A System for Wound Evaluation Support Using Depth and Image Sensors.

This paper proposes an evaluation/treatment sup-port system enabling automatic determination of wound evaluation indices from RGB-depth images and fully convolutional networks (FCNs). Segmentation experiments based on wound images and surface area determination experiments based on artificial images showed reduced errors and smaller parameters/higher levels of tissue classification than with previous approaches (proposed: 65.8 %; conventional: 60.2 %), thereby demonstrating the effectiveness of the technique.

Effects of somatosensory information provision to fingertips for mitigation of postural sway and promotion of muscle coactivation in an upright posture.

This study was conducted to investigate the effects of somatosensory information provision to the fingertips using a device proposed by the authors for mitigation of postural sway and muscle co-contraction in an upright posture, which both increase with aging and result in inefficient postural control. In the research, center of pressure (CoP) fluctuation index values and muscle co-contraction for ankle joint movement were monitored with healthy young adults in a standing position. The results showed that the proposed device helped to reduce the root mean square (RMS) of the CoP and muscle co-contraction in the right ankle joint, thereby suggesting its potential for contribution to the assistance of efficient postural control.

A novel hidden Markov model-based pattern discrimination method with the anomaly detection for EMG signals.

This paper proposes a novel sequential pattern recognition method enabling calculation of a posteriori probability for learned and unlearned classes. In this approach, probability density functions of unlearned classes are incorporated in a hiddenMarkov model to classify undefined classes via model parameter estimation using given learning samples. The technique can be applied to various pattern recognition problems such as motion classification with electromyogram (EMG) signals and in support for disease diagnosis. In the experiments conducted, motion classification from EMG signals was implemented with three subjects for eight learned/unlearned forearm motions. The proposed method produced higher levels of classification performance (learned motions: 90.13%; unlearned motions: 91.25%) than previous approaches. The results demonstrated the effectiveness of the technique.