AdelaideRMF - Robust Model Fitting Dataset

AdelaideRMF is a data set for robust geometric model fitting (homography estimation and fundamental matrix estimation). We collected a set of image pairs and manually labelled the keypoint correspondences which were obtained by SIFT matching.

Download

• Complete dataset

• Individual correspondence sets

• Original images

Format

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

Citation

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}

}

Correspondence sets

Click on each image to download.

Change log

• 16 June 2012 - original camera images are now available for download.

• 6 June 2012 - added 4 more data to AdelaideRMF (see data 35-38). adelaidermf.rar is also updated.

• 2 May 2011 - 34 data of AdelaideRMF is available online.

Enquiries

For enquires regarding this data set, please email hoisimwong@gmail.com or tat-jun.chin@adelaide.edu.au.