Implementing automated image processing increased productivity by an order of magnitude, removing 90-95% of the manual drudgery for degree level scientists so allowing them to concentrate on more value-adding tasks.  The scientists were spending hours each day counting cells viewed through a microscope.  With a camera fitted to the microscope, a viewer was developed based on a desktop computer to capture the experts’ assessment of cell images.  This information was then used to train a learning algorithm to perform the classification and counting automatically.  The algorithm included self-checking quality measures to detect borderline results and present those to the humans for manual assessment.  The result was increased productivity and improved quality and repeatability of the data being collected.

 

Designing and specifying the camera, lens and lights to capture the image data and developing and implementing algorithms to identify and count viral plaques.  Image processing lay at the heart of a

project that developed an automated assay process for live-virus vaccine quality-control.  A laborious

manual counting process was replaced by an automatic counting module.  

 

Assessed the quality and performance of the image processing software used by this company and identified alternative image processing strategies that could improve the system.  

 

Testing and selection of optical components and software tools to be used in a point of sale scanner.  Development of software to be used for process monitoring and inspection during the assembly of the optical components within the units.  This allowed the scanners to be built to meet time, budget and image quality specifications.

 

Quantified the performance of the optics of a prototype fluorimetry instrument and identified where noise was being introduced or signal strength was being lost so that the design could be improved before commercialisation and volume production.

 

Developed high-speed algorithms for inspection of pharmaceuticals.  

 

Designed, developed and implemented a system based on x-ray imaging to inspect the internal assembly of medical devices for a leading manufacturer of medical devices.  This reduced the time to market for release of a new device.

 

Developed novel optical arrangements for handling and viewing contact lenses and the associated image processing software for their automatic inspection during production.  This resulted in a patented inspection system for the client.  It is one example from a number of projects working on quality control during contact lens production for various clients.  Projects have also included systems to allow in-mould and in-pack inspection.

 

Designed novel configurations for handling and visualisation to allow machine-assisted inspection of stents.  Similar projects have been carried out for a number of different clients working with stents and other micro-scale medical devices.

 

Assessment of hardware from components to integrated systems and design of software algorithms for a number of clients on applications in the biometric monitoring and security fields.