Visual Results by Method

Results for methods appear here after users upload them and approve them for public display.


Return to numerical results table



Metrics


EPE Endpoint error over the complete frames
EPE matched Endpoint error over regions that remain visible in adjacent frames
EPE unmatched Endpoint error over regions that are visible only in one of two adjacent frames
d0-10 Endpoint error over regions closer than 10 pixels to the nearest occlusion boundary
d10-60 Endpoint error over regions between 10 and 60 pixels apart from the nearest occlusion boundary
d60-140 Endpoint error over regions between 60 and 140 pixels apart from the nearest occlusion boundary
s0-10 Endpoint error over regions with velocities lower than 10 pixels per frame
s10-40 Endpoint error over regions with velocities between 10 and 40 pixels per frame
s40+ Endpoint error over regions with velocities larger than 40 pixels per frame
All error images are normalized to be in the range between 0 and 1.

References
[1]
2bit-BM-tele
Rui Xu & David Taubman, Robust Dense Block-Based Motion Estimation Using a Two-Bit Transform on a Laplacian Pyramid, ICIP 2013 +telescopic search
[2]
A-A
Anonymous.
[3]
ADLAB-PRFlow
Anonymous.
[4]
ADW
Anonymous.
[5]
ADW-Net
Anonymous. ADW-Net,20201024 submit
[6]
AGF-Flow
Anonymous. AGF
[7]
AGIF+OF
Anonymous. Signal Processing 2015
[8]
ALNF
Anonymous. ALNF
[9]
ARFlow
Liang Liu, Jiangning Zhang, Ruifei He, Yong Liu, Yabiao Wang, Ying Tai, et al. Learning by Analogy: Reliable Supervision from Transformations for Unsupervised Optical Flow Estimation. CVPR 2020.
[10]
ARFlow-mv
Liang Liu, Jiangning Zhang, Ruifei He, Yong Liu, Yabiao Wang, Ying Tai, et al. Learning by Analogy: Reliable Supervision from Transformations for Unsupervised Optical Flow Estimation. CVPR 2020.
[11]
ARFlow-mv-ft
Liang Liu, Jiangning Zhang, Ruifei He, Yong Liu, Yabiao Wang, Ying Tai, et al. Learning by Analogy: Reliable Supervision from Transformations for Unsupervised Optical Flow Estimation. CVPR 2020.
[12]
AVG_FLOW_ROB
Average flow field of ROB2018 training set. No image information used!
[13]
AggregFlow
D. Fortun, P. Bouthemy and C. Kervrann. Aggregation of local parametric candidates and exemplar-based occlusion handling for optical flow. CVIU 2016
[14]
AnisoHuber.L1
M. Werlberger, W. Trobin, T. Pock, A. Wedel, D. Cremers, and H. Bischof. Anisotropic Huber-L1 optical flow. BMVC 2009.
[15]
AnyFlow
Anonymous. PAMI pending review
[16]
AtrousFlow
Anonymous. Real-time dense optical flow using CUDA
[17]
AugFNG_ROB
Anonymous.
[18]
AutoScaler+
Anonymous. AutoScaler+
[19]
BASELINE-Mean
[20]
BASELINE-zero
[21]
BOOM+PF.XY
Fast optical flow method (0.54s single core i7 @3.1GHz) that employs the permeability filter to interpolate NNF correspondences. The NNF is an improved version of CPM that uses a new binary descriptor termed BOOM instead of SiftFlow in order to efficiently compute matching costs. Publication: Towards Edge-Aware Spatio-Temporal Filtering in Real-Time. M. Schaffner, F. Scheidegger, L. Cavigelli, H. Kaeslin, L. Benini and A. Smolic. Accepted for publication in Trans. on Image Processing (TIP), 2017. DOI: 10.1109/TIP.2017.2757259
[22]
BOOM+PF.XYT
Fast optical flow method (0.59s single core i7 @3.1GHz) that employs the permeability filter to interpolate NNF correspondences. The NNF is an improved version of CPM that uses a new binary descriptor termed BOOM instead of SiftFlow in order to efficiently compute matching costs. Publication: Towards Edge-Aware Spatio-Temporal Filtering in Real-Time. M. Schaffner, F. Scheidegger, L. Cavigelli, H. Kaeslin, L. Benini and A. Smolic. Accepted for publication in Trans. on Image Processing (TIP), 2017. DOI: 10.1109/TIP.2017.2757259
[23]
Back2FutureFlow_UFO
J. Janai, F. Güney, A. Ranjan, M. Black and A. Geiger. Unsupervised Learning of Multi-Frame Optical Flow with Occlusions. ECCV, 2018.
[24]
C-2px
Anonymous.
[25]
C-RAFT_RVC
RVC 2020 submission.
[26]
CNet
Anonymous.
[27]
COF
H. Zimmer, A. Bruhn, J. Weickert, L. Valgaerts, A. Salgado, B. Rosenhahn, and H.-P. Seidel. Complementary optic flow. EMMCVPR 2009.
[28]
COF_2019
H. Zimmer, A. Bruhn, J. Weickert, L. Valgaerts, A. Salgado, B. Rosenhahn, and H.-P. Seidel. Complementary optic flow. EMMCVPR 2009.
[29]
COMBO
Anonymous.
[30]
CPM-Flow
Yinlin Hu, Rui Song, Yunsong Li. Efficient Coarse-to-Fine PatchMatch for Large Displacement Optical Flow. CVPR 2016.
[31]
CPM2
Coarse-to-fine PatchMatch for dense correspondence, Yunsong Li, Yinlin Hu, Rui Song, Peng Rao, Yangli Wang, submitted to T-CSVT
[32]
CPM_AUG
Anonymous. CVPR 18 submission #1939
[33]
CPNFlow
Conditional Prior Networks for Optical Flow, Yanchao Yang, Stefano Soatto; The European Conference on Computer Vision (ECCV), 2018, pp. 271-287
[34]
CRAFT
Anonymous. Cross-Attentional Flow Transformer
[35]
CVPR-1235
Anonymous.
[36]
Channel-Flow
L. Sevilla-Lara, D. Sun, E. Learned-Miller, and M. Black. Optical Flow Estimation with Channel Constancy. ECCV, pages 423-438, 2014.
[37]
Classic++
D. Sun, S. Roth, and M. Black. Secrets of optical flow estimation and their principles. CVPR, pages 2432-2439, 2010.
[38]
Classic+NL
D. Sun, S. Roth, and M. Black. Secrets of optical flow estimation and their principles. CVPR, pages 2432-2439, 2010.
[39]
Classic+NL-fast
D. Sun, S. Roth, and M. Black. Secrets of optical flow estimation and their principles. CVPR, pages 2432-2439, 2010.
[40]
Classic+NLP
D. Sun, S. Roth, and M. Black. A Quantitative Analysis of Current Practices in Optical Flow Estimation and The Principles behind Them. International Journal of Computer Vision (IJCV), 106(2):115-137, 2014.
[41]
CoT-AMFlow
H. Wang, R. Fan, and M. Liu. CoT-AMFlow: Adaptive Modulation Network with Co-Teaching Strategy for Unsupervised Optical Flow Estimation, CoRL 2020.
[42]
CompactFlow
Anonymous. ICCV submission.
[43]
CompactFlow-woscv
Anonymous.
[44]
ComponentFusion
Anonymous. Fast Optical Flow by Component Fusion. Submitted to ECCV 2014. Paper ID 941.
[45]
ContFusion
M. Stoll, D. Maurer, and A. Bruhn. Variational Large Displacement Optical Flow without Feature Matches. EMMCVPR 2017.
[46]
ContinualFlow_ROB
Michal Neoral, Jan Šochman and Jiří Matas. Continual Occlusions and Optical Flow Estimation, ACCV 2018
[47]
CosTR
Anonymous.
[48]
DA_opticalflow
Anonymous.
[49]
DCFlow
Jia Xu, René Ranftl, Vladlen Koltun. Accurate Optical Flow via Direct Cost Volume Processing. CVPR 2017.
[50]
DCFlow+KF
W. Bao, Y. Xiao, L. Chen, and Z. Gao. KalmanFlow 2.0: Efficient Video Optical Flow Estimation via Context-Aware Kalman Filtering. submitted to TIP.
[51]
DCFlow+KF2
W. Bao, Y. Xiao, L. Chen, and Z. Gao. KalmanFlow 2.0: Efficient Video Optical Flow Estimation via Context-Aware Kalman Filtering. submitted to TIP.
[52]
DCVNet
Anonymous. 0.014s with a GTX 1080ti GPU.
[53]
DDCNet_B1_ft-sintel
DDCNet B1 finetuned on Sintel
[54]
DDCNet_Multires_ft_sintel
DDCNet Multires fine tuned on Sintel
[55]
DDCNet_stacked2
Anonymous. two times down-sampling of feature maps, 128 filters each layer, fine-tuned on sintel
[56]
DDFlow
Pengpeng Liu, Irwin King, Michael R. Lyu, Jia Xu.  DDFlow: Learning Optical Flow with Unlabeled Data Distillation.  Thirty-Third AAAI Conference on Artificial Intelligence (AAAI), Jan 2019
[57]
DF
Nelson Monzón, Agustín Salgado and Javier Sánchez. Regularization Strategies for Discontinuity-Preserving Optical Flow Methods. IEEE Transactions on Image Processing, Vol 25(4), pp. 1580 - 1591, 2016
[58]
DF-Auto
Nelson Monzón, Agustín Salgado and Javier Sánchez. Regularization Strategies for Discontinuity-Preserving Optical Flow Methods. IEEE Transactions on Image Processing, Vol 25(4), pp. 1580 - 1591, 2016
[59]
DF-Beta
Nelson Monzón, Agustín Salgado and Javier Sánchez. Regularization Strategies for Discontinuity-Preserving Optical Flow Methods. IEEE Transactions on Image Processing, Vol 25(4), pp. 1580 - 1591, 2016
[60]
DICL-Flow
Displacement-Invariant Matching Cost Learning for Accurate Optical Flow Estimation (NeurIPS 2020)
[61]
DICL-Flow+
Displacement-Invariant Matching Cost Learning for Accurate Optical Flow Estimation (NeurIPS 2020)
[62]
DICL_update
Displacement-Invariant Matching Cost Learning for Accurate Optical Flow Estimation (NeurIPS 2020)
[63]
DIP
Anonymous. Deep Inverse Patch Match for High-Resolution Optical Flow
[64]
DIP-Flow
D. Maurer, M. Stoll, A. Bruhn: Directional Priors for Multi-Frame Optical Flow. BMVC 2018.
[65]
DIS-Fast
T. Kroeger, R. Timofte, D. Dai, L. Van Gool, Fast Optical Flow using Dense Inverse Search. ECCV 2016. Run-time: 0.023 s (20ms preprocessing, 3ms flow computation). Using operating point 2 of the paper.
[66]
DMF_ROB
Baseline submission for ROB flow challenge: Deepflow with Deepmatching as prior
[67]
DPCTF
Anonymous. Detail Preserving Coarse-to-Fine Matching for Stereo Matching and Optical Flow
[68]
DSPyNet
Zefeng Sun and Hanli Wang, Deeper Spatial Pyramid Network with Refined Up-Sampling for Optical Flow Estimation, 2018 Pacific Rim Conference on Multimedia (PCM'18), LNCS 11164, pp. 492-501, 2018.
[69]
DSPyNet+ft
Zefeng Sun and Hanli Wang, Deeper Spatial Pyramid Network with Refined Up-Sampling for Optical Flow Estimation, 2018 Pacific Rim Conference on Multimedia (PCM'18), LNCS 11164, pp. 492-501, 2018.
[70]
Data-Flow
Anonymous. CSAD data cost + second order smoothness
[71]
Deep+R
B. Drayer and T. Brox. Combinatorial Regularization of Descriptor Matching for Optical Flow Estimation. BMVC 2015
[72]
Deep-EIP
Anonymous. End-2-End learning for energy based inpainting of optical flow. ACCV submission 572
[73]
DeepDiscreteFlow
F. Güney and A. Geiger. Deep Discrete Flow. Asian Conference on Computer Vision (ACCV) 2016.
[74]
DeepFlow
P. Weinzaepfel, J. Revaud, Z. Harchaoui and C. Schmid. DeepFlow: Large displacement optical flow with deep matching. ICCV 2013.
[75]
DeepFlow2
J. Revaud, P. Weinzaepfel, Z. Harchaoui and C. Schmid. DeepMatching: Hierarchical Deformable Dense Matching. Submitted to IJCV.
[76]
DefFlowP
Anonymous.
[77]
Deformable_RAFT
Anonymous. RAFT with deformable
[78]
Devon
Anonymous. CVPR submission #1906
[79]
DictFlowS
Anonymous.
[80]
DiscreteFlow
M. Menze, C. Heipke and A. Geiger. Discrete Optimization for Optical Flow. GCPR 2015
[81]
DiscreteFlow+OIR
D. Maurer, M. Stoll, A. Bruhn: Order-Adaptive and Illumination Aware Variational Optical Flow Refinement. BMVC 2017.
[82]
DistillFlow
Anonymous. Unsupervise result
[83]
DistillFlow+ft
Anonymous. Supervised result.
[84]
EPIflow
Deep Epipolar Flow
[85]
EPMNet
Anonymous. The high accuracy for both small and large motion estimation are mainly cause by two contributions: firstly, we present and implement an edge preserve patch match (EPM) layer that propagates self-similarity patterns in addition to offsets. The accuracy of optical flow prediction has greatly improved by this method. Secondly, we develop a course-to-fine network architecture to tackle large displacement estimation and introduce a residual flow method to solve small displacement estimation.
[86]
EPPM
L. Bao, Q. Yang, and H. Jin. Fast Edge-Preserving PatchMatch for Large Displacement Optical Flow. CVPR 2014.
[87]
ER-FLOW2
Anonymous. Adjusted ERFlow
[88]
ERFlow
Anonymous.
[89]
EgFlow-cl
Anonymous. edge-guided, small parameter optical flow network based on CNN
[90]
EpicFlow
J. Revaud, P. Weinzaepfel, Z. Harchaoui and C. Schmid. EpicFlow: Edge-Preserving Interpolation of Correspondences for Optical Flow. CVPR 2015.
[91]
F2PD_JJN
Anonymous. With Deformable Convolutional Networks
[92]
F3-MPLF
Sathya N. Ravi, Yunyang Xiong, Lopamudra Mukherjee, Vikas Singh: Filter Flow made Practical: Massively Parallel and Lock-Free -- CVPR 2017
[93]
FALDOI
Roberto P.Palomares, Enric Meinhardt-Llopis, Coloma Ballester and Gloria Haro. FALDOI: A new minimization strategy for large displacement variational optical flow. To appear in JMIV
[94]
FAOP-Flow
Anonymous.
[95]
FC-2Layers-FF
D. Sun, J. Wulff, E. Sudderth, H. Pfister, M.J. Black. A fully connected layered model of foreground and background flow. CVPR 2013
[96]
FDFlowNet
Lingtong Kong and Jie Yang. FDFlowNet: Fast Optical Flow Estimation using a Deep Lightweight Network, ICIP 2020.
[97]
FF++_ROB
Accurate matching and robust interpolation. KITTI parameters as given by the paper. Edge detector trained on KITTI + Sintel + BSDS500.
[98]
FGI
Y. Li, D. Min, M. N. Do, and J. Lu. Fast Guided Global Interpolation for Depth and Motion. ECCV 2016
[99]
FPCR-Net
Anonymous.
[100]
FPCR-Net2
Anonymous.
[101]
FastFlow
Anonymous.
[102]
FastFlow2
Anonymous.
[103]
FastFlowNet
Lingtong Kong, Chunhua Shen and Jie Yang. FastFlowNet: A Lightweight Network for Fast Optical Flow Estimation, ICRA 2021.
[104]
FastFlowNet-ft+
Anonymous.
[105]
Flow1D
Haofei Xu, Jiaolong Yang, Jianfei Cai, Juyong Zhang, Xin Tong. High-Resolution Optical Flow from 1D Attention and Correlation. ICCV 2021, Oral
[106]
FlowFields
C. Bailer, B. Tae