Abstract
Machine learning has been well applied and recognized as an effective tool to handle a wide range of real situations, including medical applications. In this scenario, it can help to alleviate problems typically suffered by researchers in this field, such as saving time for practitioners and providing unbiased results. This tutorial is concerned with the use of machine learning techniques to solve different medical problems. We provide a survey of recent methods developed or applied to this context, together with a review of novel contributions to the ESANN 2016 special session on Machine learning for medical applications.

Abstract

Abstract
Purpose: This paper presents the Rehab@home system, a tool specifically developed for helping neurological patients performing rehabilitation exercises at home, without the presence of a physiotherapist. It is centred on the rehabilitation of balance and on the sit-to-stand (STS) movement. Method: Rehab@home is composed of two Wii balance boards, a webcam and a computer, and it has two main software applications: one for patients to perform rehabilitation exercises and another one for therapists to visualize the data of the exercises. During the exercises, data from the boards and the webcam are processed in order to automatically assess the correctness of movements. Results: Rehab@home provides exercises for the rehabilitation of balance (in sitting and in standing positions), and for the execution of the STS movement. It gives automatic feedback to the patient and data are saved for future analysis. The therapist is able to adapt the difficulty of the exercises to match with each patient's needs. A preliminary study with seven patients was conducted for evaluating their feedback. They appreciated using the system and felt the exercises more engaging than conventional therapy. Conclusions: Feedback from patients gives the hope that Rehab@home can become a great tool for complementing their rehabilitation process.

Abstract
Over the past years, volumetric cardiac imaging has matured to a modality that can be used in daily routine clinical practice. However, despite the evolution of volumetric ultrasound systems with remarkable improvement in image quality, spatio-temporal resolution of the 3D data set remains limited and inferior to what can be obtained in 2D. Further development of volumetric ultrasound is thus required. However, the development of new beam forming techniques for 3D ultrasound (US) imaging requires an open, flexible and fully programmable US platform. To date, such systems are scarce and required (custom-made) dedicated electronics. Therefore, the aim of this report is to present a novel High channel Density Programmable ULtrasound System based on consumer Electronics (HD-PULSE). © 2015 IEEE.

Abstract
Fast volumetric cardiac imaging requires to reduce the number of transmit events within a single volume. One way of achieving this is by limiting the field-of-view (FOV) of the recording to the anatomically relevant domain only (e.g. the myocardium when investigating cardiac mechanics). Although fully automatic solutions towards myocardial segmentation exist, translating that information in a fast ultrasound scan sequence is not trivial. The aim of this study was therefore to develop a methodology to automatically define the FOV from a volumetric dataset in the context of anatomical scanning. Hereto, a method is proposed where the anatomical relevant space is automatically identified as follows. First, the left ventricular myocardium is localized in the volumetric ultrasound recording using a fully automatic real-time segmentation framework (i.e. BEAS). Then, the extracted meshes are employed to define a binary mask identifying myocardial voxels only. Later, using these binary images, the percentage of pixels along a given image line that belong to the myocardium is calculated. Finally, a spatially continuous FOV that covers 'T' percentage of the myocardium is found by means of a ring-shaped template matching, giving as a result the opening angle and 'thickness' for a conical scan. This approach was tested on 27 volumetric ultrasound datasets, a T = 85% was used. The mean initial opening angle for a conical scan was of 19.67±8.53° while the mean 'thickness' of the cone was 19.01±3.35°. Therefore, a reduction of 48.99% in the number of transmit events was achieved, resulting in a frame rate gain factor of 1.96. As a conclusion, anatomical scanning in combination with new scanning sequences techniques can increase frame rate significantly while keeping information of the relevant structures for functional imaging. © 2015 IEEE.

Abstract