IV. CONCLUSIONS AND FUTURE PERSPECTIVES
The present work is part of current investigation on
algorithms for segmentation of X-ray IlCT biomedical images
and we intended to perform a viability study on the possibility
of using ANNs for more complex images.
The preliminary results obtained with ANNs that use
statistical information such as mean and standard deviation of
intensity allow us to state that ANNs should be more
investigated, representing a state-of-art technique which is a
promising method for X-ray IlCT imaging.
Future investigation on other statistical inputs such as
skewness and kurtosis will be performed. Other learning
algorithms and architectures shall also be investigated. The
application of ANNs to the segmentation of the whole scan of
this sample as well as to other samples will be assessed.