Background Diaphragm muscle atrophy during mechanical ventilation begins within 24 h and progresses rapidly with significant clinical consequences. Electrical stimulation of the phrenic nerves using invasive electrodes has shown promise in maintaining diaphragm condition by inducing intermittent diaphragm muscle contraction. However, the widespread application of these methods may be limited by their risks as well as the technical and environmental requirements of placement and care. Non‐invasive stimulation would offer a valuable alternative method to maintain diaphragm health while overcoming these limitations. Methods We applied non‐invasive electrical stimulation to the phrenic nerve in the neck in healthy volunteers. Respiratory pressure and flow, diaphragm electromyography and mechanomyography, and ultrasound visualization were used to assess the diaphragmatic response to stimulation. The electrode positions and stimulation parameters were systematically varied in order to investigate the influence of these parameters on the ability to induce diaphragm contraction with non‐invasive stimulation. Results We demonstrate that non‐invasive capture of the phrenic nerve is feasible using surface electrodes without the application of pressure, and characterize the stimulation parameters required to achieve therapeutic diaphragm contractions in healthy volunteers. We show that an optimal electrode position for phrenic nerve capture can be identified and that this position does not vary as head orientation is changed. The stimulation parameters required to produce a diaphragm response at this site are characterized and we show that burst stimulation above the activation threshold reliably produces diaphragm contractions sufficient to drive an inspired volume of over 600 ml, indicating the ability to produce significant diaphragmatic work using non‐invasive stimulation. Conclusion This opens the possibility of non‐invasive systems, requiring minimal specialist skills to set up, for maintaining diaphragm function in the intensive care setting.

No abstract

No abstract

This paper proposes the addition of a thermal camera to an RGB system with the goal of improving person and road detection reliability in unfavorable weather and illumination conditions. Custom data is gathered on an experimental vehicle and used for development and testing. For person detection, we propose a novel multi-modal approach, where bounding boxes are initially obtained from RGB and thermal images using YOLOv3-tiny. We then identify high-intensity connected components in thermal images to compensate for missed detections. Detections from the two cameras and the two algorithms are finally weighed and combined into a confidence map. Using the proposed fusion method, recall and precision are improved compared to using RGB only, without the need to retrain the network. For thermal-based road segmentation, we achieve an average precision of 94.2% after re-training MultiNet's KittiSeg decoder on a small thermal dataset, while using pre-trained weights for MultiNet's VGG-based encoder. These results show that the addition of thermal cameras to perception systems of autonomous vehicles can bring substantial benefits with minimal labelling, implementation effort and training requirements.

No abstract