Dataset Viewer

The dataset viewer is not available because its heuristics could not detect any supported data files. You can try uploading some data files, or configuring the data files location manually.

MetricStereo evaluation benchmarks

Preprocessed test data for seven of the nine benchmarks used in the MetricStereo evaluation (CVPR 2027 submission, anonymous). Each zip unpacks to <benchmark>/{images,right,gt_depth,intrinsics,manifest.csv}:

  • images/NNNN.png: left input view.
  • right/NNNN.png: captured right view (stereo benchmarks).
  • gt_depth/NNNN.npy: metric depth, float32 (depth benchmarks).
  • intrinsics/NNNN.npy: 3x3 camera matrix.
  • manifest.csv: per-pair metric baseline on the stereo benchmarks.
  • gt_disparity/NNNN.npy: ground-truth disparity in pixels of the full-resolution image, float32, invalid pixels inf (Middlebury only).

Checksums are in SHA256SUMS.

Sintel and LayeredFlow are not included, because their terms do not allow redistribution. The inference code includes a script that downloads them from the official sources and builds identical files.

Contents and licenses

The data is not relicensed. Every zip stays under the license of its original dataset, listed below. By downloading a zip you agree to those terms. The CC BY-NC-SA zips (KITTI, ETH3D) are shared under the same license.

File Benchmark Pairs Original license Source
booster.zip Booster (balanced test split) 191 CC BY-NC 4.0 https://cvlab-unibo.github.io/booster-web/
middlebury.zip Middlebury 2014 45 Free to use and publish, citation requested https://vision.middlebury.edu/stereo/data/
spring.zip Spring (test split) 1000 CC BY 4.0 https://spring-benchmark.org
kitti.zip KITTI (161-image eval split) 161 CC BY-NC-SA 3.0 https://www.cvlibs.net/datasets/kitti/
eth3d.zip ETH3D 454 CC BY-NC-SA 4.0 https://www.eth3d.net
tum.zip TUM RGB-D 1815 CC BY 4.0 https://cvg.cit.tum.de/data/datasets/rgbd-dataset
nyu.zip NYUv2 (654-image test split) 654 No license stated, citation requested https://cs.nyu.edu/~fergus/datasets/nyu_depth_v2.html

Changes from the originals

  • Images are re-encoded as PNG.
  • Booster, Middlebury and Spring are used as released (pre-rectified, native resolution).
  • Depth is converted to float32 metres.
  • Files are renumbered NNNN in a fixed order.

No other changes.

Citations

If you use this data, cite the original datasets:

@inproceedings{ramirez2022booster, title={Open Challenges in Deep Stereo: the {B}ooster Dataset}, author={Ramirez, Pierluigi Zama and Tosi, Fabio and Poggi, Matteo and Salti, Samuele and Mattoccia, Stefano and Di Stefano, Luigi}, booktitle={CVPR}, year={2022}}
@inproceedings{scharstein2014middlebury, title={High-Resolution Stereo Datasets with Subpixel-Accurate Ground Truth}, author={Scharstein, Daniel and Hirschm{\"u}ller, Heiko and Kitajima, York and Krathwohl, Greg and Ne{\v{s}}i{\'c}, Nera and Wang, Xi and Westling, Porter}, booktitle={GCPR}, year={2014}}
@inproceedings{mehl2023spring, title={Spring: A High-Resolution High-Detail Dataset and Benchmark for Scene Flow, Optical Flow and Stereo}, author={Mehl, Lukas and Schmalfuss, Jenny and Jahedi, Azin and Nalivayko, Yaroslava and Bruhn, Andr{\'e}s}, booktitle={CVPR}, year={2023}}
@inproceedings{geiger2012kitti, title={Are we ready for Autonomous Driving? The {KITTI} Vision Benchmark Suite}, author={Geiger, Andreas and Lenz, Philip and Urtasun, Raquel}, booktitle={CVPR}, year={2012}}
@inproceedings{schops2017eth3d, title={A Multi-View Stereo Benchmark with High-Resolution Images and Multi-Camera Videos}, author={Sch{\"o}ps, Thomas and Sch{\"o}nberger, Johannes L. and Galliani, Silvano and Sattler, Torsten and Schindler, Konrad and Pollefeys, Marc and Geiger, Andreas}, booktitle={CVPR}, year={2017}}
@inproceedings{sturm2012tum, title={A Benchmark for the Evaluation of {RGB-D} {SLAM} Systems}, author={Sturm, J{\"u}rgen and Engelhard, Nikolas and Endres, Felix and Burgard, Wolfram and Cremers, Daniel}, booktitle={IROS}, year={2012}}
@inproceedings{silberman2012nyu, title={Indoor Segmentation and Support Inference from {RGBD} Images}, author={Silberman, Nathan and Hoiem, Derek and Kohli, Pushmeet and Fergus, Rob}, booktitle={ECCV}, year={2012}}
Downloads last month
49