Hyper-spectral imaging is a non-invasive diagnostic technique capable of providing information about structure and composition of sub-surfaces of lamin ar materials, both biological and non-organic, in various contexts.

What is available?

The SpectraCube imaging device

• A state of the art imaging spectroscopy facility and associated software
• On-site technical support for imaging and analysis of the samples
• Provision of consultancy and support for feasibility studies from world class experts

The Equipment:

• Digital Dark Room comprising of a room painted with black non-reflective paint, Kaiser calibrated camera stand and lamp support, calibrated fluorescent light sources, a digital camera (8 bit), a digitizer board and associated software, host computer, state of the art software (Aphelion)
• The SpectraCube system made up of an interferometer, a CCD-camera, a computer and spectral image analysis software. The spectral imaging system has two viewing modes: Interferometer mode and Direct View mode. In Interferometer mode, the SpectraCube system measures the spectrum of every pixel in the image. This spectral data reveals information that conventional imaging technologies cannot detect. In Direct View mode, instead of directing light through the interferometer, the system directs light straight to the system's cooled CCD-camera, which can record image details under extremely low intensities. The result is a finely detailed high-resolution gray scale image ( http://www.spectral-imaging.com/inner/fs_technology.html).

On-site Expertise:

Spectral analysis of an image taken using the SpectraCube imaging device

There is significant work being done on the 'non -invasive' interpretation and quantitative analysis of images by our Image Understanding group in the School of Computer Science. Recent work focusing on medical images, including the characterisation of mammographic and skin lesions, to date, has involved successful collaboration between physicists, computer scientists and the medical profession, as well as link-ups with industry.

As well as the academic output, the group's project work has generated practical products. An integrated workstation for neuroclinicians developed through the EU funded projects was used for several years in hospitals in Dusseldorf and Brussels, and continues to be used at Birmingham Queen Elizabeth Hospital. A new skin imaging method developed through a PhD project in 1998 was patented and licensed (through BRDL, the Birmingham University technology transfer unit) to a small British company Astron Clinica. A commercial imaging device based on this work, SIAscope, is now used in dermatology clinics in Europe aiding in the early diagnosis of skin cancers and has just received an FDA approval in USA (www.siascope.com). The body of clinical work using the SIAscope is growing and results indicate that the technology is superior to other existing skin imaging methods. The work has achieved recognition through the award of prizes at international conferences and the commercial skin imaging device received a number of industrial awards including DTI SMART award, the Millennium Product Status and a US I.D. Design Commendation. An extensive programme of scientific and clinical research is refining and extending the applications of the technology such as to help fight serious eye diseases caused by diabetes, colon cancer and others.

The group is now refining and extending the applications of this technology in biomedical, agricultural, food environmental and manufacturing sectors, by combining image analysis with spectroscopy. This technology is generic and can be applied to a wide variety of materials in biomedical, technical, agricultural, food and environmental contexts. The following are some examples of existing applications which have already benefited from the analysis of hyperspectral data: E.Claridge@cs.bham.ac.uk

• Skin cancer diagnosis , Wound care industry , Early detection of eye disorders from fundus images
• Cytogenetics and pathology
• Quality grading and inspection of fruit (e.g. bruising,) and testing the state of ripeness of fruit and vegetables. Testing for nutrient deficiency in plants through hyperspectral leaf imaging
• Faecal contamination in chicken; fungal and bacterial contamination in fruits. Other biological tissues and samples, both human and animal, to test for pathological changes, In cell biology and genetic analysis, quantitative and qualitative analysis of samples, to enhance spectroscopy by provision of spatial information
• Jewellery industry
• Forensic analysis
• Technical products, for rapid on-line quality control of thin multi-layer paints and coatings (e.g. to check evenness of layer thickness and evenness of distribution) and of translucent objects, for non-invasive quantification of degree of contamination by deposits such as e.g. soot