Using Model-Based Approach to Confident Free-Form 3D Object Matching

An ATR system is only useful when its results are accurate to a high degree and when these results are produced in the short time frame when the target is observed. Employing 3D data in matching algorithms offers significantly more complete information asto the characteristics of the object observed, however raises the challenge of using the additional information properly to raise the matching accuracy and keeping the matching process performance acceptable for mission specific applications. We willexplore feasibility and propose innovative approaches in the following areas: (1) selecting features of 3D data and object representations that are suitable for matching; (2) algorithms for matching observed 3D data to predicted 3D data; (3) algorithms formatching observed 3D data to 3D models directly without prediction;Since 1992 IT Concepts has been actively working in the image processing and image recognition field. IT Concepts was the first company that developed high performance fingerprint recognition algorithms that could efficiently run on personal computers.This recognition technology has been widely used within the federal and state government and major corporations. IT Concepts will apply this potential to design innovative approach for high performance 3D objects matching. Accurate 3D target recognition technology would become a very successful commercial product in many areas. Increasingly, both military and civil applications rely on video cameras, remote video surveillance and automatic recognition. Presently, 2Drecognition technology is used for automatic target recognition and scene analysis. Unfortunately, for many very important applications 2D recognition technology cannot provide sufficient accuracy. If accurate 3D technology exists today the market wouldimmediately request it. 3D identification technology would bring real breakthrough to the enterprise security products based on facial recognition. Such products are in great demand and cover aspects of logical and physical access security,authentication, resource access authorization and others. Accurate 3D face identification technology would also greatly simplify and automate numerous consumer applications, for example customer loyalty programs. On the other hand generic 3D objectrecognition technology could allow for better productivity and automation of various manufacturing processes including parts production, sorting and quality control.