QuickTopic free message boards logo
Skip to Messages


Aligning Sequences and Actions by Maximizing Space-Time Correlations

^     All messages            3-18 of 18  1-2 >>
chanel fashion
02:21 AM ET (US)
has brought me much tauntingly,I looked a cross between a music,m not sure I buy the,I do not want a second longer,naturally chanel fashion
mbt shoes cheap
12:36 AM ET (US)
ugg mbt shoes cheap australiaJimmy choo just <a href="http://www.fitnessshoesforwomen.us" >
ugg boots cheap
03:12 AM ET (US)
Hollywood star and fashionable UGG snow boots on beauties esteem, let this sheep fur one shoe again came to the summit of fashion.Whether in department stores, the designer store, or in shopping website are massively filled with UGG figure.In order to avoid UGG fans buy fake snow boots, let vogue MM know more about UGG, here have deliberately UGG snow boots comprehensive explanation a!uggs on sale
spacerchaerPerson was signed in when posted
02:53 AM ET (US)
The weather in Chicago turned nasty all of a sudden a couple days ago, and I could not be HAPPIER! That is because I got to wear these Adirondack boots for the first time. You know you look good when you href="http://www.uggsinwinter.com" >ugg australia</a> discount ugg boots
Deleted by topic administrator 04-30-2010 08:02 AM
UGG Shoes
09:46 PM ET (US)
ugg bags ugg bags ugg bags ugg bags ugg bags uggs
Deleted by topic administrator 09-13-2009 02:07 AM
05:46 PM ET (US)
In response to Tom's response to me: That might work, though you'd have some interesting tradeoffs in window size.

It seems though that time is just fundamentally different in that it's ordering more than distance that matters. You see this a lot in speech recognition with Markov chains and such, where each phoneme can be of (almost) any length and it's the order of phonemenes that really matters.
Carolina GalleguillosPerson was signed in when posted
04:20 PM ET (US)
Pretty cool paper. I also wonder if outliers can be detected in another way or complementary to the concavity constraint for the quadratic approximation of the normalized correlation function.

Another question is about the computation time/space required for comparing the two sequences, in the case they want to use this for action/event recognition or any other of the applications that they mention.
02:14 PM ET (US)
re Anton: This technique might be too specific in the behaviors it would recognize, apart from it being slow. I'd be more interested in seeing the cuboid approach applied to the sequence alignment. By taking into account relative position between the cuboids, and using something like RANSAC, as mentioned earlier, to try and find a transform that aligns matching cuboid features. I'd imagine this could be more robust and efficient.
01:59 PM ET (US)
The results are impressive especially when fusing multiple sensors together for the same schene. I would be interested to see how wide the baseline between the sensors can be before this algorithm breaks down.
12:58 PM ET (US)
It'd be really interesting to see how well this performs in a behavior recognition framework. Specially when compared with cuboids, since cuboids performed better than Zelnik and Irani's "Event-Based Analysis of Video" which similarly recognized behavior regardless of differences in spatial textures such as clothes.
12:32 PM ET (US)
The results look really cool, but I imagine this takes years to run even for short sequences. I brought this up when we were talking about the cuboids paper, but I wonder if a more randomized search like RANSAC could work sufficiently well and be a lot faster.
04:01 AM ET (US)
My question is in the same vein as Nikhil. Some of techniques used to enhance convergence such as those described in "Confidence weighted regression" seem quite unique to the problem domain. I was wondering if these anyone had seen these particular techniques applied to other optimizatio problems.
03:43 AM ET (US)
I may not be understanding the paper correctly, but could this paper be described as doing something analagous to RANSAC, except in space-time, rather than just space? Did that make sense?
03:03 AM ET (US)
In response to Matt: My guess is that you could segment the sequence temporally and do alignments on sequences that could be linearly aligned if nonlinear temporal scaling was a nasty issue. As an example, say your ballerina fell twice but otherwise performed at full speed. If you segmented it properly the ground time would just be a VERY slow progression through the perfect performance with a low match metric. Just a thought mind you.
^     All messages            3-18 of 18  1-2 >>

Print | RSS Views: 2578 (Unique: 1216 ) / Subscribers: 1 | What's this?