IDEA #1LQ3XJ Medical roentgen and tomographic images processing software (LESSRAD)

LessRad Project The accuracy of diagnostics is determined mainly by two factors: • doctor’s experience; • quality, reliability and informativeness of the source data. Roentgen and, especially, tomographic images provide not indirect but rather direct data on the anatomical structures. This is their radical advantage over other methods of patients’ examination. We have developed the software to solve the following tasks: • to ensure the quality of the source data; • to control the path of formation and registration of the image; • significant improvement of informativeness of the images; • radical reduction of radiation exposure level. The quality of the source data When working with roentgen film, the sources of methodical inaccuracy are: • spread of film sensitivity; • unevenness of radiation across the area; • instability of radiation level during the exploitation of the equipment; • spread of duration of radiation exposure and duration of film development; • transformation of properties of the film developer in the process of development of a series of images; • noises of the gain-conversion channel of analog signal; • small dynamic range of analog-to-digital converter (ADC). When working with tomograms, the sources of methodical inaccuracy are: • unevenness of radiation exposure level of generator array; • unevenness of sensitivity of sensor array; • insufficient number of generator-sensor pairs; • instability of radiation level during the exploitation of the equipment; • presence of phantom images (incorrect algorithm of tomogram calculation); • noises of the gain-conversion channel of analog signal; • small dynamic range of ADC. Our software provides regimes of path quality control, as well as the possibility of image normalization that is invariant to the certain conditions of the formation of images. Increase of informativeness of images The informativeness of an image is determined by the spatial resolving power and dynamic range of signal registration. Spatial resolving power is estimated using minimum value of details that can be seen by the eye. When studying an X-ray image or tomogram on an X-ray film using negatoscope, this value is somewhere in the range of 0.1…0.5 mm. Modern flatbed scanners allow to work with the resolution of 2400 dpi, meaning that with computer processing it is possible to differentiate 0.01 mm. Dynamic range is equal to the maximum number of density gradations that transmit the image. The human eye masterfully distinguishes about 20 shades of grey. Computer monitor can transmit 30…40 gradations. At the same time, X-ray film can store up to 40 000 levels of grey. Our software can fully consider all of the information of both X-ray and tomographic images, provided the usage of 16-bit ADC. Trial operation of our software has showed that in many cases diagnosticians, who work in accordance with traditional method, find it difficult to interpret drastically increased amount of information. In order to overcome this psychological barrier, we have developed algorithms that realize “half-tone microscope”. In this case, the optional local application of “color microscope” is possible in order clarify the density relief. This technique greatly facilitates the mastering of a new method of image analysis. The reduction of patient’s radiation load (radiation exposure) According to clinical studies, the radiation load on the body is proportional to the square of the voltage on the X-ray source. Our simulation showed that during the work with 16-bit range of images registration, the informativeness does not practically change when the voltage is reduced down to 10 times. This means that biological safety can be increased up to 100 times. Additional features of the software • automatic preliminary processing of series of tomograms; when working with dozens of tomograms, this feature greatly facilitates the work of diagnostician; • automatic ignoring of text information that is put on the image; • calculation of numerical values of anomalies; • semiautomatic mode of formation of diagnostic results with the inclusion of image elements; • service tools for the work with graphics, etc.