209:
difference) and then summarise over a local image region (block base motion and filter based motion). An emerging type of matching criteria summarises a local image region first for every pixel location (through some feature transform such as
Laplacian transform), compares each summarised pixel and summarises over a local image region again. Some matching criteria have the ability to exclude points that do not actually correspond to each other albeit producing a good matching score, others do not have this ability, but they are still matching criteria.
20:
230:
208:
It can be argued that almost all methods require some kind of definition of the matching criteria. The difference is only whether you summarise over a local image region first and then compare the summarisation (such as feature based methods), or you compare each pixel first (such as squaring the
113:
between two images or video frames. The points that correspond to each other in two views (images or frames) of a real scene or object are "usually" the same point in that scene or on that object. Before we do motion estimation, we must define our measurement of correspondence, i.e., the matching
220:
is a technique used in computer vision and image processing to estimate the motion between two images or frames. It assumes that the motion can be modeled as an affine transformation (translation + rotation + zooming), which is a linear transformation followed by a translation.
114:
metric, which is a measurement of how similar two image points are. There is no right or wrong here; the choice of matching metric is usually related to what the final estimated motion is used for as well as the optimisation strategy in the estimation process.
151:
The methods for finding motion vectors can be categorised into pixel based methods ("direct") and feature based methods ("indirect"). A famous debate resulted in two papers from the opposing factions being produced to try to establish a conclusion.
193:, and match corresponding features between frames, usually with a statistical function applied over a local or global area. The purpose of the statistical function is to remove matches that do not correspond to the actual motion.
83:. The motion vectors may be represented by a translational model or many other models that can approximate the motion of a real video camera, such as rotation and translation in all three dimensions and zoom.
142:
motion vector: a two-dimensional vector used for inter prediction that provides an offset from the coordinates in the decoded picture to the coordinates in a reference picture.
569:
43:-plane of the image, combined with a lateral movement to the lower-right. This is a visualization of the motion estimation performed in order to compress an MPEG movie.
599:
521:
893:
594:
41:
826:
559:
403:
554:
589:
657:
75:
happens in three dimensions (3D) but the images are a projection of the 3D scene onto a 2D plane. The motion vectors may relate to the whole image (
674:
694:
440:
338:
281:
604:
574:
564:
544:
514:
373:
940:
759:
584:
614:
471:
308:
764:
749:
507:
491:
328:
127:
in a picture based on the position of this macroblock (or a similar one) in another picture, called the reference picture.
858:
789:
640:
609:
356:
754:
935:
930:
806:
719:
430:
250:
910:
784:
298:
161:
217:
846:
836:
579:
293:
883:
851:
630:
303:
254:
110:
390:
888:
699:
645:
105:
811:
794:
774:
744:
245:
Applying the motion vectors to an image to synthesize the transformation to the next image is called
234:
816:
679:
463:
264:. Almost all video coding standards use block-based motion estimation and compensation such as the
246:
831:
739:
724:
684:
99:
68:
260:
As a way of exploiting temporal redundancy, motion estimation and compensation are key parts of
229:
19:
769:
714:
666:
467:
457:
436:
334:
261:
166:
878:
841:
689:
549:
190:
131:
52:
873:
821:
530:
360:
48:
416:
494:." 2013 IEEE/RSJ International Conference on Intelligent Robots and Systems. IEEE, 2013.
898:
868:
779:
704:
635:
487:
419:', in IEEE Transactions on Image Processing, vol. 25, no. 3, pp. 1095-1108, March 2016.
26:
79:) or specific parts, such as rectangular blocks, arbitrary shaped patches or even per
924:
863:
63:
that describe the transformation from one 2D image to another; usually from adjacent
801:
417:
Motion
Estimation Based on Mutual Information and Adaptive Multi-scale Thresholding
175:
93:
709:
64:
123:
284:, a 3D model of a scene is reconstructed using images from a moving camera.
353:
197:
72:
429:
Borko Furht; Joshua
Greenberg; Raymond Westwater (6 December 2012).
499:
228:
80:
269:
265:
503:
374:"Latest working draft of H.264/MPEG-4 AVC on hhi.fraunhofer.de"
196:
Statistical functions that have been successfully used include
91:
More often than not, the term motion estimation and the term
97:
are used interchangeably. It is also related in concept to
406:, ICCV Workshop on Vision Algorithms, pages 267-277, 1999.
393:, ICCV Workshop on Vision Algorithms, pages 278-294, 1999
391:
Feature Based
Methods for Structure and Motion Estimation
459:
Understanding
Digital Cinema: A Professional Handbook
109:. In fact all of these terms refer to the process of
29:
23:
Motion vectors that result from a movement into the
665:
656:
623:
537:
432:Motion Estimation Algorithms for Video Compression
35:
415:Rui Xu, David Taubman & Aous Thabit Naman, '
515:
257:, because the coding is performed in blocks.
8:
662:
522:
508:
500:
435:. Springer Science & Business Media.
28:
354:Latest working draft of H.264/MPEG-4 AVC
18:
16:Process used in video coding/compression
389:Philip H.S. Torr and Andrew Zisserman:
319:
675:3D reconstruction from multiple images
695:Simultaneous localization and mapping
282:simultaneous localization and mapping
204:Additional note on the categorization
7:
492:Dense visual SLAM for RGB-D cameras
486:Kerl, Christian, Jürgen Sturm, and
760:Automatic number-plate recognition
14:
309:Scale-invariant feature transform
268:series including the most recent
765:Automated species identification
750:Audio-visual speech recognition
249:. It is most easily applied to
595:Recognition and categorization
67:in a video sequence. It is an
59:is the process of determining
1:
859:Optical character recognition
790:Content-based image retrieval
402:Michal Irani and P. Anandan:
330:Computer Vision and Robotics
169:and frequency domain methods
111:finding corresponding points
456:Swartz, Charles S. (2005).
957:
755:Automatic image annotation
590:Noise reduction techniques
363:. Retrieved on 2008-02-29.
172:Pixel recursive algorithms
941:Motion in computer vision
907:
720:Free viewpoint television
251:discrete cosine transform
785:Computer-aided diagnosis
299:Graphics processing unit
218:Affine motion estimation
213:Affine motion estimation
162:Block-matching algorithm
77:global motion estimation
847:Moving object detection
837:Medical image computing
600:Research infrastructure
570:Image sensor technology
294:Moving object detection
121:is used to represent a
884:Video content analysis
852:Small object detection
631:Computer stereo vision
304:Vision processing unit
255:video coding standards
237:
189:use features, such as
144:
44:
37:
889:Video motion analysis
700:Structure from motion
646:3D object recognition
232:
140:
106:stereo correspondence
38:
22:
812:Foreground detection
795:Reverse image search
775:Bioimage informatics
745:Activity recognition
464:Taylor & Francis
404:About Direct Methods
327:John X. Liu (2006).
235:motion interpolation
27:
879:Autonomous vehicles
817:Gesture recognition
680:2D to 3D conversion
333:. Nova Publishers.
247:motion compensation
894:Video surveillance
832:Landmark detection
740:3D pose estimation
725:Volumetric capture
685:Gaussian splatting
641:Object recognition
555:Commercial systems
359:2004-07-23 at the
238:
233:Video frames with
100:image registration
45:
33:
918:
917:
827:Image restoration
770:Augmented reality
735:
734:
715:4D reconstruction
667:3D reconstruction
560:Feature detection
442:978-1-4615-6241-2
340:978-1-59454-357-9
276:3D reconstruction
262:video compression
167:Phase correlation
134:standard defines
69:ill-posed problem
57:motion estimation
36:{\displaystyle z}
948:
936:Motion (physics)
931:Video processing
842:Object detection
807:Face recognition
690:Shape from focus
663:
550:Digital geometry
524:
517:
510:
501:
495:
484:
478:
477:
453:
447:
446:
426:
420:
413:
407:
400:
394:
387:
381:
380:
378:
370:
364:
351:
345:
344:
324:
191:corner detection
187:Indirect methods
182:Indirect methods
132:H.264/MPEG-4 AVC
53:image processing
42:
40:
39:
34:
956:
955:
951:
950:
949:
947:
946:
945:
921:
920:
919:
914:
903:
874:Robotic mapping
822:Image denoising
731:
652:
619:
585:Motion analysis
533:
531:Computer vision
528:
498:
485:
481:
474:
466:. p. 143.
455:
454:
450:
443:
428:
427:
423:
414:
410:
401:
397:
388:
384:
376:
372:
371:
367:
361:Wayback Machine
352:
348:
341:
326:
325:
321:
317:
290:
278:
243:
227:
215:
206:
184:
158:
149:
89:
49:computer vision
25:
24:
17:
12:
11:
5:
954:
952:
944:
943:
938:
933:
923:
922:
916:
915:
908:
905:
904:
902:
901:
899:Video tracking
896:
891:
886:
881:
876:
871:
869:Remote sensing
866:
861:
856:
855:
854:
849:
839:
834:
829:
824:
819:
814:
809:
804:
799:
798:
797:
787:
782:
780:Blob detection
777:
772:
767:
762:
757:
752:
747:
742:
736:
733:
732:
730:
729:
728:
727:
722:
712:
707:
705:View synthesis
702:
697:
692:
687:
682:
677:
671:
669:
660:
654:
653:
651:
650:
649:
648:
638:
636:Motion capture
633:
627:
625:
621:
620:
618:
617:
612:
607:
602:
597:
592:
587:
582:
577:
572:
567:
562:
557:
552:
547:
541:
539:
535:
534:
529:
527:
526:
519:
512:
504:
497:
496:
488:Daniel Cremers
479:
472:
448:
441:
421:
408:
395:
382:
365:
346:
339:
318:
316:
313:
312:
311:
306:
301:
296:
289:
286:
277:
274:
242:
239:
226:
223:
214:
211:
205:
202:
183:
180:
179:
178:
173:
170:
164:
157:
156:Direct methods
154:
148:
145:
88:
85:
61:motion vectors
32:
15:
13:
10:
9:
6:
4:
3:
2:
953:
942:
939:
937:
934:
932:
929:
928:
926:
913:
912:
911:Main category
906:
900:
897:
895:
892:
890:
887:
885:
882:
880:
877:
875:
872:
870:
867:
865:
864:Pose tracking
862:
860:
857:
853:
850:
848:
845:
844:
843:
840:
838:
835:
833:
830:
828:
825:
823:
820:
818:
815:
813:
810:
808:
805:
803:
800:
796:
793:
792:
791:
788:
786:
783:
781:
778:
776:
773:
771:
768:
766:
763:
761:
758:
756:
753:
751:
748:
746:
743:
741:
738:
737:
726:
723:
721:
718:
717:
716:
713:
711:
708:
706:
703:
701:
698:
696:
693:
691:
688:
686:
683:
681:
678:
676:
673:
672:
670:
668:
664:
661:
659:
655:
647:
644:
643:
642:
639:
637:
634:
632:
629:
628:
626:
622:
616:
613:
611:
608:
606:
603:
601:
598:
596:
593:
591:
588:
586:
583:
581:
578:
576:
573:
571:
568:
566:
563:
561:
558:
556:
553:
551:
548:
546:
543:
542:
540:
536:
532:
525:
520:
518:
513:
511:
506:
505:
502:
493:
489:
483:
480:
475:
473:9780240806174
469:
465:
461:
460:
452:
449:
444:
438:
434:
433:
425:
422:
418:
412:
409:
405:
399:
396:
392:
386:
383:
375:
369:
366:
362:
358:
355:
350:
347:
342:
336:
332:
331:
323:
320:
314:
310:
307:
305:
302:
300:
297:
295:
292:
291:
287:
285:
283:
275:
273:
271:
267:
263:
258:
256:
252:
248:
240:
236:
231:
224:
222:
219:
212:
210:
203:
201:
199:
194:
192:
188:
181:
177:
174:
171:
168:
165:
163:
160:
159:
155:
153:
146:
143:
139:
137:
136:motion vector
133:
128:
126:
125:
120:
119:motion vector
115:
112:
108:
107:
102:
101:
96:
95:
87:Related terms
86:
84:
82:
78:
74:
70:
66:
62:
58:
54:
50:
30:
21:
909:
802:Eye tracking
658:Applications
624:Technologies
610:Segmentation
482:
458:
451:
431:
424:
411:
398:
385:
368:
349:
329:
322:
279:
259:
253:(DCT) based
244:
241:Video coding
225:Applications
216:
207:
195:
186:
185:
176:Optical flow
150:
141:
135:
129:
122:
118:
116:
104:
98:
94:optical flow
92:
90:
76:
60:
56:
46:
710:Visual hull
605:Researchers
925:Categories
580:Morphology
538:Categories
315:References
147:Algorithms
124:macroblock
615:Software
575:Learning
565:Geometry
545:Datasets
357:Archived
288:See also
71:as the
470:
439:
337:
198:RANSAC
73:motion
65:frames
377:(PDF)
117:Each
81:pixel
468:ISBN
437:ISBN
335:ISBN
270:HEVC
266:MPEG
138:as:
130:The
103:and
51:and
490:. "
280:In
47:In
927::
462:.
272:.
200:.
55:,
523:e
516:t
509:v
476:.
445:.
379:.
343:.
31:z
Text is available under the Creative Commons Attribution-ShareAlike License. Additional terms may apply.
