17.2 Optical Flow Field Computation and Analysis

Chapter Contents (Back)
Matching, Optical Flow. Optical Flow. See also Computation for Vector Fields, Flow Fields. See also Moving Image Coding, Compression: Using Vector Fields, Flow Fields.

17.2.1 Point Matching for Optical Flow Computation

Chapter Contents (Back)
Matching, Points. Optical Flow, Features.

Enkelmann, W., Kories, R., Nagel, H.H., and Zimmermann, G.,
An Experimental Investigation of Estimation Approaches for Optical Flow Fields,
MU88(189-226). Discussion of 2 optical flow techniques, feature extraction and match and differential or gradient approaches. The conclusion is that for both methods, certain features provide dominant contributions to the optical flow computation. BibRef 8800

Prager, J.M.[John M.], and Arbib, M.A.[Michael A.],
Computing the Optic Flow: The MATCH Algorithm and Prediction,
CVGIP(24), No. 3, December 1983, pp. 271-304.
WWW Version. Using the smoothness of motion and almost rigid body constraints, the vector field (optic flow field) is computed. The algorithm also assumes that feature points are dense on the objects. Seems to produce results. The work is from 1980. See also Extracting and Labeling Boundary Segments in Natural Scenes. BibRef 8312

Buxton, B.F., Buxton, H.,
Monocular Depth Perception from Optical Flow by Space Time Signal Processing,
RoyalP(B-218), 1983, pp. 27-47. BibRef 8300

Bergmann, H.C.,
Analysis of Different Displacement Estimation Algorithms for Digital Television Signals,
ISPDSA83(215-234). BibRef 8300

Adiv, G.[Gilad],
Determining 3-D Motion and Structure from Optical Flow Generated by Several Moving Objects,
PAMI(7), No. 4, July 1985, pp. 384-401. BibRef 8507
Earlier: DARPA84(113-129). BibRef
Earlier:
Recovering 2-D Motion Parameters in Scenes Containing Multiple Moving Objects,
DARPA83(285-292). BibRef
And: COINSTR 83-11, May, 1983. BibRef
Earlier:
Recovering Motion Parameters in Scenes Containing Multiple Moving Objects,
CVPR83(399-400). Hough. Use of the generalized Hough transform to get the several moving objects. Segment objects based on 2D motion parameters. See also Inherent Ambiguities in Recovering 3-D Motion and Structure from a Noisy Flow Field. BibRef

Williams, L.R., and Hanson, A.R.,
Translating Optical Flow into Token Matches and Depth from Looming,
ICCV88(441-448).
IEEE Abstract. IEEE Top Reference. BibRef 8800
And: COINSTR 88-68, UMass. BibRef
And:
Translating Optical Flow into Token Matches,
DARPA88(970-980). BibRef
And:
Depth from Looming Structure,
DARPA88(1047-1051). Using optical flow to aid the matching of line tokens in motion pairs. BibRef

Castelow, D.A., Murray, D.W., Scott, G.L., Buxton, B.F.,
Matching Canny Edgels to Compute the Principal Components of Optic Flow,
IVC(6), No. 2, May 1988, pp. 129-136.
WWW Version. BibRef 8805

Scott, G.L.,
'Four-Line' Method of Locally Estimating Optic Flow,
IVC(5), No. 2, May 1987, pp. 67-72.
WWW Version. BibRef 8705
Earlier:
Smoothing the Optic Flow Field Under Perspective Projection,
CVPR86(504-509). Treating the flow field as 3D vectors (the projection is what we see) seems to help. BibRef

Thompson, W.B.[William B.], and Kearney, J.K.[Joseph K.],
DIDA: Dynamic Image Disparity Analysis,
TRFinal report for July 1981 to December 1982. AFWAL-TR-83-10A35. The report on an analysis of the needs for disparity analysis systems to be useful. Mostly obvious conclusions (as expected in a report of this type to a government lab) but they implemented several alternative disparity systems. The major point is that various applications have different requirements for accuracy, speed, and density of coverage; and that different methods can give you some of these. The best seems to be the gradient based methods with proper post processing to account for the variations and constrain the output. BibRef 8107

Thompson, W.B., Madarasz, R.L., and Kearney, J.K.,
Temporal Constraints for Estimating Optic Flow,
PRIP82(252-255). Relaxation. Computing optic flow using the constraints that disparity changes slowly over space and over time. Temporal continuity is used in the relaxation process (by remembering the past disparity) to restrict future matches and resolve ambiguities. This leads to a substantial speed up successive frames. BibRef 8200

Thompson, W.B., and Barnard, S.T.,
Lower-level Estimates and Interpretation of Visual Motion,
Computer(14), No. 8, August 1981, pp. 20-28. Optical Flow, Gradient Based. How to use the low-level information in motion analysis - gradients of motion, analysis of feature points (tracking-matching) and a little on 3-D analysis. Optimization versus Hough transform for solutions. BibRef 8108

Fuh, C.S., and Maragos, P.,
Motion Displacement Extimation Using an Affine Model for Image Matching,
OptEng(30), No. 7, July 1991, pp. 881-887. BibRef 9107
Earlier:
Region-Based Optical Flow Estimation,
CVPR89(130-135).
IEEE Abstract. IEEE Top Reference. Determine the flow of primitive tokens (regions) by matching. Then smooth the field. Basic. BibRef

Fuh, C.S., Maragos, P., Vinlent, M.,
Visual Motion Correspondence by Region-Based Approaches,
ACCV93(784-789). Region features only. BibRef 9300

Duncan, J.H., and Chou, T.C.,
On the Detection of Motion and the Computation of Optical Flow,
PAMI(14), No. 3, March 1992, pp. 346-352.
IEEE Abstract. IEEE Top Reference.
WWW Version. BibRef 9203
Earlier:
Temporal Edges: The Detection of Motion and the Computation of Optical Flow,
ICCV88(374-382).
IEEE Abstract. IEEE Top Reference. Use of spatio-temporal filter approach for optical flow. BibRef

Cornilleau-Pérès, V.[Valérie], Droulez, J.[Jacques],
Velocity-Based Correspondence in Stereokinetic Images,
CVGIP(58), No. 2, September 1993, pp. 137-146.
WWW Version. BibRef 9309
Earlier:
Stereo correspondence from optic flow,
ECCV90(326-330).
WWW Version. 9004 BibRef

Shapiro, V., Backalov, I., Kavardjikov, V.,
Motion Analysis Via Interframe Point Correspondence Establishment,
IVC(13), No. 2, March 1995, pp. 111-118.
WWW Version. BibRef 9503

Huang, Y., Zhuang, X.H., Yang, C.S.,
Two Block-Based Motion Compensation Methods for Video Coding,
CirSysVideo(6), No. 1, February 1996, pp. 123-126.
IEEE Top Reference. BibRef 9602

Huang, Y.[Yan], Zhuang, X.H.[Xin-Hua],
An Adaptively Refined Block Matching Algorithm for Motion Compensated Video Coding,
CirSysVideo(5), No. 1, February 1995, pp. 56-59.
IEEE Top Reference. BibRef 9502
Earlier:
Motion-partitioned adaptive block matching for video compression,
ICIP95(I: 554-557).
WWW Version. 9510 BibRef

Wu, Q.X.,
A Correlation-Relaxation-Labeling Framework for Computing Optical Flow: Template Matching from a New Perspective,
PAMI(17), No. 9, September 1995, pp. 843-853.
IEEE Abstract. IEEE Top Reference.
WWW Version. Template Matching. Clouds. Low contrast, large displacements, and non-rigid motions. BibRef 9509

Chaudhury, K., Mehrotra, R.,
A Trajectory-Based Computational Model for Optical-Flow Estimation,
RA(11), No. 5, October 1995, pp. 733-741. BibRef 9510

Bannehr, L., Rohn, M., Warnecke, G.,
A Functional Analytic Method to Derive Displacement Vector-Fields from Satellite Image Sequences,
JRS(17), No. 2, January 20 1996, pp. 383-392. BibRef 9601

Hang, H.M., Leonardi, R., Haskell, B.G., Schmidt, R.L., Bheda, H., Othmer, J.,
Digital HDTV Compression Using Parallel Motion-Compensated Transform Coders,
CirSysVideo(1), No. 2, June 1991, pp. 210-221.
IEEE Top Reference. BibRef 9106

Reusens, E.,
Joint Optimization of Representation Model and Frame Segmentation for Generic Video Compression,
SP(46), No. 1, September 1995, pp. 105-117. BibRef 9509

Pan, J.N., Shi, Y.Q., Shu, C.Q.,
Correlation Feedback Technique in Optical Flow Determination,
IP(7), No. 7, July 1998, pp. 1061-1067.
WWW Version. 9807 See also Unified Optical-Flow Field Approach To Motion Analysis from a Sequence of Stereo Images. BibRef

Chen, J.L.[Jau-Ling], Chen, P.Y.[Pei-Yin],
An efficient gray search algorithm for the estimation of motion vectors,
SMC-C(31), No. 2, May 2001, pp. 242-248.
IEEE Top Reference. 0109 BibRef

Jou, J.M., Chen, P.Y., Sun, J.M.,
The Gray Prediction Search Algorithm for Block Motion Estimation,
CirSysVideo(9), No. 6, September 1999, pp. 843.
IEEE Top Reference. BibRef 9909

Imiya, A.[Atsushi], Iwawaki, K.[Keisuke],
Voting method for the detection of subpixel flow field,
PRL(24), No. 1-3, January 2003, pp. 197-214.
HTML Version. 0211 See also Voting method for planarity and motion detection. BibRef

Sun, C.M.[Chang-Ming],
Fast optical flow using 3D shortest path techniques,
IVC(20), No. 13-14, December 2002, pp. 981-991.
WWW Version. 0212 BibRef
Earlier:
Fast Optical Flow Using Cross Correlation and Shortest-Path Techniques,
DICTA99(143-148). Shortest-Path Techniques, fast correlation See also Circular shortest path in images. See also Fast Stereo Matching Using Rectangular Subregioning and 3D Maximum-Surface Techniques. BibRef

Tordoff, B.J.[Ben J.], Murray, D.W.,
Guided-MLESAC: Faster Image Transform Estimation by Using Matching Priors,
PAMI(27), No. 10, October 2005, pp. 1523-1535.
WWW Version. 0509 BibRef
Earlier:
Guided Sampling and Consensus for Motion Estimation,
ECCV02(I: 82 ff.).
HTML Version.
HTML Version. 0205Maximum-likelihood estimation by random sampling consensus Build on MLESAC ( See also MLESAC: A New Robust Estimator with Application to Estimating Image Geometry. ). BibRef

Wills, J.[Josh], Agarwal, S.[Sameer], Belongie, S.[Serge],
A Feature-based Approach for Dense Segmentation and Estimation of Large Disparity Motion,
IJCV(68), No. 2, June 2006, pp. 125-143.
WWW Version. 0606 BibRef

Wills, J.[Josh], Belongie, S.[Serge],
A Feature-Based Approach for Determining Dense Long Range Correspondences,
ECCV04(Vol III: 170-182).
WWW Version. 0405two stage process in which a planar model is used to get an approximation for the segmentation and the gross motion, and then a spline is used to refine the fit. BibRef

Roth, S.[Stefan], Black, M.J.[Michael J.],
On the Spatial Statistics of Optical Flow,
IJCV(74), No. 1, August 2007, pp. 33-50.
WWW Version.
PDF Version. 0705 BibRef
Earlier: ICCV05(I: 42-49).
WWW Version. 0510 Award, Marr Prize, HM. Training fields from range images and moving camera. BibRef

Roth, S.[Stefan], Black, M.J.[Michael J.],
Specular Flow and the Recovery of Surface Structure,
CVPR06(II: 1869-1876).
WWW Version. 0606 BibRef


Tandjung, S.S., Soon, S.H.[Seah Hock], Kemao, Q.,
Block Motion Model for Optical Flow with Smoothness Prior Function,
ICARCV06(1-6).
WWW Version. 0612 BibRef

Tandjung, S.S., Gunawan, T., Chong, M.,
Motion Estimation Using Adaptive Blocksize Observation Model and Efficient Multiscale Regularization,
ICIP00(Vol II: 566-569).
IEEE Abstract. IEEE Top Reference. 0008 BibRef

Ben-Ari, R.[Rami], Sochen, N.A.[Nir A.],
A General Framework and New Alignment Criterion for Dense Optical Flow,
CVPR06(I: 529-536).
WWW Version. 0606 See also Geometric Approach for Regularization of the Data Term in Stereo-Vision, A. BibRef

Chapter on Optical Flow Field Computations and Use continues in
Optical Flow -- Matching Using Areas .


Last update:Aug 27, 2008 at 19:16:50