1,721,007 research outputs found
Collaborative Trajectory Mining in Smart-homes to Support Early Diagnosis of Cognitive Decline
Full text linkOur ageing world population claims for innovative tools to support healthcare and independent living. In this paper, we address this challenge by introducing a novel system to recognize symptoms of cognitive decline by exploiting modern smart-home sensors. Previous works tried to recognize wandering of elderly people in outdoor environments. However, the recognition of wandering indoors poses additional challenges. Indeed, several indoor movements resembling wandering may be actually due to the normal execution of daily living activities, or to the particular shape of the home. To address these challenges, we adopt a collaborative learning approach, using a training set of trajectories shared by individuals living in smart-homes. New wandering episodes are classified using a personalized model, built considering the homes' shape and the individuals' profiles. We apply a long-term analysis of classified wandering episodes to provide a hypothesis of diagnosis to be communicated to a medical center for further inspection. We implemented our algorithms and evaluated the system with a large dataset of real-world subjects, including people with dementia, MCI persons, and cognitively healthy people. The results indicate the potential utility of this system to support the early diagnosis of cognitive impairment
Activity Recognition in Smart Homes via Feature-Rich Visual Extraction of Locomotion Traces
Full text linkThe proliferation of sensors in smart homes makes it possible to monitor human activities, routines, and complex behaviors in an unprecedented way. Hence, human activity recognition has gained increasing attention over the last few years as a tool to improve healthcare and well-being in several applications. However, most existing activity recognition systems rely on cameras or wearable sensors, which may be obtrusive and may invade the user's privacy, especially at home. Moreover, extracting expressive features from a stream of data provided by heterogeneous smart-home sensors is still an open challenge. In this paper, we investigate a novel method to detect activities of daily living by exploiting unobtrusive smart-home sensors (i.e., passive infrared position sensors and sensors attached to everyday objects) and vision-based deep learning algorithms, without the use of cameras or wearable sensors. Our method relies on depicting the locomotion traces of the user and visual clues about their interaction with objects on a floor plan map of the home, and utilizes pre-trained deep convolutional neural networks to extract features for recognizing ongoing activity. One additional advantage of our method is its seamless extendibility with additional features based on the available sensor data. Extensive experiments with a real-world dataset and a comparison with state-of-the-art approaches demonstrate the effectiveness of our method
Exploiting Feature Selection in Human Activity Recognition: Methodological Insights and Empirical Results Using Mobile Sensor Data
Full text linkHuman Activity Recognition (HAR) using mobile sensor data has gained increasing attention over the last few years, with a fast-growing number of reported applications. The central role of machine learning in this field has been discussed by a vast amount of research works, with several strategies proposed for processing raw data, extracting suitable features, and inducing predictive models capable of recognizing multiple types of daily activities. Since many HAR systems are implemented in resource-constrained mobile devices, the efficiency of the induced models is a crucial aspect to consider. This paper highlights the importance of exploiting dimensionality reduction techniques that can simplify the model and increase efficiency by identifying and retaining only the most informative and predictive features for activity recognition. More in detail, a large experimental study is presented that encompasses different feature selection algorithms as well as multiple HAR benchmarks containing mobile sensor data. Such a comparative evaluation relies on a methodological framework that is meant to assess not only the extent to which each selection method is effective in identifying the most predictive features but also the overall stability of the selection process, i.e., its robustness to changes in the input data. Although often neglected, in fact, the stability of the selected feature sets is important for a wider exploitability of the induced models. Our experimental results give an interesting insight into which selection algorithms may be most suited in the HAR domain, complementing and significantly extending the studies currently available in this field
Explainable AI-powered Graph Neural Networks for HD EMG-Based Gesture Intention Recognition
Full text linkThe ability to recognize fine-grained gestures enables several applications in different domains, including healthcare, robotics, remote control, and human-computer interaction. Traditional gesture recognition systems rely on data acquired from cameras, depth sensors, or smart gloves. More recently, techniques for recognizing gestures based on signals acquired by high-density (HD) EMG electrodes worn on the forearm have been proposed. An advantage of these techniques is that they do not rely on the use of external devices, and they are feasible also to people who underwent amputation. Unfortunately, the extraction of complex features from raw HD EMG signals may introduce delays that deter the real-time requirements of the system. To address this issue, in a preliminary investigation we proposed to use graph neural networks for gesture recognition from raw HD EMG data. In this paper, we extend our previous work by exploiting Explainable AI algorithms to automatically refine the graph topology based on the data in order to improve recognition rates and reduce the computational cost. We performed extensive experiments with a large dataset collected from 20 volunteers regarding the execution of 65 fine-grained gestures, comparing our technique with state-of-the-art methods based on handcrafted features and different machine learning algorithms. Experimental results show that our technique outperforms the state of the art in terms of recognition performance while incurring significantly lower computational cost at run-time
Towards Vision-based Analysis of Indoor Trajectories for Cognitive Assessment
No full textThe rapid increase of the senior population in our societies calls for innovative tools to early detect symptoms of cognitive decline. To this aim, several methods have been recently proposed that exploit Internet of Things data and artificial intelligence techniques to recognize abnormal behaviors. In particular, the analysis of position traces may enable early detection of cognitive decline. However, indoor movement analysis introduces several challenges. Indeed, indoor movements are constrained by the ambient shape and by the presence of obstacles, and are affected by variability of activity execution. In this paper, we propose a novel method to identify abnormal indoor movement patterns that may indicate cognitive decline according to well known clinical models. Our method relies on trajectory segmentation, visual feature extraction from trajectory segments, and vision-based deep learning on the edge. In order to avoid privacy issues, we rely on indoor localization technologies without the use of cameras. Preliminary experimental results with a real-world dataset gathered from cognitively healthy persons and people with dementia show that this research direction is promising
TraMiner: Vision-Based Analysis of Locomotion Traces for Cognitive Assessment in Smart-Homes
Full text linkThe rapid increase in the senior population is posing serious challenges to national healthcare systems. Hence, innovative tools are needed to early detect health issues, including cognitive decline. Several clinical studies show that it is possible to identify cognitive impairment based on the locomotion patterns of the elderly. In this work, we investigate the use of sensor data and deep learning to recognize those patterns in instrumented smart-homes. In order to get rid of the noise introduced by indoor constraints and activity execution, we introduce novel visual feature extraction methods for locomotion data. Our solution relies on locomotion trace segmentation, image-based extraction of salient features from locomotion segments, and vision-based deep learning. We carried out extensive experiments with a large dataset acquired in a smart-home test bed from 153 seniors, including people with cognitive diseases. Results show that our system can accurately recognize the cognitive status of the senior, reaching a macro-F1 score of 0.873 for the three categories that we target: cognitive health, mild cognitive impairment, and dementia. Moreover, an experimental comparison shows that our system outperforms state-of-the-art methods
Towards EEG-based Performance Assessment in Dataset Annotation Tasks
No full textArtificial intelligence is more and more adopted to complement human activity in solving complex tasks in several domains, including healthcare, security, finance, and automation. In order to be effective, several artificial intelligence tools rely on large training sets of carefully annotated data. Since labeling is mostly performed manually, it is a costly and error-prone process. Hence, there is increasing interest in devising innovative tools to support the annotation task. In this paper, we report an initial investigation on the application of EEG data mining for evaluating the performance of humans carrying out image annotation tasks. Our approach relies on a cheap portable EEG sensor and on supervised learning methods. We collected a dataset from five volunteers, and performed an initial evaluation of our technique. The achieved results are promising, and pave the way to several research directions. To the best of our knowledge, our work is the first one applying EEG data mining for assessing the performance of labelers
Sensor-Based Locomotion Data Mining for Supporting the Diagnosis of Neurodegenerative Disorders: A Survey
Full text linkLocomotion characteristics and movement patterns are reliable indicators of neurodegenerative diseases (NDDs). This survey provides a systematic literature review of locomotion data mining systems for supporting NDD diagnosis. We discuss techniques for discovering low-level locomotion indicators, sensor data acquisition and processing methods, and NDD detection algorithms. The survey presents a comprehensive discussion on the main challenges for this active area, including the addressed diseases, locomotion data types, duration of monitoring, employed algorithms, and experimental validation strategies. We also identify prominent open challenges and research directions regarding ethics and privacy issues, technological and usability aspects, and availability of public benchmarks
Ensembling classical machine learning and deep learning approaches for morbidity identification from clinical notes
Full text linkThe past decade has seen an explosion of the amount of digital information generated within the healthcare domain. Digital data exist in the form of images, video, speech, transcripts, electronic health records, clinical records, and free-text. Analysis and interpretation of healthcare data is a daunting task, and it demands a great deal of time, resources, and human effort. In this paper, we focus on the problem of co-morbidity recognition from patient’s clinical records. To this aim, we employ both classical machine learning and deep learning approaches.We use word embeddings and bag-of-words representations, coupled with feature selection techniques. The goal of our work is to develop a classification system to identify whether a certain health condition occurs for a patient by studying his/her past clinical records. In more detail, we have used pre-trained word2vec, domain-trained, GloVe, fastText, and universal sentence encoder embeddings to tackle the classification of sixteen morbidity conditions within clinical records. We have compared the outcomes of classical machine learning and deep learning approaches with the employed feature representation methods and feature selection methods. We present a comprehensive discussion of the performances and behaviour of the employed classical machine learning and deep learning approaches. Finally, we have also used ensemble learning techniques over a large number of combinations of classifiers to improve the single model performance. For our experiments, we used the n2c2 natural language processing research dataset, released by Harvard Medical School. The dataset is in the form of clinical notes that contain patient discharge summaries. Given the unbalancedness of the data and their small size, the experimental results indicate the advantage of the ensemble learning technique with respect to single classifier models. In particular, the ensemble learning technique has slightly improved the performances of single classification models but has greatly reduced the variance of predictions stabilizing the accuracies (i.e., the lower standard deviation in comparison with single classifiers). In real-life scenarios, our work can be employed to identify with high accuracy morbidity conditions of patients by feeding our tool with their current clinical notes. Moreover, other domains where classification is a common problem might benefit from our approach as well
- …
