| EPSRC Reference: |
EP/G037590/1 |
| Title: |
Bayesian Inference for Multi-object Tracking with Application to Single Molecule Fluorescence Microscopy |
| Principal Investigator: |
Singh, Dr SS |
| Other Investigators: |
|
| Researcher Co-Investigators: |
|
| Project Partners: |
|
| Department: |
Engineering |
| Organisation: |
University of Cambridge |
| Scheme: |
First Grant Scheme |
| Starts: |
01 September 2009 |
Ends: |
28 February 2013 |
Value (£): |
290,844
|
| EPSRC Research Topic Classifications: |
| Analytical Science |
Image & Vision Computing |
| Statistics & Appl. Probability |
|
|
| EPSRC Industrial Sector Classifications: |
| Aerospace, Defence and Marine |
|
|
| Related Grants: |
|
| Panel History: |
| Panel Date | Panel Name | Outcome |
|
29 Jan 2009
|
ICT Prioritisation Panel (January 2009)
|
Announced
|
|
|
Summary on Grant Application Form |
|
This proposal concerns the Bayesian analysis of data arising from multi-object tracking problems.As a specific example of object tracking, consider a surveillance system comprised of a network of cameras. The aim is to detect new objects that enter their field of view, classify them (pedestrians or vehicles, in the latter case, the type of vehicle), follow their trajectories and ultimately infer their intent.With advances in computing, networking and sensor technologies, surveillance systems are becoming more pervasive and sophisticated.A recent example is the Cambridge Transport Information Monitoring Environment which aims to monitor and improve the efficiency of the public transport infrastructure in Cambridge, e.g. by predicting the onset of congestion on roads and reschedule services as necessary.If current trends persist, it is likely that the amount of such data that is collected will only increase in years to come and developing efficient statistical methods for analysing it will continue to be an important challenge.This project aims to develop a flexible and robust methodology for Bayesian inferencefor multi-object tracking to meet the demands of existing and theever-increasing complexity of new applications.It will consider the computational and statisticalaspects of the problem and the developed methodology will be applied primarily to a new application area, namely Single Molecule Fluo-rescence Microscopy (SMFM).SMFM is an emerging area of study in Biology which is contributing tothe quantitative understanding of the fundamental processes in living cells.Experimentation techniques are advanced but the statisticalanalysis of the data, which is needed to support the scientific investigations, has not been addressed. The multi-object tracking framework is ideally suited to this application and our developed computational methodology will be specialized for it.In the methodological work, we will consider issues of generic and practical importance. Currently,there are three main challenges which will be addressed in this project. The first is computing the complex posterior probability distribution from which all inference is to be based.Getting the right answer to the practitioner's problem rests on a comprehensive solution for this problemwhich is presently unavailable. This proposal aims systematically address this using Monte Carlo as a primarymeans for computation.The second is calibrating the statistical models in multi-object tracking which are complex and have many tunable parameters.This problem is important as analysis based on models that are not properly calibrated will yieldincorrect results but unfortunately has received very little attention.The third is model assessment, which involves determining whether or not a proposed model adequately fits the data, and determining whether or not different models would be more suitable.In multi-object tracking, the data structures and the models are complex and routine methods for Bayesian model assessment cannot be immediately applied. A dedicated set of diagnostic tools need to be developed.
|
|
Key Findings |
|
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk
|
|
Potential use in non-academic contexts |
|
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk
|
|
Impacts |
|
| Description |
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk |
| Summary |
|
| Date Materialised |
|
|
|
|
Sectors submitted by the Researcher |
|
This information can now be found on Gateway to Research (GtR) http://gtr.rcuk.ac.uk
|
| Project URL: |
|
| Further Information: |
|
| Organisation Website: |
http://www.cam.ac.uk |