The BreastScreening project aims to develop an improved breast cancer Computer-Aided Diagnosis (CAD) system. The system incorporates text, MammoGraphy (MG), UltraSound (US) and Magnetic Resonance Imaging (MRI), using Deep Belief Networks (DBN).
This system optimally integrates information from any combination of the above data sources, so as to improve the efficiency, sensitivity and specificity of breast cancer diagnosis. The project is a collaboration between Instituto Superior Técnico, University of Adelaide and University of Queensland. Also, the BreastScreening project receives contributions from both MIDA and MIMBCD-UI projects.
With the BreastScreening project we have two research questions: 1) can we use Deep ConvNet models pre-trained from Computer Vision (CV) datasets on medical image analysis applications; and 2) can we use Deep ConvNet to analyse unregistered medical images.
We present a novel methodology for the automated detection of breast lesions from Dynamic Contrast-Enhanced Magnetic Resonance (DCE-MRI) volumes. Our method, based on Deep Reinforcement Learning (DRL), significantly reduces the inference time for lesion detection compared to an exhaustive search, while retaining state-of-art accuracy.
In short, the BreastScreening project is an automated analysis of Multi-Modal Medical Data using Deep Belief Networks (DBN). This project is an ARC Discovery Project (DP140102794) aiming at: (i) automatically detect and segment suspicious regions from different breast imaging data, e.g., MG, US and MRI; (ii) estimate BI-RADS scores (from 0 to 6) from the segmentations and the patient’s clinical records; (iii) retrieve similar cases from our database, given the above estimation results; and (iv) automatically extract relevant features from any combination of input data types.
Our Researchers and Fellows:
Our Support Researchers, Fellows and Interns:
Our Doctors, Physicians and Clinicians:
Our Sponsors, Donors and Honors:
