coordinate system to those
of the "preoperative" MR image.
- "US_denoised" contains the denoised version of the reconstructed US volumes using the NL-means method.
- "MR_cropped" and "CT_cropped" contain the resliced MR and CT volume for each US acquisition. The cropped MR and CT were resliced according to the transformation of the neuronavigation system. Therefore, the volumes are roughly aligned and not perfectly registered. The transformation stored in file US.mat describes the transformation matrix that was used to reslice the MR according to the FOV of the US. The transformation was used as an input of the VTKReslice class.
|Catheter 1||Catheter 2||MR,CT and US Images|
|0ml||0ml||All US, cropped MR and cropped CT volumes (around 200M)|
|5ml||0ml||All US, cropped MR and cropped CT volumes (around 200M)|
|5ml||5ml||All US, cropped MR and cropped CT volumes (around 200M)|
|5ml||10ml||All US, cropped MR and cropped CT volumes (around 200M)|
|10ml||10ml||All US, cropped MR and cropped CT volumes (around 200M)|