|
Written by Tom SF Haines
|
|
Thursday, 03 November 2011 |
The 13th International Conference on Computer Vision (2011) is just days away, and my paper promises code.
So, my paper is:
Delta-Dual Hierarchical Dirichlet Processes: A pragmatic abnormal behaviour detector
by me, Tom. S. F. Haines, and Tao. Xiang. The above links through to the pdf, as you would expect, and you can also get the poster from here. The code is available from my Google code project, which is accessible from the menu.
The approach itself is an extremely complex topic model (Non-parametric and Bayesian.) designed with video data in mind. Its stated purpose, which we demonstrate, is to be able to do semi-supervised learning of behaviour that is normal, except that it is happening in an unusual context. An example of this is we might want to differentiate between people crossing the road and people crossing the road whilst traffic continues to drive across the crossing. The results are somewhat weak however, due to the difficulty of finding real world examples of such behaviours with enough examples for training/testing. I am also not convinced that it fully converges, or that taking a single sample from such a complex Gibbs sampler is sensible (Which is all we can do given the run time.). The complexity is its main failing however - there are so many random variables being sampled, often using extremely complex techniques, that no sane person would ever reimplement this algorithm. Given how much work was involved I'm still somewhat surprised I got it implemented and working in the first place!
|
|
Last Updated ( Thursday, 03 November 2011 )
|
|
|
Written by Tom SF Haines
|
|
Wednesday, 24 August 2011 |
Next week is the British Machine Vision Conference 2011, at which I have a paper - Active Learning using Dirichlet Processes for Rare Class Discovery and Classification by T. S. F. Haines and T. Xiang. The paper can be obtained from
http://thaines.com/content/research/2011_bmvc/dp_al.pdf
whilst the source code can be obtained from my code store under the directory dp_al - click on the source code menu item to head there.
Edit: You can download my slides, with an additional bonus slide, from
http://thaines.com/content/research/2011_bmvc/dp_al_pres.pdf
and a video of me giving the talk will be available at some point, once the bmvc people set it up - at that time I'll stick a link up here.
Edit 2: You can now watch the video over at the bmva website:
http://www.bmva.org/bmvc/2011/proceedings/paper9/index.html
Whilst potentially conceptually tricky it is a very easy to implement active learning method. It assumes that the categories have been drawn from a Dirichlet process, which is to say that there is a Dirichlet process mixture model over the samples in the pool. As assumptions go it is a pretty weak one, and yet the moment you make it you gain the probability that an entity in the pool belongs to a previously unseen class. An old if mostly unexplored concept (The entropy method of active learning tends to do a little better.) is to select from the pool the entity which the classifier is most likely to classify incorrectly. On its own this idea only targets entities to refine the boundaries of existing categories, but if we include the Dirichlet process assumption, such that an entity could be misclassified either because it is the wrong side of a boundary or because it is an example of something new, its character changes. We now have an algorithm (Refereed to as P(wrong) in the paper.) that selects entities to discover new classes and refine the boundaries of the existing classes, that is particularly suited to situations where there are rare classes. It balances these two tasks without any real parameter tuning. This makes it a perfect candidate for many real world problems, such as automated surveillance, where entirely new behaviours regularly appear but boundary refinement of the existing classes remains important.
|
|
Last Updated ( Friday, 23 December 2011 )
|
|
|
Written by Tom SF Haines
|
|
Sunday, 08 May 2011 |
Sometime last year I published a paper in which I indicated that the code would be available on my website. Being that the paper was in some minor workshop I might of delayed.. slightly. Well, better late than never, and I guess this is a good time to put the paper, and other related stuff, online as well.
The paper itself is Video Topic Modelling with Behavioural Segmentation by T. S. F. Haines and T. Xiang, and appears in the ACM Workshop on Multimodal Pervasive Video Analysis, 2010. You may download it as a pdf by clicking on the papers title.
The code itself can be found over at my Google code project, which is linked from the source code link on this website. You can also download the presentation I gave here, which includes various video files demonstrating it in action on the mile end data set. Additionally, there is a 2 minute long demonstration video I created, which can be obtained here.
|
|
Last Updated ( Sunday, 08 May 2011 )
|
|
|
Written by Tom SF Haines
|
|
Sunday, 16 May 2010 |
Well, my PhD is finished, other than some more paperwork and the ceremony, which means two things:
1) I have put my thesis online - you may download it here. Alternatively you can download it directly from the University of York, from here.
2) The complete source code for it is available. Actually, its been online for some time now, but I never announced it - just added a link to the links section. But you can obtain it all from here.
|
|
Last Updated ( Monday, 24 May 2010 )
|
|
|
Written by Tom SF Haines
|
|
Tuesday, 07 October 2008 |
Well, next week is the European Conference on Computer Vision, and as I am presenting there it is time to put up the relevant stuff.
|
|
Last Updated ( Monday, 01 December 2008 )
|
|
Read more...
|
|
|
|
