Faster, more accurate 3D scanning and Face Recognition
Summary
The Geometric Modelling and Pattern Recognition (GMPR) Group at Sheffield Hallam University has developed and patented internationally-known line projection technologies for fast 3D scan, reconstruction and recognition. We have built up an extensive library of 3D faces from which we are continually developing our face recognition methods for a range of applications. This has led to the licensing of patented technology to European and the US companies and its use in EU-funded technology projects. Within the region, the technology has also been used to bring to life the Museums Sheffield Metalwork Collection on the web, and is helping to turn a landmark sculpture project and gateway to South Yorkshire and the Sheffield City Region into a reality.
Our research
The GMPR research team has made three significant advances
- The development of unique structured light techniques for 3D scanning and reconstruction
- Pattern recognition using the 3D data through the development of face recognition algorithms to allow high levels of success, but requiring only a small number of features to represent the subject
- The development of mathematical models for 3D data compression
Professor Marcos A Rodrigues and Dr Alan Robinson's research breakthrough came in 2004 with their patented uncoded multiple line projection technique. The technique could provide accurate real time 3D reconstruction from a 2D image more quickly and easily than previous state-of-the-art technology based on single stripe systems or coded light patterns.
Once this method had been developed, the research began to focus on improving resolution and since 2011, the GMPR scanner has been able to offer a resolution of 0.25mm in the horizontal (4 vertices per millimetre) and 0.5mm in the vertical direction (2 vertices per millimetre) for an object scanned within 1 metre cubed. An accuracy of 40 micrometres can be achieved for an object scanned at 10mm from the surface. The scanning speed is typically < 5 milliseconds, depending on the capture speed of the image sensor.
There is a strong continuing interest in active structured light technology as evidenced by the Kinect box from Microsoft using coded structured light. GMPR technology has greater accuracy and measurement density than Kinect, making it appropriate to applications in the medical, industrial inspection, quality control, and security domains.
The research was partially guided by the National Science Foundation Face Recognition Grand Challenge set in 2005 to improve 2D face recognition success rates by using 3D face measurements. In 2010-11 GMPR demonstrated real-time face recognition: in just over one second, the group's face tracking software can automatically take an image, process it, reconstruct the image in 3D, normalise it to compare to a database and recognise that face in a database.
Sheffield Hallam University have offered invaluable assistance to our project, providing crucial 3D data that will be used to model a full scale version of the figure.
Yorkshire Man of Steel team
The impact
Companies have been attracted by the speed and accuracy of GMPR scanning techniques, pattern recognition algorithms and methods for face recognition in 3D and in 2D. In Europe and in the USA, technology patents have been issued to commercialise the research in a wide range of contexts: in 3D and 2D face recognition for security applications and for various purposes in both medical and entertainment markets.
Two EU-funded projects have also seen the benefits of these technologies extend to small and medium enterprises (SMEs) across Europe. In the MARWIN project, they form a critical part of a cognitive 3D based vision system designed to increase the productivity and quality of robotic welding assemblies. GMPR technology is also essential to the ADMOS project, which provides intelligent analytics on outdoor media by tracking the approximate age and gender of passers-by and whether or not they have noticed a particular advert.
These technologies have also made a difference to local cultural projects. In 2011, with funding from the JISC e-Content Programme, GMPR 3D scanned the internationally important Museums Sheffield Metalwork Collection. This project produced a digital record of the designated-status collection now available on the web. In 2012, GMPR technology was used in the development of a new Man of Steel stainless steel landmark and visitor centre for the Yorkshire and Sheffield region which celebrates the community's connections with the steel industry.
