This paper presents a low-cost rehabilitation support system for lower extremity muscle strengthening exercises: LOWER-LIMB FOLLOW-UP. The system includes two main modules: a goal-oriented serious computer game module for the patients and another that provides feedback to the physiotherapist. By analyzing the surface electromyography signals obtained from the relevant muscles of the patient, it is ensured that the patient gains points in serious computer game. The raw surface electromyography signals generated while playing the games and the result of the signal analysis are saved in the database. Feature extraction methods are used for the electromyography signal analysis. The physiotherapists can access the database via a web-based application and obtain information about their patients' performance. Muscle strengthening exercises frequently recommended, such as active knee extension exercise for quadriceps, active knee flexion exercise for hamstring and terminal extension exercise for vastus medialis obliquus muscle groups, are selected for this study.
{"title":"LOWER-LIMB FOLLOW-UP: A Surface Electromyography Based Serious Computer Game and Patient Follow-Up System for Lower Extremity Muscle Strengthening Exercises in Physiotherapy and Rehabilitation","authors":"Tugba Günaydin, R. Arslan","doi":"10.1109/CBMS.2019.00103","DOIUrl":"https://doi.org/10.1109/CBMS.2019.00103","url":null,"abstract":"This paper presents a low-cost rehabilitation support system for lower extremity muscle strengthening exercises: LOWER-LIMB FOLLOW-UP. The system includes two main modules: a goal-oriented serious computer game module for the patients and another that provides feedback to the physiotherapist. By analyzing the surface electromyography signals obtained from the relevant muscles of the patient, it is ensured that the patient gains points in serious computer game. The raw surface electromyography signals generated while playing the games and the result of the signal analysis are saved in the database. Feature extraction methods are used for the electromyography signal analysis. The physiotherapists can access the database via a web-based application and obtain information about their patients' performance. Muscle strengthening exercises frequently recommended, such as active knee extension exercise for quadriceps, active knee flexion exercise for hamstring and terminal extension exercise for vastus medialis obliquus muscle groups, are selected for this study.","PeriodicalId":311634,"journal":{"name":"2019 IEEE 32nd International Symposium on Computer-Based Medical Systems (CBMS)","volume":"32 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134256818","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The volume and quality of patient data stored and collected have drastically grown in the last years. Such data can be analyzed by machine learning algorithms to improve health and well-being. However, while the distribution of data is benefitial, it should be performed in a way that preserves patient privacy. It would be expected to obtain useful information from the use of machine learning algorithms applied to both anonymized and non-anonymized datasets. However, those algorithms can generate lower quality results (even invalid ones) due to information loss during the anonymization process. We aim to analyze the relationship between anonymization and data utility/information loss, through the use of different algorithms and information loss metrics. With that aim, we plan to 1) analyze how real algorithms used on real data are affected by different anonymization techniques; 2) to use the lessons learned to design useful metrics for measuring the information loss after annonymization; and 3) to validate the proposed metrics by testing them in other environments with different types of data. The expected contributions of the research will be to obtain more information about how anonymization techniques affect the data usefulness, together with additional knowledge about the more suitable machine learning algorithms to be used to anonymized data, and a set of metrics to measure the usefulness of anonymized data would be developed
{"title":"Towards the Analysis of How Anonymization Affects Usefulness of Health Data in the Context of Machine Learning","authors":"Fer Carmona, J. Conesa, Jordi Casas-Roma","doi":"10.1109/CBMS.2019.00126","DOIUrl":"https://doi.org/10.1109/CBMS.2019.00126","url":null,"abstract":"The volume and quality of patient data stored and collected have drastically grown in the last years. Such data can be analyzed by machine learning algorithms to improve health and well-being. However, while the distribution of data is benefitial, it should be performed in a way that preserves patient privacy. It would be expected to obtain useful information from the use of machine learning algorithms applied to both anonymized and non-anonymized datasets. However, those algorithms can generate lower quality results (even invalid ones) due to information loss during the anonymization process. We aim to analyze the relationship between anonymization and data utility/information loss, through the use of different algorithms and information loss metrics. With that aim, we plan to 1) analyze how real algorithms used on real data are affected by different anonymization techniques; 2) to use the lessons learned to design useful metrics for measuring the information loss after annonymization; and 3) to validate the proposed metrics by testing them in other environments with different types of data. The expected contributions of the research will be to obtain more information about how anonymization techniques affect the data usefulness, together with additional knowledge about the more suitable machine learning algorithms to be used to anonymized data, and a set of metrics to measure the usefulness of anonymized data would be developed","PeriodicalId":311634,"journal":{"name":"2019 IEEE 32nd International Symposium on Computer-Based Medical Systems (CBMS)","volume":"10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129042769","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Wellington S. Silva, Daniel L. Jasbick, R. E. Wilson, P. M. A. Marques, A. Traina, Lúcio F. D. Santos, A. E. Jorge, Daniel de Oliveira, M. Bedo
Tissue segmentation in photographs of lower limb chronic ulcers is a non-intrusive approach that supports dermatological analyses. This paper presents 2PLA, a method that combines supervised and unsupervised learning strategies for enhancing the segmentation of dermatological wounds. Given an ulcer photo captured according to a fixed protocol, 2PLA first phase performs a pixelwise classification of points of interest, whereas pre-processing filters are employed for the smoothing of image noise. The cleaned image is further sent to the 2PLA divide-and-conquer second phase. It builds upon SLIC superpixel construction algorithm for dividing the lower limb into regions of interest with well-defined borders, and clusters the superpixels by taking advantage of the similarity-based DBSCAN algorithm. We set up the phases of our method by using a real annotated set of dermatological wounds, and empirical evaluations on representative samples up to 100,000 points showed a compact Multi-Layer Perceptron with Levenberg-Marquardt training algorithm (Cohen-Kappa = .971, Sensitivity = .98, and Specificity = .98) outperformed other classifiers as 2PLA first phase. Additionally, experimental trials on DBSCAN with five distance functions (L1, L2, Loo, Canberra, and BrayCurtis) indicated L1 function provided fewer groups in comparison to the competitors, and the number of clusters was an exponential decay to the similarity ratio. Accordingly, we used the elbow criterion for finding the L1-based DBSCAN threshold as 2PLA second phase parameterization. We evaluated the fine-tuned setting of our method over a labeled set of ulcer images, and wounded tissues were segmented within a .05 Mean Absolute Error ratio. These results illustrate the impact of learning parameters on 2PLA as well as the method efficacy for wound segmentation.
{"title":"A Two-Phase Learning Approach for the Segmentation of Dermatological Wounds","authors":"Wellington S. Silva, Daniel L. Jasbick, R. E. Wilson, P. M. A. Marques, A. Traina, Lúcio F. D. Santos, A. E. Jorge, Daniel de Oliveira, M. Bedo","doi":"10.1109/CBMS.2019.00076","DOIUrl":"https://doi.org/10.1109/CBMS.2019.00076","url":null,"abstract":"Tissue segmentation in photographs of lower limb chronic ulcers is a non-intrusive approach that supports dermatological analyses. This paper presents 2PLA, a method that combines supervised and unsupervised learning strategies for enhancing the segmentation of dermatological wounds. Given an ulcer photo captured according to a fixed protocol, 2PLA first phase performs a pixelwise classification of points of interest, whereas pre-processing filters are employed for the smoothing of image noise. The cleaned image is further sent to the 2PLA divide-and-conquer second phase. It builds upon SLIC superpixel construction algorithm for dividing the lower limb into regions of interest with well-defined borders, and clusters the superpixels by taking advantage of the similarity-based DBSCAN algorithm. We set up the phases of our method by using a real annotated set of dermatological wounds, and empirical evaluations on representative samples up to 100,000 points showed a compact Multi-Layer Perceptron with Levenberg-Marquardt training algorithm (Cohen-Kappa = .971, Sensitivity = .98, and Specificity = .98) outperformed other classifiers as 2PLA first phase. Additionally, experimental trials on DBSCAN with five distance functions (L1, L2, Loo, Canberra, and BrayCurtis) indicated L1 function provided fewer groups in comparison to the competitors, and the number of clusters was an exponential decay to the similarity ratio. Accordingly, we used the elbow criterion for finding the L1-based DBSCAN threshold as 2PLA second phase parameterization. We evaluated the fine-tuned setting of our method over a labeled set of ulcer images, and wounded tissues were segmented within a .05 Mean Absolute Error ratio. These results illustrate the impact of learning parameters on 2PLA as well as the method efficacy for wound segmentation.","PeriodicalId":311634,"journal":{"name":"2019 IEEE 32nd International Symposium on Computer-Based Medical Systems (CBMS)","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122707011","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
P. Kroupa, Caroline Morton, K. L. Calvez, Matt Williams
Survival prediction is a key task in medicine. Existing models are based on statistical techniques, such as the Cox models and there is limited work on the application of machine learning. In this paper we demonstrate that the K-Nearest Neighbour algorithm can be used for survival prediction. We show that its performance is as good as that of standard techniques, and that it provides a clear interpretation of the results. We show that pre-processing methods improve performance, and evaluate the performance across 20 different datasets with differing properties to show that the model performs well under various conditions. For low event rate datasets we show that KNN can outperform the Cox model.
{"title":"Assessing K-Nearest Neighbours Algorithm for Simple, Interpretable Time-to-Event Survival Predictions Over a Range of Simulated Datasets","authors":"P. Kroupa, Caroline Morton, K. L. Calvez, Matt Williams","doi":"10.1109/CBMS.2019.00080","DOIUrl":"https://doi.org/10.1109/CBMS.2019.00080","url":null,"abstract":"Survival prediction is a key task in medicine. Existing models are based on statistical techniques, such as the Cox models and there is limited work on the application of machine learning. In this paper we demonstrate that the K-Nearest Neighbour algorithm can be used for survival prediction. We show that its performance is as good as that of standard techniques, and that it provides a clear interpretation of the results. We show that pre-processing methods improve performance, and evaluate the performance across 20 different datasets with differing properties to show that the model performs well under various conditions. For low event rate datasets we show that KNN can outperform the Cox model.","PeriodicalId":311634,"journal":{"name":"2019 IEEE 32nd International Symposium on Computer-Based Medical Systems (CBMS)","volume":"71 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126834361","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
H. Kondylakis, L. Koumakis, Dimitrios G. Katehakis, A. Kouroubali, K. Marias, M. Tsiknakis, P. Simos, E. Karademas
Breast cancer is the most common cancer disease in women and is rapidly becoming a chronic illness due recent advances in treatment methods. As such, coping with cancer has become a major socio-economic challenge leading to an increasing need for predicting resilience of women to the variety of stressful experiences and practical challenges they face. In this paper, we present the data infrastructure developed for this purpose, demonstrating the various components that will contribute to the developing the resilience trajectory predictor. Special emphasis is given to the semantic tier, presenting the project solution already implemented for effectively collecting, ingesting, cleaning, modelling and processing data that will be used throughout the lifetime of the project.
{"title":"Developing a Data Infrastructure for Enabling Breast Cancer Women to BOUNCE Back","authors":"H. Kondylakis, L. Koumakis, Dimitrios G. Katehakis, A. Kouroubali, K. Marias, M. Tsiknakis, P. Simos, E. Karademas","doi":"10.1109/CBMS.2019.00134","DOIUrl":"https://doi.org/10.1109/CBMS.2019.00134","url":null,"abstract":"Breast cancer is the most common cancer disease in women and is rapidly becoming a chronic illness due recent advances in treatment methods. As such, coping with cancer has become a major socio-economic challenge leading to an increasing need for predicting resilience of women to the variety of stressful experiences and practical challenges they face. In this paper, we present the data infrastructure developed for this purpose, demonstrating the various components that will contribute to the developing the resilience trajectory predictor. Special emphasis is given to the semantic tier, presenting the project solution already implemented for effectively collecting, ingesting, cleaning, modelling and processing data that will be used throughout the lifetime of the project.","PeriodicalId":311634,"journal":{"name":"2019 IEEE 32nd International Symposium on Computer-Based Medical Systems (CBMS)","volume":"106 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127532911","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Lung cancer is the leading cause of cancer mortality, accounting for approximately 20% of all cancer-related deaths. Patients diagnosed in the early stages have a 1-year survival rate of 81-85% while in an advanced stage have 15-19% chances of survival. Therefore, it is very necessary to diagnose lung cancer in early stages in malignant or benign, when the nodules are still very small, but it is a complex task even for experienced specialists and presents some challenges. To assist specialists, computer-aided diagnosis systems have been used to improve the accuracy in the diagnosis. In this paper, we exploit the use of a technique of hyperparameter tuning to find the best architecture of a Convolutional Neural Network to classify small pulmonary nodules balanced with diameter 5-10mm. The best results achieved were an error rate of 12%, sensitivity of 94%, specificity of 83%, accuracy of 88% and F-measure of 89%
{"title":"Efficient Hyperparameter Optimization of Convolutional Neural Networks on Classification of Early Pulmonary Nodules","authors":"Lucas L. Lima, J. Ferreira, M. C. Oliveira","doi":"10.1109/CBMS.2019.00039","DOIUrl":"https://doi.org/10.1109/CBMS.2019.00039","url":null,"abstract":"Lung cancer is the leading cause of cancer mortality, accounting for approximately 20% of all cancer-related deaths. Patients diagnosed in the early stages have a 1-year survival rate of 81-85% while in an advanced stage have 15-19% chances of survival. Therefore, it is very necessary to diagnose lung cancer in early stages in malignant or benign, when the nodules are still very small, but it is a complex task even for experienced specialists and presents some challenges. To assist specialists, computer-aided diagnosis systems have been used to improve the accuracy in the diagnosis. In this paper, we exploit the use of a technique of hyperparameter tuning to find the best architecture of a Convolutional Neural Network to classify small pulmonary nodules balanced with diameter 5-10mm. The best results achieved were an error rate of 12%, sensitivity of 94%, specificity of 83%, accuracy of 88% and F-measure of 89%","PeriodicalId":311634,"journal":{"name":"2019 IEEE 32nd International Symposium on Computer-Based Medical Systems (CBMS)","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131371138","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In this paper we propose a new approach for detecting temporary cognitive overload. Due to the raising propagation of wearable devices with various integrated sensors, the idea is to detect such overload situations based on acceleration data out of these sensors at task relevant body parts. We executed an experiment in order to investigate the performance differences of people in a relaxed state and under cognitive load. The loaded state was simulated in a dual-task test. Additionally, we analyzed changes in the participants' motion behaviors at their hips and both of their wrists. We could show, that dual-task measuring is a suitable way for generating ground truth data for cognitive load. For this reason we used the study's data also as ground truth for the subsequent developed classification system. After investigating different features from the data we could discriminate the two states ("relaxed" and "loaded") with an accuracy of 90% and an MCC of 0.7986, which indicates a high correlation between ground truth and classified data. That outperforms other ACC based systems and approaches the performance of vital parameter based ones. Moreover, it could be shown that the dominant hand's data have greater influence to the results than the recessive one's. However, using data from both hands leads to further improvements.
{"title":"Using Acceleration Data for Detecting Temporary Cognitive Overload in Health Care Exemplified Shown in a Pill Sorting Task","authors":"L. Kohout, Manuel Butz, W. Stork","doi":"10.1109/CBMS.2019.00015","DOIUrl":"https://doi.org/10.1109/CBMS.2019.00015","url":null,"abstract":"In this paper we propose a new approach for detecting temporary cognitive overload. Due to the raising propagation of wearable devices with various integrated sensors, the idea is to detect such overload situations based on acceleration data out of these sensors at task relevant body parts. We executed an experiment in order to investigate the performance differences of people in a relaxed state and under cognitive load. The loaded state was simulated in a dual-task test. Additionally, we analyzed changes in the participants' motion behaviors at their hips and both of their wrists. We could show, that dual-task measuring is a suitable way for generating ground truth data for cognitive load. For this reason we used the study's data also as ground truth for the subsequent developed classification system. After investigating different features from the data we could discriminate the two states (\"relaxed\" and \"loaded\") with an accuracy of 90% and an MCC of 0.7986, which indicates a high correlation between ground truth and classified data. That outperforms other ACC based systems and approaches the performance of vital parameter based ones. Moreover, it could be shown that the dominant hand's data have greater influence to the results than the recessive one's. However, using data from both hands leads to further improvements.","PeriodicalId":311634,"journal":{"name":"2019 IEEE 32nd International Symposium on Computer-Based Medical Systems (CBMS)","volume":"237 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133484475","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
F. Buendía, Joaquín Gayoso-Cabada, J. A. J. Méndez, J. Sierra
In this paper, we describe how to transform unstructured free-text clinical corpora, made from reports written in natural language and complementary assets (e.g., medical images, laboratory results, etc.), into collections of digital objects compatible with Clavy, a tool for managing reconfigurable digital collections. It will allow healthcare experts to subsequently reorganize the resulting collections to adapt them to their specific needs. The transformation will be achieved through the use of MetaMap, a robust tool for mapping clinical texts into the UMLS (Unified Medical Language System) thesaurus. Thus, by processing reports with MetaMap, we will be able to extract a significant set of corpus-specific UMLS terms, grouped according to relevant semantic types, which will be used to support a preliminary organization of the resources in the Clavy collection. We illustrate the viability of the approach with the generation of a reconfigurable Clavy collection from the Indiana Chest X-ray corpus of radiology reports and images. On the basis of this case study, we also discuss the strengths and weaknesses of the approach proposed.
{"title":"Transforming Unstructured Clinical Free-Text Corpora into Reconfigurable Medical Digital Collections","authors":"F. Buendía, Joaquín Gayoso-Cabada, J. A. J. Méndez, J. Sierra","doi":"10.1109/CBMS.2019.00105","DOIUrl":"https://doi.org/10.1109/CBMS.2019.00105","url":null,"abstract":"In this paper, we describe how to transform unstructured free-text clinical corpora, made from reports written in natural language and complementary assets (e.g., medical images, laboratory results, etc.), into collections of digital objects compatible with Clavy, a tool for managing reconfigurable digital collections. It will allow healthcare experts to subsequently reorganize the resulting collections to adapt them to their specific needs. The transformation will be achieved through the use of MetaMap, a robust tool for mapping clinical texts into the UMLS (Unified Medical Language System) thesaurus. Thus, by processing reports with MetaMap, we will be able to extract a significant set of corpus-specific UMLS terms, grouped according to relevant semantic types, which will be used to support a preliminary organization of the resources in the Clavy collection. We illustrate the viability of the approach with the generation of a reconfigurable Clavy collection from the Indiana Chest X-ray corpus of radiology reports and images. On the basis of this case study, we also discuss the strengths and weaknesses of the approach proposed.","PeriodicalId":311634,"journal":{"name":"2019 IEEE 32nd International Symposium on Computer-Based Medical Systems (CBMS)","volume":"208 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115541531","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
F. D. Casagrande, O. O. Nedrejord, Wonho Lee, E. Zouganeli
We report work in progress from interdisciplinary research on Assisted Living Technology in smart homes for older adults with mild cognitive impairments or dementia. We present our field trial, the set-up for collecting and storing data from real homes, and preliminary results on action recognition using low resolution depth video cameras. The data have been collected from seven apartments with one resident each over a period of two weeks. We propose a pre-processing of the depth videos by applying an Infinite Response Filter (IIR) for extracting the movements in the frames prior to classification. In this work we classify four actions: TV interaction (turn it on/ off and switch over), standing up, sitting down, and no movement. Our first results indicate that using the IIR filter for movement information extraction improves accuracy and can be an efficient method for recognizing actions. Our current implementation uses a convolutional long short-term memory (ConvLSTM) neural network, and achieved an average peak accuracy of 86%.
{"title":"Action Recognition in Real Homes using Low Resolution Depth Video Data","authors":"F. D. Casagrande, O. O. Nedrejord, Wonho Lee, E. Zouganeli","doi":"10.1109/CBMS.2019.00041","DOIUrl":"https://doi.org/10.1109/CBMS.2019.00041","url":null,"abstract":"We report work in progress from interdisciplinary research on Assisted Living Technology in smart homes for older adults with mild cognitive impairments or dementia. We present our field trial, the set-up for collecting and storing data from real homes, and preliminary results on action recognition using low resolution depth video cameras. The data have been collected from seven apartments with one resident each over a period of two weeks. We propose a pre-processing of the depth videos by applying an Infinite Response Filter (IIR) for extracting the movements in the frames prior to classification. In this work we classify four actions: TV interaction (turn it on/ off and switch over), standing up, sitting down, and no movement. Our first results indicate that using the IIR filter for movement information extraction improves accuracy and can be an efficient method for recognizing actions. Our current implementation uses a convolutional long short-term memory (ConvLSTM) neural network, and achieved an average peak accuracy of 86%.","PeriodicalId":311634,"journal":{"name":"2019 IEEE 32nd International Symposium on Computer-Based Medical Systems (CBMS)","volume":"120 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128086260","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Muntazir Mehdi, Denis Schwager, R. Pryss, W. Schlee, M. Reichert, F. Hauck
Tinnitus is a disorder that is not entirely understood, and many of its correlations are still unknown. On the other hand, smartphones became ubiquitous. Their modern versions provide high computational capabilities, reasonable battery size, and a bunch of embedded high-quality sensors, combined with an accepted user interface and an application ecosystem. For tinnitus, as for many other health problems, there are a number of apps trying to help patients, therapists, and researchers to get insights into personal characteristics but also into scientific correlations as such. In this paper, we present the first approach to an app in this context, called TinnituSense that does automatic sensing of related characteristics and enables correlations to the current condition of the patient by a combined participatory sensing, e.g., a questionnaire. For tinnitus, there is a strong hypothesis that weather conditions have some influence. Our proof-of-concept implementation records weather-related sensor data and correlates them to the standard Tinnitus Handicap Inventory (THI) questionnaire. Thus, TinnituSense enables therapists and researchers to collect evidence for unknown facts, as this is the first opportunity to correlate weather to patient conditions on a larger scale. Our concept as such is limited neither to tinnitus nor to built-in sensors, e.g., in the tinnitus domain, we are experimenting with mobile EEG sensors. TinnituSense is faced with several challenges of which we already solved principle architecture, sensor management, and energy consumption.
{"title":"Towards Automated Smart Mobile Crowdsensing for Tinnitus Research","authors":"Muntazir Mehdi, Denis Schwager, R. Pryss, W. Schlee, M. Reichert, F. Hauck","doi":"10.1109/CBMS.2019.00026","DOIUrl":"https://doi.org/10.1109/CBMS.2019.00026","url":null,"abstract":"Tinnitus is a disorder that is not entirely understood, and many of its correlations are still unknown. On the other hand, smartphones became ubiquitous. Their modern versions provide high computational capabilities, reasonable battery size, and a bunch of embedded high-quality sensors, combined with an accepted user interface and an application ecosystem. For tinnitus, as for many other health problems, there are a number of apps trying to help patients, therapists, and researchers to get insights into personal characteristics but also into scientific correlations as such. In this paper, we present the first approach to an app in this context, called TinnituSense that does automatic sensing of related characteristics and enables correlations to the current condition of the patient by a combined participatory sensing, e.g., a questionnaire. For tinnitus, there is a strong hypothesis that weather conditions have some influence. Our proof-of-concept implementation records weather-related sensor data and correlates them to the standard Tinnitus Handicap Inventory (THI) questionnaire. Thus, TinnituSense enables therapists and researchers to collect evidence for unknown facts, as this is the first opportunity to correlate weather to patient conditions on a larger scale. Our concept as such is limited neither to tinnitus nor to built-in sensors, e.g., in the tinnitus domain, we are experimenting with mobile EEG sensors. TinnituSense is faced with several challenges of which we already solved principle architecture, sensor management, and energy consumption.","PeriodicalId":311634,"journal":{"name":"2019 IEEE 32nd International Symposium on Computer-Based Medical Systems (CBMS)","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2019-06-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127628715","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：