Method

FastFlowNet: A Lightweight Network for Fast Optical Flow Estimation [FastFlowNet]
https://github.com/ltkong218/FastFlowNet

Submitted on 12 Sep. 2020 14:38 by
Lingtong Kong (SJTU)

Running time:0.01 s
Environment:NVIDIA GTX 1080 Ti

Method Description:
Dense optical flow estimation plays a key role in
many robotic vision tasks. It has been predicted
with satisfying accuracy than traditional methods
with advent of deep learning. However, current
networks often occupy large number of parameters
and require heavy computation costs. These
drawbacks have hindered applications on power- or
memory-constrained mobile devices. To deal with
these challenges, in this paper, we dive into
designing efficient structure for fast and
accurate optical flow prediction. Our proposed
FastFlowNet works in the well-known coarse-to-fine
manner with innovations of head enhanced pooling
pyramid (HEPP) feature extractor, center dense
dilated correlation (CDDC) layer and efficient
shuffle block decoder (SBD). Experiments on both
synthetic Sintel and real-world KITTI datasets
demonstrate the effectiveness of proposed
approaches, which consumes only 1/10 computation
of comparable networks to get 90% of their
performance.
Parameters:
FastFlowNet
Latex Bibtex:
@inproceedings{kong2021fastflownet,
author={Lingtong Kong and Chunhua Shen and Jie
Yang},
booktitle={IEEE International Conference on
Robotics and Automation (ICRA)},
title={FastFlowNet: A Lightweight Network for Fast
Optical Flow Estimation},
year={2021},}

Detailed Results

This page provides detailed results for the method(s) selected. For the first 20 test images, the percentage of erroneous pixels is depicted in the table. We use the error metric described in Object Scene Flow for Autonomous Vehicles (CVPR 2015), which considers a pixel to be correctly estimated if the disparity or flow end-point error is <3px or <5% (for scene flow this criterion needs to be fulfilled for both disparity maps and the flow map). Underneath, the left input image, the estimated results and the error maps are shown (for disp_0/disp_1/flow/scene_flow, respectively). The error map uses the log-color scale described in Object Scene Flow for Autonomous Vehicles (CVPR 2015), depicting correct estimates (<3px or <5% error) in blue and wrong estimates in red color tones. Dark regions in the error images denote the occluded pixels which fall outside the image boundaries. The false color maps of the results are scaled to the largest ground truth disparity values / flow magnitudes.

Test Set Average

Error Fl-bg Fl-fg Fl-all
All / All 11.20 11.30 11.22
All / Est 11.20 11.30 11.22
Noc / All 6.29 7.78 6.56
Noc / Est 6.29 7.78 6.56
This table as LaTeX

Test Image 0

Error Fl-bg Fl-fg Fl-all
All / All 4.83 33.67 8.79
All / Est 4.83 33.67 8.79
Noc / All 4.32 33.67 8.77
Noc / Est 4.32 33.67 8.77
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 1

Error Fl-bg Fl-fg Fl-all
All / All 5.97 15.29 7.01
All / Est 5.97 15.29 7.01
Noc / All 5.65 15.29 6.84
Noc / Est 5.65 15.29 6.84
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 2

Error Fl-bg Fl-fg Fl-all
All / All 7.63 6.92 7.59
All / Est 7.63 6.92 7.59
Noc / All 6.07 6.92 6.12
Noc / Est 6.07 6.92 6.12
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 3

Error Fl-bg Fl-fg Fl-all
All / All 21.09 42.32 23.05
All / Est 21.09 42.32 23.05
Noc / All 18.19 31.09 19.30
Noc / Est 18.19 31.09 19.30
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 4

Error Fl-bg Fl-fg Fl-all
All / All 10.84 16.09 11.71
All / Est 10.84 16.09 11.71
Noc / All 8.11 14.46 9.29
Noc / Est 8.11 14.46 9.29
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 5

Error Fl-bg Fl-fg Fl-all
All / All 6.66 0.07 6.07
All / Est 6.66 0.07 6.07
Noc / All 4.44 0.07 3.99
Noc / Est 4.44 0.07 3.99
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 6

Error Fl-bg Fl-fg Fl-all
All / All 9.44 0.23 8.47
All / Est 9.44 0.23 8.47
Noc / All 6.27 0.23 5.55
Noc / Est 6.27 0.23 5.55
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 7

Error Fl-bg Fl-fg Fl-all
All / All 2.14 30.49 7.69
All / Est 2.14 30.49 7.69
Noc / All 2.14 31.08 7.55
Noc / Est 2.14 31.08 7.55
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 8

Error Fl-bg Fl-fg Fl-all
All / All 1.67 4.67 2.22
All / Est 1.67 4.67 2.22
Noc / All 1.67 4.67 2.22
Noc / Est 1.67 4.67 2.22
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 9

Error Fl-bg Fl-fg Fl-all
All / All 1.70 20.81 6.58
All / Est 1.70 20.81 6.58
Noc / All 1.70 20.81 6.58
Noc / Est 1.70 20.81 6.58
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 10

Error Fl-bg Fl-fg Fl-all
All / All 4.20 4.93 4.37
All / Est 4.20 4.93 4.37
Noc / All 3.51 4.93 3.88
Noc / Est 3.51 4.93 3.88
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 11

Error Fl-bg Fl-fg Fl-all
All / All 4.55 3.86 4.43
All / Est 4.55 3.86 4.43
Noc / All 4.40 3.87 4.29
Noc / Est 4.40 3.87 4.29
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 12

Error Fl-bg Fl-fg Fl-all
All / All 5.60 2.56 5.40
All / Est 5.60 2.56 5.40
Noc / All 3.41 2.56 3.34
Noc / Est 3.41 2.56 3.34
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 13

Error Fl-bg Fl-fg Fl-all
All / All 5.98 3.83 5.71
All / Est 5.98 3.83 5.71
Noc / All 2.22 1.13 2.09
Noc / Est 2.22 1.13 2.09
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 14

Error Fl-bg Fl-fg Fl-all
All / All 7.85 4.44 7.79
All / Est 7.85 4.44 7.79
Noc / All 3.99 4.44 4.00
Noc / Est 3.99 4.44 4.00
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 15

Error Fl-bg Fl-fg Fl-all
All / All 15.25 5.52 14.37
All / Est 15.25 5.52 14.37
Noc / All 11.01 5.52 10.38
Noc / Est 11.01 5.52 10.38
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 16

Error Fl-bg Fl-fg Fl-all
All / All 21.40 9.10 19.59
All / Est 21.40 9.10 19.59
Noc / All 10.52 9.10 10.27
Noc / Est 10.52 9.10 10.27
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 17

Error Fl-bg Fl-fg Fl-all
All / All 16.89 6.24 15.78
All / Est 16.89 6.24 15.78
Noc / All 8.95 6.24 8.60
Noc / Est 8.95 6.24 8.60
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 18

Error Fl-bg Fl-fg Fl-all
All / All 27.15 99.79 61.66
All / Est 27.15 99.79 61.66
Noc / All 15.61 99.32 38.70
Noc / Est 15.61 99.32 38.70
This table as LaTeX

Input Image

Flow Result

Flow Error


Test Image 19

Error Fl-bg Fl-fg Fl-all
All / All 15.23 2.28 13.76
All / Est 15.23 2.28 13.76
Noc / All 7.52 2.28 6.76
Noc / Est 7.52 2.28 6.76
This table as LaTeX

Input Image

Flow Result

Flow Error




eXTReMe Tracker