| EPSRC Reference: |
GR/J10464/01 |
| Title: |
LOW COMPLEXITY ROBUST OBJECT RECOGNITION ALGORITHMS |
| Principal Investigator: |
Yates, Professor R |
| Other Investigators: |
|
| Researcher Co-Investigators: |
|
| Project Partners: |
|
| Department: |
Electronic and Electrical Engineering |
| Organisation: |
University of Sheffield |
| Scheme: |
Standard Research (Pre-FEC) |
| Starts: |
01 June 1993 |
Ends: |
31 May 1995 |
Value (£): |
86,346
|
| EPSRC Research Topic Classifications: |
|
| EPSRC Industrial Sector Classifications: |
|
| Related Grants: |
|
| Panel History: |
|
|
Summary on Grant Application Form |
|
To develop algorithms for robust object recognition and localisation where robust means that algorithmic performance can be quantified for a particular application. An additional objective is to investigate the suitability of these algorithms for implementation in VLSI hardware. Progress:The two year duration of this grant is now nearing completion and in this time we have carried out all of the algorithmic work outlined in the objectives of the original grant proposal and in addition, considerable hardware research not funded by this grant. In particular we have continued to evaluate and extend the original object recognition work and have identified a whole family of geometric histograms with differing levels of invariance characteristics. In addition we have shown that this representation of shape is complete, in the sense that the original object can be recovered from a set of histograms describing a shape [1] [2]. We have also evaluated the potential scalability of the method for application to large model data bases and are currently in a position to show that the representation is sufficiently powerful to discriminate between in excess of one million object fragments. We have also designed a VLSI device and evaluated several commercially available chips for this purpose [3]. We have developed a derivation for the statistical origin of the Bhattacharrya measure for comparison of probability density distributions. This gives a rigorous underpinning to the use of geometric histograms for object recognition. We have also extended the original methods to improve robustness by the use of probabilistic generalised hough transform for object location. This delivers meaningful statistical measures which can be thresholded in a meaningful manner when wishing to determine an objects presence in a scene. The work has identified several new avenues for possible research for the use of these methods in the recognition of deformable shape and autonomous map building and path planning. [1] P A Riocreux, N A Thacker and R B Yates, An Analysis of Pairwise Geometric Histograrns for View-Based Object Recognition. Proc BMVC, York, Sept 1994. [2] N A Thacker, P A Riocreux and R B Yates, Assessing the Completeness Properties of Pairwise Geometric Histograms. Accepted for publication in Image and Vision Computing 1994. [3]N A Thacker, P Courtney, S N Walker, S J Evans and R B Yates, Specification and Design of a General Purpose Image Processing Chip. Proc ICPR 3, pp266-273, Jerusalem, Oct 1994.
|
|
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.shef.ac.uk |