The study and development of a decision support system for doctors and students is presented, aiming to ease the diagnosis of the cervix cancer through an automated smart system: using a web application, a medical expert can upload colposcopie images feeding an expert system that carries out a deep analysis by a processing system on images uploaded by doctors; results coming out of the processing system are presented by a user friendly system suggesting the decision to the the medical expert who is able to confirm or change it, and annotate information. If changed, the diagnosis is sent to the expert system, developed on a reinforcement learning scheme, to tune decision parameters and enhance detection rates. The paper present the work in progress preliminary results of the system being developed
CADdy – Colposcopy Learning Machine for Computer Aided Diagnosis / Traversi, M.; Falagario, M.; Guaragnella, C.. - (2013), pp. 330-333. (Intervento presentato al convegno IEEE Third International Conference on Consumer Electronics, ICCEBerlin 2013 tenutosi a Berlin, Germany nel September 8-11, 2013) [10.1109/ICCE-Berlin.2013.6697965].
CADdy – Colposcopy Learning Machine for Computer Aided Diagnosis
Guaragnella, C.
2013-01-01
Abstract
The study and development of a decision support system for doctors and students is presented, aiming to ease the diagnosis of the cervix cancer through an automated smart system: using a web application, a medical expert can upload colposcopie images feeding an expert system that carries out a deep analysis by a processing system on images uploaded by doctors; results coming out of the processing system are presented by a user friendly system suggesting the decision to the the medical expert who is able to confirm or change it, and annotate information. If changed, the diagnosis is sent to the expert system, developed on a reinforcement learning scheme, to tune decision parameters and enhance detection rates. The paper present the work in progress preliminary results of the system being developedI documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
