|
View: |
Part 1: Document Description
|
|
Citation |
|
|---|---|
|
Title: |
Spring: A High-Resolution High-Detail Dataset and Benchmark for Scene Flow, Optical Flow and Stereo |
|
Identification Number: |
doi:10.18419/darus-3376 |
|
Distributor: |
DaRUS |
|
Date of Distribution: |
2023-03-14 |
|
Version: |
2 |
|
Bibliographic Citation: |
Mehl, Lukas; Schmalfuss, Jenny; Jahedi, Azin; Nalivayko, Yaroslava; Bruhn, Andrés, 2023, "Spring: A High-Resolution High-Detail Dataset and Benchmark for Scene Flow, Optical Flow and Stereo", https://doi.org/10.18419/darus-3376, DaRUS, V2 |
|
Citation |
|
|
Title: |
Spring: A High-Resolution High-Detail Dataset and Benchmark for Scene Flow, Optical Flow and Stereo |
|
Identification Number: |
doi:10.18419/darus-3376 |
|
Authoring Entity: |
Mehl, Lukas (Institute for Visualization and Interactive Systems, University of Stuttgart) |
|
Schmalfuss, Jenny (Institute for Visualization and Interactive Systems, University of Stuttgart) |
|
|
Jahedi, Azin (Institute for Visualization and Interactive Systems, University of Stuttgart) |
|
|
Nalivayko, Yaroslava (Institute for Visualization and Interactive Systems, University of Stuttgart) |
|
|
Bruhn, Andrés (Institute for Visualization and Interactive Systems, University of Stuttgart) |
|
|
Grant Number: |
251654672 |
|
Distributor: |
DaRUS |
|
Access Authority: |
Bruhn, Andrés |
|
Depositor: |
Mehl, Lukas |
|
Date of Deposit: |
2023-03-08 |
|
Holdings Information: |
https://doi.org/10.18419/darus-3376 |
|
Study Scope |
|
|
Keywords: |
Computer and Information Science, Computer Vision, Optical Flow, Computer Stereo Vision, Benchmark |
|
Abstract: |
The Spring dataset contains files for scene flow, optical flow and stereo estimation. For easier handling, we organized them into sub-directories: </br> <i>train</i> split: <ul> <li><code>train_frame_left.zip</code>: left camera frames</li> <li><code>train_frame_right.zip</code>: right camera frames</li> <li><code>train_disp1_left.zip</code>: left-to-right disparity in the reference frame</li> <li><code>train_disp1_right.zip</code>: right-to-left disparity in the reference frame</li> <li><code>train_disp2_FW_left.zip</code>: left-to-right disparity in the future/forward target frame</li> <li><code>train_disp2_BW_left.zip</code>: left-to-right disparity in the past/backward target frame</li> <li><code>train_disp2_FW_right.zip</code>: right-to-left disparity in the future/forward target frame</li> <li><code>train_disp2_BW_right.zip</code>: right-to-left disparity in the past/backward target frame</li> <li><code>train_flow_FW_left.zip</code>: left forward optical flow</li> <li><code>train_flow_BW_left.zip</code>: left backward optical flow</li> <li><code>train_flow_FW_right.zip</code>: right forward optical flow</li> <li><code>train_flow_BW_right.zip</code>: right backward optical flow</li> <li><code>train_cam_data.zip</code>: camera data: intrinsics, extrinsics, focal distance</li> <li><code>train_maps.zip</code>: additional maps: detail, match, rigid, sky</li> </ul> <i>test</i> split: <ul> <li><code>test_frame_left.zip</code>: left camera frames</li> <li><code>test_frame_right.zip</code>: right camera frames</li> <li><code>test_cam_data.zip</code>: camera data: intrinsics</li> </ul> </br> File formats: <ul> <li>images and maps are given in png format</li> <li>optical flow files are given in HDF5 file format and named <code>.flo5</code></li> <li>disparity files are given in HDF5 file format and named <code>.dsp5</code></li> </ul> For the project website see <a href="https://spring-benchmark.org">spring-benchmark.org</a>. |
|
Methodology and Processing |
|
|
Sources Statement |
|
|
Data Sources: |
The Spring movie assets (https://cloud.blender.org/spring) by Blender Foundation are licensed under CC BY 4.0 |
|
Data Access |
|
|
Other Study Description Materials |
|
|
Related Publications |
|
|
Citation |
|
|
Title: |
Lukas Mehl, Jenny Schmalfuss, Azin Jahedi, Yaroslava Nalivayko, Andrés Bruhn: Spring: A High-Resolution High-Detail Dataset and Benchmark for Scene Flow, Optical Flow and Stereo. Proc. IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2023. |
|
Identification Number: |
2303.01943 |
|
Bibliographic Citation: |
Lukas Mehl, Jenny Schmalfuss, Azin Jahedi, Yaroslava Nalivayko, Andrés Bruhn: Spring: A High-Resolution High-Detail Dataset and Benchmark for Scene Flow, Optical Flow and Stereo. Proc. IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2023. |
|
Label: |
test_cam_data.zip |
|
Text: | |
|
Notes: |
application/zip |
|
Label: |
test_frame_left.zip |
|
Text: | |
|
Notes: |
application/zip |
|
Label: |
test_frame_right.zip |
|
Text: | |
|
Notes: |
application/zip |
|
Label: |
train_cam_data.zip |
|
Text: | |
|
Notes: |
application/zip |
|
Label: |
train_disp1_left.zip |
|
Text: | |
|
Notes: |
application/zip |
|
Label: |
train_disp1_right.zip |
|
Text: | |
|
Notes: |
application/zip |
|
Label: |
train_disp2_BW_left.zip |
|
Text: | |
|
Notes: |
application/zip |
|
Label: |
train_disp2_BW_right.zip |
|
Text: | |
|
Notes: |
application/zip |
|
Label: |
train_disp2_FW_left.zip |
|
Text: | |
|
Notes: |
application/zip |
|
Label: |
train_disp2_FW_right.zip |
|
Text: | |
|
Notes: |
application/zip |
|
Label: |
train_flow_BW_left.zip |
|
Text: | |
|
Notes: |
application/zip |
|
Label: |
train_flow_BW_right.zip |
|
Text: | |
|
Notes: |
application/zip |
|
Label: |
train_flow_FW_left.zip |
|
Text: | |
|
Notes: |
application/zip |
|
Label: |
train_flow_FW_right.zip |
|
Text: | |
|
Notes: |
application/zip |
|
Label: |
train_frame_left.zip |
|
Text: | |
|
Notes: |
application/zip |
|
Label: |
train_frame_right.zip |
|
Text: | |
|
Notes: |
application/zip |
|
Label: |
train_maps.zip |
|
Notes: |
application/zip |