Studio di fattibilità sul rilevamento di enfisema in immagini ottenute mediante TAC a spirale con bassa dose

Amendola T., Chimenti M., Fantini E., Tampucci M., Miniati M., Monti S.

Image Processing  D.0 Software. J.3  LIFE AND MEDICAL SCIENCE 

In a previous study we developed a software able to detect anomalies in CT high definition images of the thorax, possibly related to pulmonary emphysema; in this study we evaluate the possibility of applying this software to the CT images obtained in a large scale screening. In fact, theses images have been produced by a low-dose computer tomography and are affected by a noise greater than that of previously processed images. In this study we used a photometric images analysis, which gives the amount of pixels with luminance value less than a predefined threshold, and the blob analysis, which using the same threshold detects the binary objects composed of adjacent pixels and gives up their size and location. The paper reports the results obtained according two decision thresholds: we show that the photometric analysis of low-dose CT images gives no information about the presence of pulmonary emphysema, while the blob analysis gives useful descriptors