Preliminary Results on Contrast Enhanced Ultrasound Video Stream Diagnosis Using Deep Neural Architectures

In recent years, the interest in the deep learning (DL) paradigm has grown dramatically. Due to its ability to automatically extract characteristics and recognize patterns from data, this technology has been particularly successful in the field of medical imaging. The current paper proposes a fully automatic computer-aided diagnosis DL-based system for the investigation of contrast enhanced ultrasound (CEUS) video streams in order to identify specific focal liver lesions (FLL). Our solution, based on a combination of a convolutional neural network (CNN) with a recurrent neural network (RNN) showed promising results (87% accuracy) when it was evaluated against the SYSU dataset.