A multifunctional, wearable sensor based on a reduced oxide graphene (rGO) film onto a porous inverse opal acetylcellulose (IOAC) film has been developed and can perform simultaneous, in situ monitoring of various human motions and ion concentrations in sweat. The rGO film is used as a strain-sensing layer for monitoring human motion via its resistance change, whereas the porous IOAC film is used as a flexible microstructured substrate not only for high sensitive motion sensing, but also for collection and analysis of ion concentrations in sweat by its simple colorimetric changes or reflection-peak shifts. Studies on humans demonstrated that the devices have excellent capability for monitoring various human motions, such as finger bending motion, wrist bending motion, head rotation motion and various small-scale motions of the throat. Simultaneous, in situ analysis of the ion concentration in sweat during these motions shows that the IOAC substrate can detect a wide range of NaCl concentrations in sweat from normal 30 to 680 mM under the conditions of severe dehydration. This investigation provides new horizons toward the design and fabrication of multifunctional, wearable health monitoring devices and the proposed wearable sensor shows promising applications in healthcare and preventive medicine.
In this study, a multifunctional wearable sensing device based on two different graphene films is fabricated and can achieve the simultaneous detection of physiological signals and volatile organic compound (VOC) biomarkers without mutual signal interference. The wearable device was designed with two sensing components: on the upper layer of the device, four kinds of porphyrin-modified reduced graphene oxide (rGO) films were prepared and used for a sensor array that could sufficiently react with VOC vapors to achieve highly sensitive detection. A porous rGO film was designed on the underlayer of the device and used as a strain-sensing matrix, which could be closely attached to the skin to achieve a highly sensitive detection of the physiological signal. A polyimide film between the two sensing components was used not only as a flexible substrate, but also as a protective layer to avoid the porous rGO film's response to VOC molecules. Investigation of the detection ability showed that the porous rGO strain-sensing matrix can achieve a higher gauge factor (282.28) than the unstructured rGO counterpart (8.96) and is more desirable for the detection of physiological motion. In contrast, the porphyrin-modified rGO sensor array displayed a superior response to VOC vapors, and eight different VOC biomarkers could be detected and discriminated using the as-prepared sensor array together with a pattern recognition approach. The multifunctional sensing devices displayed excellent ability for the detection of a variety of human physiological signals, such as pulse and respiration rates. Simultaneous analysis of simulated diabetic breath samples, simulated nephrotic breath samples, and breath samples exhaled by healthy individuals using our wearable device exhibited clear identification and discrimination. Our study provides new insights into fabrication and design of multifunctional sensing devices without signal interference, and the application of the proposed devices are promising in preventive medicine and health care.
Semantic segmentation with deep learning has achieved great progress in classifying the pixels in the image. However, the local location information is usually ignored in the high-level feature extraction by the deep learning, which is important for image semantic segmentation. To avoid this problem, we propose a graph model initialized by a fully convolutional network (FCN) named Graph-FCN for image semantic segmentation. Firstly, the image grid data is extended to graph structure data by a convolutional network, which transforms the semantic segmentation problem into a graph node classification problem. Then we apply graph convolutional network to solve this graph node classification problem. As far as we know, it is the first time that we apply the graph convolutional network in image semantic segmentation. Our method achieves competitive performance in mean intersection over union (mIOU) on the VOC dataset(about 1.34% improvement), compared to the original FCN model.
Progress in cross-lingual modeling depends on challenging, realistic, and diverse evaluation sets. We introduce Multilingual Knowledge Questions and Answers (MKQA), an open- domain question answering evaluation set comprising 10k question-answer pairs aligned across 26 typologically diverse languages (260k question-answer pairs in total). Answers are based on heavily curated, language- independent data representation, making results comparable across languages and independent of language-specific passages. With 26 languages, this dataset supplies the widest range of languages to-date for evaluating question answering. We benchmark a variety of state- of-the-art methods and baselines for generative and extractive question answering, trained on Natural Questions, in zero shot and translation settings. Results indicate this dataset is challenging even in English, but especially in low-resource languages.1
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.