New device uses a standard digital camera with a calibrated light source and measures the way in which different wavelengths of light are absorbed, much like a spectrometer
Astron Clinica has used Matlab from MathWorks to develop an imaging technology that is designed for the early diagnosis of skin cancer.
The non-contact Siascope uses an optical modelling technique to model the transport of light through tissue and map out the underlying micro-architecture of the skin.
This helps to identify lesions such as basal cell carcinoma.
Matlab significantly reduced the time taken to develop the technology and simplified its calibration and analysis during early trials. The Siascope helps with the early diagnosis of skin cancer by enabling better visualisation and detection of regions of tissue with abnormal blood flow.
The device uses a standard digital camera with a calibrated light source and measures the way in which different wavelengths of light are absorbed by the skin, much like a spectrometer.
Matlab was used to construct a function that links the red, green and blue elements of every pixel in the image to a model of the underlying histology of the skin.
This data is then displayed as an image, known as a Siagraph, that maps out the concentrations of melanin and blood in the epidermal layer.
"Constructing the modelling function was a difficult task, but using Matlab really simplified the amount of work we had to do.
It reduced the time taken to develop the technology by almost five-fold compared to what it would have taken in C," said Steve Preece, a lead scientist at Astron Clinica.
"As a result of the development work we've done we've realised what a useful tool Matlab is and we're now trying to extend it to other similar products".
As the Siascope was developed, Matlab enabled the company to visually assess the accuracy of the calibration procedure for the imaging system. This was done by comparing the predictions of the model with image data for the skin.
This was important in the early stages of development and was something that couldn't have been done in C.
Matlab was also used to analyse the skin tissue images taken during the initial trials of the technology and will be used to process the images of different skin lesions, such as basal cell carcinomas, when the technology is used commercially. This is done by comparing the model predictions of skin colouration with the image data collected.
If any pathology is present in the tissue, the model detects this as different colours in the image.
Having developed the technology, Astron Clinica believes it can extend its use to new medical imaging applications, such as cosmetics testing or skin disorders, because the underlying mathematics for many different functions is very similar.
