[Mp4-tech] best motion estimation

[Andrew Krupiczka]

Hi kk,
>From our somewhat limited experiences I'd also suggest to look at frequency based
"translation motion measurements" approach, namely phase correlation method, 
originally developed by Thomas from BBC R&D in late 80s.
As the major advantage here I regard the fact that complexity of doesn't depend on the 
motion searching range making it very desirable for higher video resolutions.
Also because of normalization stage involved this algorithm is completely insensitive 
to illumination changes i.e. still correctly measures occuring motions
what might become a serious issue for algorithms operating in spatial domain.
Regards,
Andrew
> -----Original Message-----
> From: Tourapis Alexandros [mailto:alexandros.tourapis thomson.net]
> Sent: Tuesday, June 29, 2023 6:04 PM
> To: karunakar A.K.; mp4-tech lists.mpegif.org
> Subject: RE: [Mp4-tech] best motion estimation
> 
> 
> It depends on how/what you quantify/qualify as "best". There are many
> items you need to consider, implementation complexity (hardware vs
> software), memory requirements, quality peformance, speed up increase
> etc. Other constrains within your system could also affect 
> your decision
> of the ME scheme, while depending on the codec you use, i.e. 
> MPEG4-AVC,
> other features such as Mode decision are critical. There is a book by
> Peter Kuhn named "Algorithms, Complexity Analysis and Vlsi 
> Architectures
> for Mpeg-4 Motion Estimation" which provides some overview, 
> but it might
> be considered a bit outdated while you can definitely find much better
> references online. For certain applications (especially in hardware)
> Hierarchical algorithms can be for certain applications (esp in
> hardware) the better option, while for others you can go for 
> predictive
> algorithms (although you can claim that maybe I am biased on this
> prediction). DS in general is not considered a very good scheme. 
> 
> You can see the following :
>  
> M.Gallant, G.Cote, and F.Kossentini, "A Computation Constrained
> Block-based Motion Estimation Algorithm for Low Bit Rate 
> Video Coding",
> IEEE Transactions on Image Processing, 8(12), Dec'99 
> A.M.Tourapis, O.C.Au, and M.L.Liou, "ighly efficient predictive zonal
> algorithms for fast block-matching motion estimation,"IEEE 
> Transactions
> on Circuits and Systems for Video Technology, Vol.12, Iss.10, Pages:
> 934- 47, Oct'02
> A.M.Tourapis, "Enhanced Predictive Zonal Search for Single 
> and Multiple
> Frame Motion Estimation," in proceedings of Visual Communications and
> Image Processing 2002 (VCIP-2002), pp. 1069-79, San Jose, CA, Jan'02
> H.Y. Cheong, A.M. Tourapis, "Fast Motion Estimation within the H.264
> codec," in proceedings of 2003 IEEE International Conference on
> MultiMedia & Expo (ICME-2003), Baltimore, MD, July 6-9, 2003
> 
> There are also several algorithms in the literature with tiny 
> variations
> (e.g. using different patterns for example) but more or less 
> the concept
> is the same. Select a set of predictors, examine their reliability,
> select the best one(s), and if necessary refine using certain patterns
> within a small area around them. 
> 
> Best regards
> 
> Alexis
> 
> 
> 
> 
> 
> ________________________________
> 
> From: mp4-tech-bounces lists.mpegif.org
> [mailto:mp4-tech-bounces lists.mpegif.org] On Behalf Of karunakar A.K.
> Sent: Tuesday, June 29, 2023 10:33 AM
> To: mp4-tech lists.mpegif.org
> Subject: [Mp4-tech] best motion estimation
> 
> 
> Please tell me which is the best block based motion estimation
> algorithm? I consider the fast dimond search, is it true or not?
> with regards
> kk
> 
> Yahoo! India Matrimony
> <http://in.rd.yahoo.com/specials/mailtg/*http://yahoo.shaadi.c
om/india-m
atrimony/> : Find your partner online
<http://in.rd.yahoo.com/specials/mailtg2/*http://yahoo.shaadi.com/india-
matrimony/community.php> .
_______________________________________________
NOTE: Please use clear subject lines for your posts. Include [audio, [video], [systems], [general] or another apppropriate identifier to indicate the type of question you have.
Note: Conduct on the mailing list is subject to the Antitrust guidelines found at http://www.mpegif.org/public/documents/vault/mp-out-30042-Antitrust.php