Spring Dataset and Benchmark Logo

Spring: L. Mehl, J. Schmalfuss, A. Jahedi, Y. Nalivayko, A. Bruhn — University of Stuttgart

RobustSpring: V. Oei, J. Schmalfuss, L. Mehl, M. Bartsch, S. Agnihotri, M. Keuper, A. Bulling, A. Bruhn — University of Stuttgart, University of Mannheim, MPI for Informatics

Not logged in | Login | Create Account

Show: All | Accuracy | Robustness

💡 Please note that methods marked "submitted by spring team" have not been finetuned on Spring.

Name 1px ▲
total 		1px
low-det. 		1px
high-det. 		1px
matched 		1px
unmat. 		1px
rigid 		1px
non-rig. 		1px
not sky 		1px
sky 		1px
s0-10 		1px
s10-40 		1px
s40+ 		EPE Fl WAUC robust
ΔEPE 		robust
ΔFl 		robust
Δ1px
 
1
RoCo-23-SPEAR-L-K1  
2.836 2.514 53.563 2.357 22.655 1.242 14.894 2.887 2.073 0.854 3.452 17.756 0.268 1.039 95.818 0.658 2.162 8.146
Anonymous.
2
RoCo-23  
2.838 2.515 53.661 2.360 22.626 1.247 14.878 2.876 2.257 0.863 3.464 17.664 0.269 1.044 95.851 n/a n/a n/a
Anonymous.
3
RoCo-23-SPEAR-L-K3  
2.846 2.525 53.412 2.366 22.704 1.252 14.904 2.900 2.027 0.865 3.451 17.776 0.273 1.036 95.669 0.550 1.813 6.186
Anonymous.
4
RoCo-14 CAR-WAFT  
3.151 2.798 58.640 2.579 26.793 1.322 16.986 3.277 1.239 1.099 4.182 17.531 0.298 1.195 95.038 1.586 5.091 11.129
 
5
FreeFlow-L   code
3.192 2.855 56.187 2.616 26.996 1.430 16.522 3.247 2.348 0.936 3.607 20.941 0.278 1.048 95.235 1.613 6.123 15.281
V. Bargatin, A. Yakovenko, K. Abud, and D. Vatolin. "FreeFlow: A Bias-free Hierarchical Transformer for Optical Flow Estimation", ECCV 2026.
6
RoCo-14-Test  
3.226 2.873 58.802 2.642 27.378 1.363 17.317 3.324 1.725 1.112 4.236 18.181 0.299 1.195 94.931 1.597 5.356 11.825
 
7
ARFlow  
3.265 2.926 56.655 2.717 25.955 1.436 17.108 3.325 2.362 1.079 4.160 19.133 0.353 1.212 95.283 n/a n/a n/a
Anonymous.
8
TDFlow  
3.283 2.943 56.757 2.725 26.346 1.469 17.007 3.308 2.891 1.073 4.166 19.381 0.363 1.219 95.210 n/a n/a n/a
Anonymous.
9
MEMFOF   code
3.289 2.947 57.246 2.751 25.551 1.446 17.236 3.327 2.723 1.084 4.202 19.270 0.355 1.238 95.186 n/a n/a n/a
V. Bargatin, E. Chistov, A. Yakovenko, and D. Vatolin. "MEMFOF: High-Resolution Training for Memory-Efficient Multi-Frame Optical Flow Estimation", ICCV 2025.
10
RoCo-19-Q4-Independent [SF]  
3.295 2.961 55.952 2.719 27.129 1.431 17.399 3.375 2.078 1.118 3.977 19.663 0.328 1.255 95.113 n/a n/a n/a
 
11
WAFT-DAv2-a2   code
3.298 2.947 58.502 2.731 26.732 1.433 17.407 3.362 2.319 1.126 4.197 19.030 0.304 1.197 94.990 n/a n/a n/a
WAFT: Warping-Alone Field Transforms for Optical Flow. Wang et al. ICLR 2026
12
FreeFlow-M   code
3.392 3.043 58.277 2.786 28.466 1.493 17.753 3.475 2.134 0.983 3.855 22.291 0.346 1.171 94.919 2.046 7.265 17.398
V. Bargatin, A. Yakovenko, K. Abud, and D. Vatolin. "FreeFlow: A Bias-free Hierarchical Transformer for Optical Flow Estimation", ECCV 2026.
13
ACT  
3.433 3.067 61.027 2.853 27.436 1.423 18.638 3.493 2.515 1.175 4.496 19.445 0.351 1.309 94.863 n/a n/a n/a
Anonymous.
14
GeoMFlow  
3.439 3.095 57.612 2.869 27.025 1.612 17.264 3.323 5.208 1.061 4.235 21.182 0.351 1.238 95.208 n/a n/a n/a
Anonymous.
15
DPFlow   code
3.442 3.102 56.941 2.859 27.563 1.500 18.132 3.522 2.218 1.188 3.998 20.786 0.340 1.311 94.980 n/a n/a n/a
H. Morimitsu et al. "DPFlow: Adaptive Optical Flow Estimation with a Dual-Pyramid Framework". CVPR 2025.
16
test  
3.481 3.131 58.631 2.899 27.562 1.567 17.961 3.492 3.315 1.126 4.203 21.230 0.510 1.378 94.757 1.656 4.480 10.962
 
17
FAST-Huge  
3.494 3.150 57.677 2.912 27.549 1.502 18.563 3.517 3.138 1.038 4.025 22.597 0.323 1.212 94.837 n/a n/a n/a
Anonymous.
18
MS-RAFT-3D+ [SF]   code
3.520 3.159 60.379 3.005 24.814 1.310 20.242 3.669 1.261 1.028 4.160 22.638 0.359 1.431 95.108 n/a n/a n/a
J. Schmid, A. Jahedi, N. Berenguel Senn and A. Bruhn. "MS-RAFT-3D: A Multi-scale Architecture for Recurrent Image-based Scene Flow". ICIP 2025.
19
MEMFOF (w/o ft)   code
3.600 3.254 58.072 3.049 26.384 1.510 19.416 3.708 1.961 1.315 4.574 20.081 0.432 1.353 94.481 n/a n/a n/a
V. Bargatin, E. Chistov, A. Yakovenko, and D. Vatolin. "MEMFOF: High-Resolution Training for Memory-Efficient Multi-Frame Optical Flow Estimation", ICCV 2025.
20
RoCo-45  
3.640 3.294 58.133 3.085 26.625 1.530 19.607 3.744 2.060 1.325 4.561 20.521 0.431 1.356 94.460 n/a n/a n/a
 
21
OFM (w/o ft)  
3.660 3.311 58.700 3.064 28.315 1.535 19.742 3.762 2.107 1.241 4.428 21.823 0.468 1.477 94.462 n/a n/a n/a
Anonymous.
22
SEA-RAFT  
3.686 3.322 60.991 3.025 31.057 1.561 19.767 3.757 2.615 1.241 4.760 21.234 0.363 1.347 94.534 2.960 9.050 17.520
💡 submitted by spring team | Wang, Yihan, Lahav Lipson, and Jia Deng. "Sea-raft: Simple, efficient, accurate raft for optical flow." In ECCV, 2024.
23
SEA-RAFT(M)   code
3.686 3.323 60.986 3.025 31.058 1.561 19.769 3.757 2.616 1.241 4.760 21.237 0.363 1.347 94.534 n/a n/a n/a
SEA-RAFT: Simple, Efficient, Accurate RAFT for Optical Flow. Wang et al. ECCV 2024
24
MS-RAFT-3D [SF]   code
3.703 3.341 60.571 3.154 26.406 1.495 20.404 3.771 2.669 1.103 4.186 24.146 0.377 1.409 94.835 n/a n/a n/a
J. Schmid, A. Jahedi, N. Berenguel Senn and A. Bruhn. "MS-RAFT-3D: A Multi-scale Architecture for Recurrent Image-based Scene Flow". ICIP 2025.
25
GeoFlow-M  
3.716 3.384 55.951 3.156 26.882 1.733 18.717 3.479 7.322 0.989 3.944 25.914 0.376 1.323 94.689 n/a n/a n/a
Anonymous.
26
RoCo-19-WAFT+  
3.743 3.385 60.188 3.138 28.780 1.631 19.725 3.811 2.723 1.315 4.604 21.733 0.364 1.349 93.776 2.603 7.735 15.163
 
27
RoCo-19-WAFT  
3.805 3.451 59.663 3.208 28.526 1.662 20.025 3.838 3.306 1.351 4.618 22.143 0.384 1.367 93.607 n/a n/a n/a
 
28
GeoFlow  
3.822 3.483 57.178 3.256 27.232 1.786 19.223 3.591 7.327 1.082 4.184 25.768 0.421 1.385 94.514 n/a n/a n/a
Anonymous.
29
SEA-RAFT(S)   code
3.904 3.536 61.951 3.172 34.228 1.662 20.871 3.974 2.855 1.264 4.871 23.378 0.377 1.389 94.182 n/a n/a n/a
SEA-RAFT: Simple, Efficient, Accurate RAFT for Optical Flow. Wang et al. ECCV 2024
30
MegaFlow (w/o ft)  
3.949 3.594 59.743 3.298 30.868 1.737 20.678 4.048 2.435 1.390 4.971 22.587 0.398 1.417 92.595 n/a n/a n/a
Anonymous.
31
StreamFlow   code
4.152 3.790 61.297 3.424 34.304 1.986 20.544 3.986 6.678 1.236 4.381 27.935 0.467 1.424 94.404 n/a n/a n/a
Sun Shangkun, et al. "StreamFlow: Streamlined Multi-Frame Optical Flow Estimation for Video Sequences". NeurIPS 2024.
32
PanMatch (w/o ft)  
4.221 3.868 59.787 3.590 30.324 1.823 22.364 4.313 2.828 1.360 4.298 27.954 0.397 1.456 93.316 n/a n/a n/a
Anonymous.
33
DPFlow (w/o ft)   code
4.277 3.935 58.071 3.660 29.796 1.985 21.615 4.099 6.978 1.653 4.376 25.958 0.479 1.602 93.822 n/a n/a n/a
H. Morimitsu et al. "DPFlow: Adaptive Optical Flow Estimation with a Dual-Pyramid Framework". CVPR 2025.
34
MemFlow   code
4.482 4.119 61.703 3.742 35.115 2.391 20.306 3.934 12.809 1.305 4.437 31.184 0.471 1.416 93.855 n/a n/a n/a
Qiaole Dong, Yanwei Fu. MemFlow: Optical Flow Estimation and Prediction with Memory. CVPR 2024.
35
CroCo-Flow   code
4.565 4.209 60.594 3.848 34.200 2.194 22.501 4.479 5.868 1.225 4.332 33.134 0.498 1.508 93.660 n/a n/a n/a
CroCo v2: Improved Cross-view Completion Pre-training for Stereo Matching and Optical Flow. Weinzaepfel et al. ICCV 2023.
36
MatchFlow (w/o ft)   code
4.584 4.229 60.421 3.858 34.600 2.336 21.591 4.294 8.985 1.407 4.603 31.113 0.453 1.505 93.389 n/a n/a n/a
Tingman Yan and Tao Liu and Xilian Yang and Qunfei Zhao and Zeyang Xia, "MatchAttention: Matching the Relative Positions for High-Resolution Cross-View Matching", arXiv:2510.14260, 2025.
37
RoCo-22_PairBridge  
4.765 4.415 59.840 4.131 30.984 2.835 19.360 4.445 9.627 1.267 4.338 35.177 0.388 1.313 93.704 1.644 5.128 11.618
 
38
BHSFlow  
4.777 4.415 61.824 3.999 36.981 2.284 23.637 4.665 6.483 1.705 4.918 30.094 0.570 1.703 92.695 n/a n/a n/a
Anonymous.
39
RPKNet (w/o ft)   code
4.809 4.460 59.716 4.171 31.198 2.298 23.802 4.478 9.834 1.665 4.757 31.249 0.657 1.756 92.638 n/a n/a n/a
H. Morimitsu, X. Zhu, X. Ji, and X. Yin. "Recurrent Partial Kernel Network for Efficient Optical Flow Estimation". In The 38th Annual AAAI Conference on Artificial Intelligence (AAAI), 2024.
40
FreeFlow-S   code
5.087 4.714 63.896 4.317 36.962 2.627 23.698 4.723 10.618 1.214 6.275 34.277 0.533 1.452 90.196 2.841 9.117 19.638
V. Bargatin, A. Yakovenko, K. Abud, and D. Vatolin. "FreeFlow: A Bias-free Hierarchical Transformer for Optical Flow Estimation", ECCV 2026.
41
StreamFlow-0-shot   code
5.215 4.869 59.550 4.559 32.343 2.865 22.987 4.435 17.059 2.597 4.492 29.067 0.606 1.856 93.253 n/a n/a n/a
Sun Shangkun, et al. "StreamFlow: Streamlined Multi-Frame Optical Flow Estimation for Video Sequences". NeurIPS 2024.
42
Win-Win  
5.371 5.003 63.211 4.624 36.274 2.706 25.531 4.965 11.535 1.318 4.854 40.679 0.475 1.621 92.720 n/a n/a n/a
Win-Win: Training High-Resolution Vision Transformers from Two Windows. Leroy et al. ICLR 2024.
43
MS-RAFT+   code
5.724 5.370 61.497 5.041 33.954 3.047 25.973 4.840 19.150 2.055 5.022 38.315 0.643 2.189 92.888 3.620 12.210 23.390
💡 submitted by spring team | A. Jahedi, M. Luz, M. Rivinius, L. Mehl, and A. Bruhn. "MS-RAFT+: High Resolution Multi-Scale RAFT " International Journal of Computer Vision (IJCV), 2023
44
MemFlow(w/o ft)   code
5.759 5.394 63.348 5.107 32.755 3.293 24.422 4.494 24.990 2.918 4.820 32.071 0.627 2.114 92.253 n/a n/a n/a
Qiaole Dong, Yanwei Fu. MemFlow: Optical Flow Estimation and Prediction with Memory. CVPR 2024.
45
U²Flow (w/o ft)   code
6.323 5.944 66.019 5.381 45.311 3.556 27.257 5.464 19.382 3.407 6.543 30.131 0.608 1.878 89.322 n/a n/a n/a
Xunpei Sun et al., "U²Flow: Uncertainty-Aware Unsupervised Optical Flow Estimation". CVPR 2026.
46
SMURF (w/o ft)  
6.426 6.044 66.516 5.490 45.123 3.580 27.954 5.528 20.068 3.437 6.589 30.988 0.659 2.098 90.449 n/a n/a n/a
submitted by U²Flow author. Stone, Austin, et al. "Smurf: Self-teaching multi-frame unsupervised raft with full-image warping." Proceedings of the IEEE/CVF conference on Computer Vision and Pattern Recognition. 2021.
47
FlowFormer   code
6.510 6.144 64.219 5.766 37.294 3.527 29.084 5.500 21.858 3.381 5.530 35.344 0.723 2.384 91.679 3.770 11.210 21.530
💡 submitted by spring team | Z. Huang, X. Shi, C. Zhang, Q. Wang, K. C. Cheung, H. Qin, J. Dai, and H. Li. "FlowFormer: A Transformer Architecture for Optical Flow." In European Conference on Computer Vision (ECCV), 2022.
48
FlowNet2   code
6.710 6.346 64.061 5.691 48.892 3.711 29.404 6.039 16.908 1.862 5.816 49.693 1.040 2.823 90.907 7.010 11.090 18.840
💡 submitted by spring team | E. Ilg, N. Mayer, T. Saikia, M. Keuper, A. Dosovitskiy, and T. Brox. "FlowNet 2.0: Evolution of Optical Flow Estimation with Deep Networks." In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2017.
49
RoCo-Spring Team Baselines-Optical Flow  
6.765 6.401 64.185 6.006 38.169 3.960 27.987 5.284 29.286 3.071 5.321 41.579 1.493 3.096 91.039 4.360 9.727 18.760
💡 submitted by spring team | Z. Teed, and J. Deng. "RAFT: Recurrent All-Pairs Field Transforms for Optical Flow." In European Conference on Computer Vision (ECCV), 2020.
50
RoCo-45  
6.766 6.401 64.231 6.007 38.165 3.961 27.987 5.284 29.292 3.072 5.322 41.582 1.513 3.100 91.037 n/a n/a n/a
 
51
RAFT   code
6.790 6.426 64.087 5.999 39.481 4.107 27.088 5.250 30.183 3.134 5.301 41.403 1.476 3.198 90.920 5.640 11.470 20.180
💡 submitted by spring team | Z. Teed, and J. Deng. "RAFT: Recurrent All-Pairs Field Transforms for Optical Flow." In European Conference on Computer Vision (ECCV), 2020.
52
RoCo-19-SEARAFT-Finetune  
6.997 6.635 63.889 6.045 46.369 4.372 26.856 5.595 28.304 2.591 5.086 49.021 2.651 3.463 91.430 n/a n/a n/a
Anonymous.
53
GMA   code
7.074 6.699 66.203 6.281 39.892 4.276 28.247 5.614 29.263 3.645 5.389 40.327 0.914 3.079 90.722 4.030 10.950 21.470
💡 submitted by spring team | S. Jiang, D. Campbell, Y. Lu, H. Li, and R. Hartley. "Learning to Estimate Hidden Motions with Global Motion Aggregation." In IEEE/CVF International Conference on Computer Vision (ICCV), 2021.
54
RoCo-19-DEFOM2K-DPFlow [SF]  
7.658 7.341 57.556 6.970 36.128 4.934 28.268 7.794 5.588 1.783 9.506 51.801 0.631 1.778 90.072 n/a n/a n/a
 
55
RoCo-45  
7.762 7.397 65.216 6.983 39.977 4.583 31.817 6.213 31.302 4.288 5.702 42.397 1.765 3.583 89.986 n/a n/a n/a
 
56
RoCo-Spring Team Baselines-Scene Flow [SF]  
9.095 8.711 69.526 8.294 42.225 6.240 30.689 6.637 46.449 5.600 7.349 43.056 1.250 3.839 88.037 n/a n/a n/a
💡 submitted by spring team | Z. Teed, and J. Deng. "RAFT-3D: Scene Flow using Rigid-Motion Embeddings." In CVPR, 2021. and L. Lipson, Z. Teed, and J. Deng. "RAFT-Stereo: Multilevel Recurrent Field Transforms for Stereo Matching." In 3DV, 2021.
57
SplatFlow3D (C+T) + LEAStereo (Things); Two-frame [SF]   code
9.203 8.845 65.532 8.301 46.520 6.319 31.025 7.841 29.901 4.411 7.288 54.475 0.707 2.903 88.424 n/a n/a n/a
 
58
GMFlow   code
10.355 9.935 76.613 9.060 63.949 6.800 37.258 8.952 31.680 5.412 9.901 52.944 0.945 2.952 82.337 2.980 14.680 40.890
💡 submitted by spring team | H. Xu, J. Zhang, J. Cai, H. Rezatofighi, and D. Tao. "GMFlow: Learning Optical Flow via Global Matching." In IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022.
59
RAFT-3D (K) [SF]   code
13.962 13.539 80.464 12.963 55.254 8.932 52.013 11.822 46.479 8.895 14.726 54.283 2.528 6.889 81.267 5.030 17.360 31.200
💡 submitted by spring team | Z. Teed, and J. Deng. "RAFT-3D: Scene Flow using Rigid-Motion Embeddings." In IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2021.
60
PWOC-3D [SF]   code
14.461 14.066 76.598 13.263 64.017 10.002 48.200 12.986 36.884 5.212 15.469 89.110 2.813 5.412 81.149 n/a n/a n/a
R. Saxena, R. Schuster, O. Wasenmuller, and D. Stricker. "PWOC-3D: Deep Occlusion-Aware End-to-End Scene Flow Estimation." In IEEE Intelligent Vehicles Symposium (IV), 2019.
61
M-FUSE (F) [SF]   code
20.374 19.993 80.398 19.382 61.415 15.312 58.668 18.381 50.653 9.734 29.588 84.458 2.948 8.791 76.550 n/a n/a n/a
💡 submitted by spring team | L. Mehl, A. Jahedi, J. Schmalfuss, and A. Bruhn. "M-FUSE: Multi-frame Fusion for Scene Flow Estimation." In IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2023.
62
M-FUSE (K) [SF]   code
20.979 20.600 80.743 19.942 63.882 15.953 59.005 19.500 43.455 10.131 30.966 84.713 2.526 8.480 76.182 3.390 11.170 22.000
💡 submitted by spring team | L. Mehl, A. Jahedi, J. Schmalfuss, and A. Bruhn. "M-FUSE: Multi-frame Fusion for Scene Flow Estimation." In IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2023.
63
CamLiFlow (F) [SF]   code
24.012 23.694 74.084 23.112 61.234 21.203 45.265 21.791 57.763 15.394 33.769 69.710 27.774 17.216 74.082 n/a n/a n/a
💡 submitted by spring team | H. Liu, T. Lu, Y. Xu, J. Liu, W. Li, and L. Chen. "CamLiFlow: Bidirectional Camera-LiDAR Fusion for Joint Optical Flow and Scene Flow Estimation." In IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022.
64
SPyNet   code
29.963 29.661 77.450 28.783 78.766 26.442 56.601 25.832 92.738 24.803 24.201 88.714 4.162 12.866 67.150 4.290 19.180 38.320
💡 submitted by spring team | A. Ranjan, and M. J. Black. "Optical Flow Estimation using a Spatial Pyramid Network." In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2017.
65
RoCo-29  
34.890 34.640 74.307 34.211 63.005 32.435 53.467 35.519 25.333 29.753 36.584 73.291 1.411 5.967 67.517 3.328 17.722 40.303
 
66
RAFT-3D (F) [SF]   code
48.066 47.883 76.933 47.662 64.791 48.200 47.056 48.798 36.942 42.335 68.531 40.645 4.784 34.921 50.686 n/a n/a n/a
💡 submitted by spring team | Z. Teed, and J. Deng. "RAFT-3D: Scene Flow using Rigid-Motion Embeddings." In IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2021.
67
StructFlow  
54.013 53.800 87.530 53.224 86.674 51.569 72.505 51.543 91.556 50.058 53.038 89.747 13.850 38.349 50.149 n/a n/a n/a
Anonymous.
68
CamLiFlow (K) [SF]   code
69.685 69.533 93.651 69.179 90.611 67.381 87.114 67.724 99.492 62.899 79.497 99.903 127.387 60.485 30.635 n/a n/a n/a
💡 submitted by spring team | H. Liu, T. Lu, Y. Xu, J. Liu, W. Li, and L. Chen. "CamLiFlow: Bidirectional Camera-LiDAR Fusion for Joint Optical Flow and Scene Flow Estimation." In IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2022.
69
PWCNet   code
82.265 82.268 81.747 82.069 90.400 82.817 78.090 81.575 92.761 81.402 82.189 89.693 2.288 4.889 45.670 7.250 16.440 31.710
💡 submitted by spring team | D. Sun, X. Yang, M. Liu, and J. Kautz. "PWC-Net: CNNs for Optical Flow Using Pyramid, Warping, and Cost Volume." In IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2018.
70
RoCo-19-WAFT  
n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a n/a