The aim of the research is to develop intelligent imaging sensors that would find a wide area of applications in medicine, non-destructive testing, and security. These new sensors will revolutionise the way in which information is collected and processed because of their ability to respond to changing conditions during image recording and hence optimise their performance in real time. Thus, the region of high information content within the subject is investigated most closely whilst the remaining regions of low interest are recorded with the minimum amount of radiation. The result would be maximum information collected with the minimum penalty – the ideal imaging system. 

As a simple example, consider the left image which is a standard mammogram. The right image has been simulated from the left image assuming that the patient was scanned from left to right with a line scan mammographic imaging system using an I-ImaS sensor.

The intelligence used to ‘control’ the scanning parameters in the ‘I-ImaS image’ was derived from considering the mean value of the standard deviation (SD) of a strip of information. As each strip of information running from top to bottom of the image was ‘collected’ the X-ray tube current was increased or decreased according to the SD of the strip. If the SD increased then the electrical current (mA) was increased as this could imply a more variable and hence interesting region. The radiation dose saving using this simple measure was 18.9% and all the important features for carrying out the diagnosis are retained.

For more information contact: Joar Martin Østby