Computer Assisted Decisions in Medicine based on Machine Learning Techniques on Biomedical Data and Signals

 In Medicine, computational methods could be used for learning mechanisms that will help medical doctors to induce knowledge from examples, from previous experience or from merged data, and, at the same time, to update already available intelligent systems.
 
In particular, innovative supervised machine learning methods are useful to support clinical decisions, in such cases where there is a lack of formal models, the knowledge about the application domain is poorly defined, and then the standard protocol solutions are not still available or rule based validated.
 
We show some our studies where both 2D and 3D medical images (CT-RMN-PET) and EEG signals visualization, filtering, processing and understanding, were combined with soft-computing and machine learning techniques for decision-making. The obtained results have provided an experimental added advantage for clinical applications and allowed us to propose innovative frameworks and carry out new experiments to extract new features useful for early diagnosis, better prognosis and optimized/personalized therapy.
 
Finally, following the experience in designing and testing CAD based on Medical Images and EEG signals, we present new experimental protocols focusing on some new trends of Human Machine/Computer Interaction scenarios and their results as solutions for studying elicited behaviors and movement disorders.


 

---