The home of the Adelaide RMF datasets is https://www.ai4space.group/research/adelaidermf — but the links there are broken. The first contact listed on that page — Hoi Sim Wong - kindly emailed me a Dropbox link to the data. The data almost completely corresponds to another copy I found at data/AdelaideRMF in https://github.com/trungtpham/RCMSA, commit 0d5f3e3, but the copies here are the copies fro the Dropbox link from the main author.

In the home page above above, the format of the files is given as:

Each Matlab .mat file contains:

• img1 - left image
• img2 - right image
• data - keypoint correspondences (x1,y1,1,x2,y2,1) where (x1,y1) in img1 and (x2,y2) in img2
• score - SIFT correspondences matching score
• label - 0 indicates gross outliers and others (e.g., 1, 2, ...) indicate the structure membership

For citation of these datasets, see the main page above. In particular:

If you use this dataset in a publication, please cite:

Hoi Sim Wong, Tat-Jun Chin, Jin Yu and David Suter, Dynamic and
Hierarchical Multi-Structure Geometric Model Fitting, International
Conference on Computer Vision (ICCV), Barcelona, Spain, 2011.

@inproceedings{wong11,
  author = {Hoi Sim Wong and Tat-Jun Chin and Jin Yu and David Suter},
  booktitle = {International Conference on Computer Vision (ICCV)},
  title = {Dynamic and hierarchical multi-structure geometric model fitting},
  year = {2011}
}

To run the version of scikit-image that I am using:

git clone https://github.com/matthew-brett/scikit-image
# Currently commit 8cd967682
git co refactor-fundamental

and then follow the instructions to install a development version.