IEEE International Conference on E-health Networking, Application & Services (IEEE Healthcom 2021)

Session OR-2: Opening Remarks
Session KN3: Keynote 3
Session VCB-3: Virtual Coffee Break Session VLB-2: Virtual Lunch Break Session VCB-4: Virtual Coffee Break Session VDB-2: Virtual Dinner Break
Session S4-R1: Medical, Biomedical, and Health Informatics Session S4-R2: E-Health Services and Applications Session S4-R3: Communications and Networking Session S5-R1: Medical, Biomedical, and Health Informatics Session S5-R2: E-Health Services and Applications Session S5-R3: Communications and Networking Session S6-R1: Medical, Biomedical, and Health Informatics Session S6-R2: Communications and Networking
Session T2: Tutorial: Transdisciplinary Methods For Interdisciplinary Research in IoT-Based Cyber-Physical Systems and Human-Machine Interfaces
Session D2-R1: Medical, Biomedical, and Health Informatics Session D2-R2: E-Health Services and Applications Session D2-R3: Communications and Networking

Session S4-R1

Medical, Biomedical, and Health Informatics

Conference
10:30 AM — 12:30 PM HKT
Local
Mar 1 Mon, 9:30 PM — 11:30 PM EST

A Novel Method to Quantify Chest Compressions and Physical Ergonomics During Neonatal Resuscitation

Patrick Scott and Md Shaad Mahmud (University of New Hampshire, USA)

1
Initiation of breathing is critical in the physiologic adaptation from intra-uterine to extra-uterine life. Approximately 10% of newborns require some assistance and stimulation to initiate breathing at birth and about 3% require extensive resuscitative measures. For decades, standardized training and algorithmic approach such as Neonatal Resuscitation Program (NRP) has been used to guide and assess health care professionals caring for neonates at birth. However, despite this, neonatal resuscitation remains a complex endeavor prone to a relatively high frequency (up to 23%) of error. Physical ergonomics is the key science that is focused on ensuring work environment adaptation to human talents, abilities, skills, and limits. We propose a novel method to achieve real-time quantitative assessment of the chest compression quality for neonates, such as ergonomics, compression rate and depth in order to improve performance of medical personnel. In order for the device to be both accurate and completely non-invasive, a high precision laser depth sensor is used to collect data, which is plotted in real time and undergoes analysis after a trial. The real-time feedback algorithm proved to be reliable and accurate when tested retrospectively with human trial. Along with closely analyzing the compression characteristics it it also important to monitor the upper body ergonomics as it heavily plays into the effectiveness of the administered CPR. This is to be done through flexible angular displacement sensors placed on upper body joints to closely monitor changes in upper body positioning.

A 3D-printed EEG based prosthetic arm

Josue Fuentes-Gonzalez, Andres Infante-Alarcón, Víctor Asanza and Francis Loayza (Escuela Superior Politécnica del Litoral, Ecuador)

1
Nowadays, with the use of 3D printers, the upper limb prosthesis is more available, mainly the myoelectric controlled devices; however, there are some cases where the myoelectric signal is difficult to detect due to the patient's injury. In this case, the Electroencephalography (EEG) controlled prosthesis is more suitable. In this work, we reported a case of a man of 64 years old who received a 13.2 kV. electric shock in his right hand, registering a low electromyographic signal. The prosthetic arm was designed using Blender software, sized and adjusted using the left hand as a reference. All the prosthetic components were 3D printed: hinges with flexible material and the rest with polylactide (PLA). We used three actuators placed on the forearm, which connected through nylon thread to the fingers: one motor for the thumb and the other two motors connected to two fingers for each one. We used the Neurosky Mindwave 2 equipped with a frontal sensor and wireless data transmission and a control unit to obtain the EEG data. The data acquired from the EEG device was filtered and adjusted to the patient's capabilities to find a thresholding value that will be used as a control parameter. As a result, the 3D printed prosthesis was fitted to the patient's stump. The patient was able to perform opening and closing movements of the hand with a force of 11.0 N, able to grab objects used in daily life. We can demonstrate the feasibility of simple control of prostheses by recording EEG data, especially for those whose electromyography signals are challenging to read.

Prediction of COVID-19 using time-sliding window: a case study with Piauí State - Brazil

Patrick Santos, Lucas Souza, Samuel Lelis and Hector Ribeiro (Federal University of Piauí, Brazil); Fabbio Borges (USP, Brazil); Romuere R. V. e Silva and Antônio Oseas (Universidade Federal do Piauí, Brazil); Flávio Araújo (Federal University of Piauí, Brazil); Ricardo A. L. Rabelo (Federal University of Piaui (UFPI), Brazil); Joel J. P. C. Rodrigues (Federal University of Piauí (UFPI), Brazil & Instituto de Telecomunicações, Portugal)

1
COVID-19 is an infectious disease caused by a type of coronavirus recently discovered, called SARS-CoV-2. It has infected more than 20 million people worldwide and it is responsible for more than 737,000 deaths. This work presents a study that explores linear regression mechanisms combined with a sliding and cumulative time window approach to provide inputs to assist in decision making for public policies, within the scope of the COVID-19 pandemic evolution, whether they are hardening or easing the isolation. Data from five states of Brazil were collected and applied a Ridge regression to predict the curve behavior of cases and deaths of COVID-19. As a result, an Explained Variance Status (EVS) up to 0.998 and 0.999 is presented, considering cases and deaths, respectively. It was concluded that sliding time window bring more information about the infection than cumulative, since public policy changes in a few time-lapse.

Automatic Diagnostic of the Presence of Exudates in Retinal Images Using Deep Learning

Deusimar Sousa (Federal University of Piauí, Brazil); Antônio Oseas (Universidade Federal do Piauí, Brazil); Ricardo A. L. Rabelo (Federal University of Piaui (UFPI), Brazil); Joel J. P. C. Rodrigues (Federal University of Piauí (UFPI), Brazil & Instituto de Telecomunicações, Portugal)

1
Diabetes is one of the fastest-growing chronic diseases in the world. Diabetic retinopathy, a complication of Diabetes that affects vision, and if not treated promptly, can lead to total blindness of the patient. This abnormality has no cure, but if discovered in its early stages, there is a high chance that the patient will not reach total blindness. Detection of retinal background exudates is essential for the early diagnosis of diabetic retinopathy. In this paper, we present a deep learning model with a Convolutional Neural Network to diagnose exudates' presence or absence. The best results are about 99.52% sensitivity, 100% specificity, and about 99.76% accuracy for 1,608 images. Thus, the authors believe the proposed method can integrate a clinical system.

Automatic Segmentation of Melanoma Skin Cancer Using Deep Learning

Rafael Luz Araújo (Federal University of Piaui, Brazil); Ricardo A. L. Rabelo (Federal University of Piaui (UFPI), Brazil); Joel J. P. C. Rodrigues (Federal University of Piauí (UFPI), Brazil & Instituto de Telecomunicações, Portugal); Romuere R. V. e Silva (Universidade Federal do Piauí, Brazil)

2
Segmentation is a crucial step to obtain success for classifying medical images. However, it is a highly complex task due to the abnormal shapes and the presence of other artifacts. In this study, a melanoma segmentation approach based on deep learning is proposed. In conjunction with post-processing techniques, the proposed modified U-net network has proven to be highly effective in lesions segmentation. The experiments were performed in two public datasets (PH2 and DermIS) and reached an average Dice coefficient of 0.933 in the PH2 dataset and Dice = 0.872 in the DermIS dataset. Considering the high-performance methodologies available in the literature, the proposed solution is very promising, surpassing other methods with very promising results.

A Deep Learning Based Ternary Task Classification System Using Gramian Angular Summation Field in fNIRS Neuroimaging Data

Sajila D Wickramaratne and Md Shaad Mahmud (University of New Hampshire, USA)

1
Functional near-infrared spectroscopy (fNIRS) is a non-invasive, economical method used to study its blood flow pattern. These patterns can be used to classify tasks a subject is performing. Currently, most of the classification systems use simple machine learning solutions for the classification of tasks. These conventional machine learning methods, which are easier to implement and interpret, usually suffer from low accuracy and undergo a complex preprocessing phase before network training. The proposed method converts the raw fNIRS time series data into an image using Gramian Angular Summation Field. A Deep Convolutional Neural Network (CNN) based architecture is then used for task classification, including mental arithmetic, motor imagery, and idle state. Further, this method can eliminate the feature selection stage, which affects the traditional classifiers' performance. This system obtained 87.14% average classification accuracy higher than any other method for the dataset.

Session S4-R2

E-Health Services and Applications

Conference
10:30 AM — 12:30 PM HKT
Local
Mar 1 Mon, 9:30 PM — 11:30 PM EST

Generating a Visual-Inertial Odometry Dataset based on a Helmet Prototype for Recognizing Human Activities

Kazi Md Shahiduzzaman (Huazhong University of Science and Technology & Jatiya Kabi Kazi Nazrul Islam University, China); Xiaojun Hei and Wenqing Cheng (Huazhong University of Science and Technology, China)

1
Human activity recognition (HAR) is an important area for elderly care. Because with an effective HAR clinical management authorities can monitor movement, abnormality, human behavior, chronicle diseases, and suddenly fall remotely. HAR may also reduce the workload of a caregiver. Our research mainly focuses on HAR for sudden fall detection and prediction. Usually, raw signals or features extracted from raw signals are used in HAR developmental works, which can increase false alarm rates (FAR). Besides, it is hard to differentiate various human activities through the illustration of this time-series signal. If these activities can be patterned in regular shape and can be expressed with a simple mathematical equation, then the recognition algorithm can not only detect daily activities but also predict them. Therefore, we will present a new and much effective technical way by using visual-inertial odometry (VIO) for human activity recognition in this paper. We consider walking, running and jumping activities to show our claims. From the results, we can see that considered human activities are easy to differentiate. 'Goodness of fit' of these activities will show how we could model mathematically them.

Recommender System for Postpartum Depression Monitoring based on Sentiment Analysis

Marcílio Carneiro and Mário W. L. Moreira (Federal Institute of Education, Science, and Technology of Ceará, Brazil); Silas Santiago Lopes Pereira (Federal Institute of Ceará, Brazil); Erica Gallindo (Federal Institute of Education, Science, and Technology of Ceará, Brazil); Joel J. P. C. Rodrigues (Federal University of Piauí (UFPI), Brazil & Instituto de Telecomunicações, Portugal)

1
Emotions influence all aspects of human behavior. All of these aspects shape people's lives, directly impacting their ways of life. Some diseases are directly linked to emotions. Among them, depression is one of the diseases with the greatest impact on society. Hence, faced with this problem, the objective of this study is to present a context-aware solution based on text mining for gestational depression prevention. This system uses text mining to analyze documents filled from pregnant women in order to identify their feelings through natural language processing techniques and probabilistic algorithms. As a case study, the analyzed texts were obtained from forms answered by pregnant women. The model performance is evaluated using metrics associated with the confusion matrix. The results show that the proposed model has achieved a reliable performance in all metrics, mainly when classifying new cases. Thus, the results obtained by the model can be used as support to health professionals in monitoring high-risk pregnancies.

Automatic Segmentation of Lung Nodules in CT Images Using Deep Learning

Acucena Soares and Thiago Lima (Federal University of Piauí, Brazil); Ricardo A. L. Rabelo (Federal University of Piaui (UFPI), Brazil); Joel J. P. C. Rodrigues (Federal University of Piauí (UFPI), Brazil & Instituto de Telecomunicações, Portugal); Flávio Araújo (Federal University of Piauí, Brazil)

1
Lung cancer is one of the leading death causes by cancer worldwide. Early diagnosis increases the patient's cure chances. This diagnosis is made by computed tomography, an imaging exam that provides accurate information about the nodule. However, it depends on many external factors, from equipment quality to the fatigue of expert who analyzes. Image processing techniques might be great allies in early nodule detection, once it has no human limitations. This study presents an evaluation of two deep learning approaches, 3D U-Net and 3D V-Net, with different configurations of architectures, parameters, and data augmentation distribution applied to pulmonary nodules segmentation. The best results obtained mean an IoU of 0.74 and 0.99 for 3D U-Net and 3D V-Net, respectively. The second network obtained the best results because it is a much more robust network than the 3D U-Net, since it is a network developed for volumetric data processing.

Automatic Identification of Metastasis in Histopathological Images Using Deep Learning

Daniel S Luz (Universidade Federal do Piauí (UFPI) & Instituto Federal do Piauí (IFPI), Brazil); Renesio Costa (Universidade Federal do Piauí (UFPI), Brazil); Ricardo A. L. Rabelo (Federal University of Piaui (UFPI), Brazil); Joel J. P. C. Rodrigues (Federal University of Piauí (UFPI), Brazil & Instituto de Telecomunicações, Portugal); Flávio Araújo (Federal University of Piauí, Brazil)

2
Metastatic tumor is one that spreads from its place of origin to other parts of the body. A tumor formed by metastatic cancer cells is called a metastatic tumor or metastasis. The early identification of these tumors is essential to increase the chances of success in treating the disease. However, for this identification it is necessary to analyze extensive tissues of the affected organs, which is a tiring and error-prone task. In this paper, it is present three deep learning strategies for automatic identification of metastasis in histopathological images. For the development and evaluation of these strategies it was used the PCam database, which is composed of 327,680 color images extracted from histopathological exams of sections of lymph nodes. The obtained results using the fine tuning technique are promising, showing that deep learning models can be used for metastasis identification.

Characterising Edge-Cloud Data Transmission for Patient-centric Healthcare Systems

Zheng Li and Francisco Millar-Bilbao (University of Concepción, Chile)

1
Benefiting from the modern information and communication technologies, the healthcare provisioning is actively evolving along a trend toward patient centricity, and the de facto solution seems to be technological collaboration and cooperation within a three-tier architecture. Given the largely distributed tiers and the components, a hot research focus is on minimizing the data transmission latency to improve the quality of healthcare services, especially in the time-critical situations. It has been identified that a typical performance bottleneck is the communication between the last two tiers that are generally represented by the edge and the cloud nowadays. Thus, optimising the edge-cloud data traffic becomes valuable and crucial to addressing the performance bottleneck, while characterising the edge-cloud data transmission plays a prerequisite role in the optimisation efforts. Unlike the existing studies that mainly emphasise the intuitive features (e.g., the overall big data volume), our work argues and reveals the importance of identifying the practical characteristics of edge-cloud data transmission at runtime, w.r.t. different workload regimes and heterogeneous edge nodes.

Texture Maps as Input in 3D CNNs Applied to Classify Nodules in CT Images

Helio Junior (Universidade Federal do Piauí (UFPI), Brazil); Flávio Araújo (Federal University of Piauí, Brazil); Ricardo A. L. Rabelo (Federal University of Piaui (UFPI), Brazil); Joel J. P. C. Rodrigues (Federal University of Piauí (UFPI), Brazil & Instituto de Telecomunicações, Portugal); Romuere R. V. e Silva (Universidade Federal do Piauí, Brazil)

1
Lung cancer is the leading cause of cancer-related death worldwide. Early diagnosis of pulmonary nodules on chest CT scans provides a chance to design an effective treatment. The focus of this study is the classification problem of benign and malignant pulmonary nodules in CT images. Thus, it is proposed to apply texture maps directly to the 3D nodules as a previous of the feature extraction process. For this, the local binary patterns (LBP), with branches, such as using neighbors with borders (LBP-6), average dimensions (LBP-M), and a 3×3×3 neighborhood (LBP-3×), to highlight the nodule texture. Convolutional Neural Networks, such as DenseNet, ResNet, and LeNet, were used as attribute extractors using the 3D texture maps computed. Then, those deep features are used as input to train a Random Forest classifier. In the experiments, it is used LIDC-IDRI image database. The LIDC-IDRI database was used with two segmentation process, one made by radiologists, present in the base itself (B1), and one performed automatically by a third party (B2). In B1, the best result was the original nodules' attributes extracted with the DenseNet architecture reaching an accuracy of 0.8371, a specificity of 0.9130, sensitivity of 0.7328, and Kappa of 0.6591. In B2, the best result was a combination of attributes of the original nodule combined with the extracted LBP-6 with LeNet architecture that reached an accuracy of 0.9037, a specificity of 0.8453, sensitivity 0.9266, and Kappa of 0.7641. In conclusion, it is possible to improve the classification accuracy by including a texture map computation as part of the process.

Session S4-R3

Communications and Networking

Conference
10:30 AM — 12:30 PM HKT
Local
Mar 1 Mon, 9:30 PM — 11:30 PM EST

Segmented Encryption: A Quality and Safety Supervisory Model for Herbal Medicine Based on Blockchain Technology

Jiameng Liu, Shaoliang Peng, Jiawei Luo, Zhuo Tang and Hao Liu (Hunan University, China)

1
The quality of herbal medicine has an important impact on human health. In this paper, we proposed a blockchain-based herbal quality and safety supervisory model for the current frequent herbal counterfeiting phenomenon. We manage the production, processing, and trading processes of herbal medicines by exploiting the blockchain's immutable and traceable properties. We proposed a segmented encryption method for information to encrypt the private information of enterprises. We use shared cloud storage to reduce waste of local storage space, and we proposed a verifiable random chain cutting mechanism based on the verifiable random function to handle the redundant blocks of the chain. Our article addressed the problem of herbal source falsification. Automated recording of key factors such as soil and temperature that affect the quality of herbal medicines is done from the seedling stage. Our herbal quality and safety supervisory model used blockchain technology to increase control over the production and distribution of herbal products, reduce herbal counterfeiting, and improve the efficiency of the system.

Efficiently recognition of vaginal micro-ecological environment based on Convolutional Neural Network

Shaoliang Peng, Hao Huang, Minxia Cheng and Yaning Yang (Hunan University, China); Fei Li (Chinese Academy of Sciences, China)

1
Vaginal diseases caused by vaginal micro-ecological abnormalities mainly include Vulvovaginal Candidiasis (VVC), Aerobic Vaginitis (AV), and Bacterial Vaginosis (BV). Severe cases can lead to poor pregnancy outcomes and infertility. AI-based technologies are being deployed with an expectation to relieve doctors of routine, tedious work when implemented correctly in daily microscopy of vaginal micro-ecological abnormalities. In this paper, we built a clinical image dataset of the Gram stain of the vaginal discharge. By comparing the performance of state of art convolutional neural network models, we found the fine-tuning Inception ResNet V2 shows the best classification performance for vaginal diseases. It achieves 96%, 94%, 86% AUC in VVC, AV, BV classification respectively. The result shows that compared with human visual inspection, the method based on deep learning greatly improves the screening sensitivity. Besides, we found that transfer learning can reduce the required manual labeling by roughly 73% (about more than one thousand samples). But for BV, which is difficult to diagnose for both humans and AI. Unlike AV and VVC, it requires more labeled data and is insensitive to the transfer fine-tuning.

A Novel Gait Prediction Method for Soft Exosuit Base on Limit Cycle and Neural Network

Lingxing Chen (Shenzhen Institutes of Adanced Technology, Chinese Academy of Sciences, China); Chen Chunjie (Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences, China); Tao Fang (Shenzhen Institutes of Adanced Technology, Chinese Academy of Sciences, China); Zhang Yu (Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences, China); Youfu Liu (Shaanxi University of Science & Technology & Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences, China)

1
The important purpose of soft exosuit is to decrease the energy consumption of users by providing assistance. If there is something wrong with the judgment of the human gait during the assisting process, the assisting effect will be poor and even people's walking will be affected. A novel method is designed to predict human gait information in the paper, which utilizes the curve shape and mathematical characteristics of the Rayleigh oscillator equation in limit cycle to fit gait information. Only 1/4 gait cycle data is needed to input into the trained neural network to output an Rayleigh oscillator equation that can better predict the remaining gait cycle information. Data of four subjects on different terrains which include flat ground and upstairs are collected. Experiment results showed that the trained ANN model costs about 0.006s in CPU, it has a prediction speed similar to traditional prediction methods. Simultaneously the Rayleigh oscillator equation has good performance in predicting gait information, and it can show a higher stability and accuracy compared with traditional gait prediction methods.

A Spatial Attention based Convolutional Neural Network for Gesture recognition with HD-sEMG signals

Sirong Hao, Ruxin Wang, Yishan Wang, Ye Li (Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences Shenzhen, China)

1
Recently, surface electromyogram (sEMG) has a trend with an increasing number of electrodes to compose a 2-dimension (D) electrode array, which is called high density sEMG (HD-sEMG). However, gesture recognition algorithm with HD-sEMG is still a challenge especially in real time recognition application. This paper researched several spatial attention modules and embedded them to the input layer of neural net-work. In this way, we can re-weight the input channel to get a better accuracy, robustness and interpretability. By utilizing the Group Convolution Neural Network (CNN), the gesture classification accuracy is improved by 4.44% and 2.71% in CapgMyo and CSL-HDEMG dataset respectively. This method is so efficient that it achieves only with 128 parameters, barely increasing the computational overhead. Meanwhile, we compared the performance in 1-D, 2-D and 3-D CNN, and found that our 1-D group CNN has great advantages in total computational overhead without the loss of accuracy. It provides a practical solution for real time gesture recognition application.

Predicting functional elements in non-coding regions based on deep learning

Yunhao Liu and Shaoliang Peng (Hunan University, China); Wenjie Shu (Beijing Institute of Radiation Medicine, China); Bin Jiang (Hunan University of China, China); Chao Yang and Kun Xie (Hunan University, China)

1
Accurate recognition and annotation of the important functional elements in the genome is an important prerequisite to understand the coding mode of complex regulatory networks in the one-dimensional genome .Despite rapid advances in sequencing and recognition technologies, accurately calling non-coding variant effects from large-scale sequence reads remains challenging.Here we present a deep neural network-based algorithmic framework, DeepMSA, which directly learns a regulatory sequence code from large-scale chromatin-profiling data,enabling to evaluate chromatin effects caused by SNP(single nucleotide polymorphism).

Session S5-R1

Medical, Biomedical, and Health Informatics

Conference
2:00 PM — 3:30 PM HKT
Local
Mar 2 Tue, 1:00 AM — 2:30 AM EST

A Novel Soft Exosuit Based on Biomechanical Analysis for Assisting Lower Extremity

Youfu Liu (Shaanxi University of Science & Technology & Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences, China); Chen Chunjie (Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences, China); Jun Lu (Shaanxi University of Science & Technology, China); Liu Yida, Zhang Yu and Zhuo Wang (Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences, China)

1
In the past two decades, with growing focus of lower limb exoskeleton, large variety of rigid exoskeletons were designed for medical rehabilitation and the other purposes. Compared with the rigid exoskeletons, which added extra inertial to low limb and restricted wearers' movement, the soft wearable lower extremity exoskeleton minimized the impact of these factors on the human body locomotion. In this paper, a novel design of exosuit that provided assistive force for both hip extension and flexion through the variation in hip joint dynamics during strides was proposed. Based on the change of hip moment, an assistance strategy and assistance force curve were came up with. PD type iterative learning control (ILC) method was introduced to reduce the error caused by wearing position and biological characteristics to improve assistance performance. To evaluate assistance performance, the metabolic cost of four subjects wearing exosuit and walking on treadmill at 5km/h in the situations of that without assistance, assisting both hip extension and hip flexion, and assisting hip extension respectively was measured. Compared with assisting hip extension only and wearing exosuit with no assistance, results indicated that the decrease in average net metabolic cost of assisting both hip extension and flexion was 0.445W/kg and 1.027W/kg respectively, corresponding to the average net metabolic cost rate decrease was 7.45\% and 15.67\%.

A Method for Recognition of Dynamic Hand Gestures Based on Wrist Tendon Sounds

Bailin He (Shenzhen Institutes of Advanced Technology,University of Chinese Academy of Sciences, China); Can Wang (Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, China); Hao Wang (Samsung Research Institute China-Shenzhen (SRC-SZ), Samsung Electronics,China); Mengyao Li (Shenzhen Institutes of Advanced Technology, University of Chinese Academy of Sciences, China); Shengcai Duan (Shenzhen Institutes of Advanced Technology,University of Chinese Academy of Sciences, China)

1
Hand gesture recognition are widely applied in game industry and rehabilitation engineering. Some recognition methods based on physiological data are attracting attention. In this paper, a novel method for dynamic hand gesture recognition based on sounds of wrist tendon is proposed. Bone conduction sensors which automatically filter the ambient noise is utilized to collect the wrist tendon sounds of three types of dynamic hand gestures. Continuous Wavelet Transform(CWT) is utilized extract the features of sounds and Support Vector Machine(SVM) is utilized to train the classifier model. Five healthy subjects participated the experiment and the average test accuracy of three classes achieves 92.33%. The proposed method provides an effective perspective for hand gesture recognition.

Multiple Time Scale Motion Images for Action Recognition

Qin Cheng and Ziliang Ren (Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, China); Jianming Liu (GuiLin University of Electronic Technology, China); Jun Cheng (Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, China)

1
This paper proposes a simple and effective approach for RGB-based action recognition using multiple streams Convolutional Neural Networks (ConvNets). We utilize a new method to represent temporal structure of RGB videos, named Motion Image (MI), which is constructed from the difference between frames with a certain time scale. Considering the different duration of different actions, multiple time scale sampling MIs can obtain more temporal information. Furthermore, we adopt multiple streams ConvNets, including MIs and RGB streams, to learn spatial-temporal features for action recognition. Our approach has been evaluated on UCF-101 and HMDB-51, and the experimental results demonstrate the effectiveness and significantly improve action recognition rate at a small computational cost.

Phase-Sensitive Model for Temporal Action Proposal Generation

Shijie Sun (Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, China); Qingsong Zhao (School of Electronics and Information Engineering, Tongji University, China); Ziliang Ren, Lei Wang and Jun Cheng (Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, China)

1
Temporal action proposal generation is an important and challenging task, aiming to localize the position where an action or event may occur in an untrimmed video. In this paper, we propose an efficient and end-to-end framework to generate temporal action proposals, named Phase-Sensitive Model (PSM), which fully understands all phases of temporal information. In particular, the PSM consists two modules: Boundary Phase Classification (BPC) and Action Phase Classification (APC). The BPC aims to provide two temporal boundary phase confidence maps by rich local information, while the APC is designed to generate an action phase confidence map by global features. Moreover, we introduce a new method boundary probability calculation to get the final score. Our experiments on ActivityNet-1.3 show a significant improvement with remarkable efficiency and generalizability.

Session S5-R2

E-Health Services and Applications

Conference
2:00 PM — 3:30 PM HKT
Local
Mar 2 Tue, 1:00 AM — 2:30 AM EST

Prognosticating Colorectal Cancer Recurrence using Machine Learning Techniques

Danliang Ho (National University of Singapore, Singapore); Dawn Qingqing Chong, Brenda Tay and Iain BeeHuat Tan (National Cancer Center Singapore, Singapore); Mehul Motani (National University of Singapore, Singapore)

0
Colorectal cancer is among the top three most commonly occurring cancers worldwide, and around 30-40% of patients treated by curative intent surgery will experience cancer recurrence. Proactive prognostication would enable clinicians to better plan treatment modality and intensity, and follow-up frequency to reduce recurrence. Here, we study the application of machine learning models to predict cancer recurrence in a cohort of 904 post-resection colorectal cancer patients. We employ heterogeneous structured and temporal clinical features including demographic and diagnostic information, tumour stage and location details, biochemistry and molecular typing results, as well as surgical details and treatment parameters. We characterize the performance of multiple machine learning classifiers including logistic regression, support vector machine, gradient boosting and multi-layer perceptron on structured data. Our best model achieved a sensitivity of 80.7% and a specificity of 88.2%. This is comparable to and even exceeding the performance of carcinoembryonic antigen (CEA), a tumour marker commonly used in the clinic for colorectal cancer monitoring. We also demonstrate feasibility for accurate forecasting of recurrence up to 4 months in advance, as well as the possibility of predicting recurrence as early as 6 months post-surgery. Our results have positive implications for better management of colorectal cancer patients in the post-resection setting.

Usage Prediction and Effectiveness Verification of App Restriction Function for Smartphone Addiction

Katsuki Yasudomi (Waseda University, Japan); Toshitaka Hamamura, Masaru Honjo and Akio Yoneyama (KDDI Research, Inc., Japan); Masato Uchida (Waseda University, Japan)

0
In recent years, there has been a growing problem of smartphone addiction. As the excessive use of smartphones has negatively impacted our daily lives, many apps for reducing smartphone addiction have been developed around the world. In this study, we focus on the app restriction function, which is one of the key features of digital medicines for smartphone addiction, and analyze the usage of the function and verify its effectiveness. The results showed significant differences in both psychological and behavioral aspects between those who used the app restriction function and those who did not. Specifically, we found that the app restriction function was more likely to be used by those who were more aware of their smartphone addiction. We also found that the app restriction function was effective in lessening smartphone usage time, especially when the smartphone addiction is relatively moderate.

Blockchain-empowered Contact Tracing for COVID-19 Using Crypto-spatiotemporal Information

Zheng Wen and Keping Yu (Waseda University, Japan); Xin Qi (Global Information and Telecommunication Institute, Waseda University, Japan); Toshio Sato (Waseda University, Japan); Yutaka Katsuyama (Global Information and Telecommunication Institute Waseda University, Japan); Takuro Sato and Wataru Kameyama (Waseda University, Japan); Fumiyuki Kato, Yang Cao and Masatoshi Yoshikawa (Kyoto University, Japan); Min Luo and Jun Hashimoto (EY Japan, Japan)

0
The pandemic of the COVID-19 [1] has re-awakened people that viruses are still the greatest threat to human society. Quarantining the patients and tracking close contacts has been used for hundreds of years in the battle between humans and the plague, which are still useful today. In the information society, we can employ information communications technology (ICT) to suppress the spread of epidemics and lower the epidemic curve. By using spatiotemporal information, we can trace the trajectories of patients and their close contacts. However, spatiotemporal information also involves personal privacy, and it has become a topic of concern about whether people's privacy should be sacrificed for epidemic control. In this paper, we propose a close contact tracing solution based on crypto-spatiotemporal information (CSI). First, the solution encrypts spatiotemporal information to protect personal privacy. Then, it uses a blockchain platform to realize the proof of CSI and uses Intel SGX [2] based trusted execution environment [3] to perform close contact judgment. Finally, it can trace close contacts while protecting personal privacy. The evaluation results indicate that the advantages and efficiency of the proposed scheme are significant.

Optimizing Broad Learning System Hyper-parameters through Particle Swarm Optimization for Predicting COVID-19 in 184 Countries

Choujun Zhan (South China Normal University, China); Wu Z Dong (University of Nanfang, China); Quansi Wen and Ying Gao (South China University of Technology, China); Haijun Zhang (Shenzhen Graduate School, Harbin Institute of Technology, China)

0
The Coronavirus Disease 2019 (COVID-19) began to outbreak since December 2019 and widely spread to more than 184 countries. How to accurately predict the spread of COVID-19 in a country or a city is one of the essential issues for controlling the pandemic. This article establishes a general model that can predict the trend of COVID-19 in a country based on historical COVID-19 data in 184 countries. First, Savitzky-Golay (S-G) filter is utilized to detect multiple waves of COVID-19 in a region. Then, a PSOSIR (particle swarm optimization susceptible infected recovery) model is provided for data augmentation. Finally, a novel PSO-BLS (particle swarm optimization broad learning system) is proposed to predict the pandemic trend. Experimental results show that compared with the deep learning models (ANN, CNN, LSTM, and GRU), PSO-BLS has higher accuracy and stability in predicting the number of COVID-19 active infection and removed cases. The R2 of the predicted number of active infection and removed cases can reach 0.998 and 0.995, respectively.

Session S5-R3

Communications and Networking

Conference
2:00 PM — 3:30 PM HKT
Local
Mar 2 Tue, 1:00 AM — 2:30 AM EST

Remote Monitoring Framework for Elderly Care Home Centers in UAE

Raafat Aburukba and Assim Sagahyroon (American University of Sharjah, United Arab Emirates); Loay Kamel, Abdulla Al-Shamsi, Hussain Surti and Eisa Sajwani (AUS, United Arab Emirates)

0
In an effort to improve the quality of life of its residents, elderly care centers in the UAE are exploring the use of remote monitoring techniques as a viable and cost effective option. Currently, elderlies within the elderly home care center might be in different places unmonitored. Moreover, at night, the number of caregivers or attending nurses is usually limited where at times making the continuous monitoring of all the residents a very difficult if not impossible task. Elderlies would be highly susceptible to emergencies such as wandering off the care home center, falling and possibly abnormal changes in their vital signs. Thus there is critical need for continuous and reliable monitoring of residents in these centers preferably with real-time built-in alerting mechanism in case of emergencies.. This work proposes a framework that enables the data collection from wearable sensor devices and customizes the elderly monitoring based on defined rules for each individual. Moreover, the work provides a dashboard for caregivers to visually monitor all elderlies' conditions and activities in the care center. This project is carried out in cooperation with one of the largest center in Dubai, UAE. It is validated by a prototype implementation.

Energy-Efficient Networks Selection Based Deep Reinforcement Learning for Heterogeneous Health Systems

Zina Chkirbene (Qatar University & Electrical Engineering, Qatar); Amr Mohamed (Qatar University, Qatar); Aiman Erbad (Hamad Bin Khalifa University, Qatar); Mohsen Guizani (Qatar University, Qatar)

1
Smart health systems improve the existing health services by integrating information and technology into health and medical practices. However, smart healthcare systems are facing major challenges including limited network resources, energy allocation, and latency. In this paper, we leverage the dense heterogeneous network (HetNet) architecture over 5G network to enhance network capacity and provide seamless connectivity for smart health systems. The network selection and energy allocation in HetNets are important factors in this regard due to their significant impact on system performance. Inspired by the success of Deep Reinforcement Learning (DRL) in solving complicated control problems, we present a novel DRL model for energy-efficient network selection in heterogeneous health systems. The proposed model selects the set of networks to be used for data transmission with adaptive compression at the edge with an optimal energy allocation policy for all the network participants. Our experimental results show that the proposed DRL model has a good performance compared to the existing state of art techniques while meeting different users' demands in highly dynamic environments.

A drug information embedding method based on graph convolution neural network

Xiaoyi Feng and Shaoliang Peng (Hunan University, China); Fei Li (Computer Network Information Center, Chinese Academy of Sciences, China); Ying Xu and Xiangxiang Zeng (Hunan University, China); Donqing Wei (Shanghai Jiao Tong University, China); Yunhao Liu (Hunan University, China)

0
New drug development is an extremely time-consuming and high-risk process. It has been widely valued by the biomedical industry to fully explore the new uses of existing drugs and repositioning. How to find drug disease with potential therapeutic relationship from a large number of unproven relationship pairs is the research focus of drug reorientation. With the help of machine learning model, we can improve the enrichment degree of potential drug disease relationship pairs, and reduce the false positive rate of prediction. In the past few years, a series of graph based convolutional network models have been developed to calculate the information latent feature representation of nodes and links. Researchers at home and abroad have done a lot of research on network embedding technology based on biomedical data, and have achieved a series of important research results. Among them, the research methods used can be divided into two categories: one is the traditional machine learning algorithm based on artificial feature extraction, the other is the method based on deep learning. For example, some experts proposed a new graph convolution network (GCN) with parts of existing models, Deep Drug Repositioning and DTINet based on node characteristics and their connections, which can be used for node classification. Aiming at the problem of imbalance of drug information data samples, the invention provides a drug relocation method based on deep learning multi-source heterogeneous network. In order to avoid the limitations of traditional feature extraction methods, such as highly dependent on the experience and knowledge of medical staff, strong subjectivity, consuming a lot of time and energy to complete, and extracting high-quality features with distinguishing features often exists In this paper, with the help of graph convolution encoder model and variational auto encoder neural network, we can automatically learn the characteristics of multi-source and heterogeneous drug low-dimensional network, and complete the drug relocation of drug disease association prediction.

GeoAI-based Epidemic Control with Geo-Social Data Sharing on Blockchain

Shaoliang Peng (National University of Defense Technology & State Key Laboratory of High Performance Computing, China); Liang Bai (Hunan University, China); Li Xiong (The Second Xiangya Hospital of Central South University, China); Qiang Qu (Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, China); Xiaolan Xie (Guilin University of Electronic Technology, China); Shulin Wang (Hunan University, China)

0
Epidemics especially those caused by major contagious diseases have entailed huge losses in human history. The fights have thus never stopped to prevent pandemics. Due to its acute outbreak, is generally susceptible to the population regardless of ages, so strict quarantine of the infections becomes the most effective means for the epidemic control, which has been proved in the prevention of other contagious diseases such as SARS and H1N1. The key strategy widely used to find infected and suspected patients is still the epidemiological tracking of confirmed cases. However, this may fail to identify infections especially when patients do not show any symptoms. Therefore, the approach to rapid, effective, and simple infection identification is essential to prevent the spread of a contagious disease. This paper proposes to leverage a social apps and Geospatial artificial intelligence (GeoAI) with Blockchain to effectively identify infections with privacy concern. Since people widely use social apps, a large scale of social data with geospatial information could be easily collected and kept on Blockchain with privacy preservation, which thus provides a framework of decentralized, tamper-proof, and privacy-preserved information sharing. With the support of GeoAI, which analyzes the spatial distribution of diseases from the shared data, we could study the influence factors based on spatial propagation of contagious diseases for infection identification. Since WeChat is widely used in China, we take COVID-19 as an example to use the experiments on real-life datasets demonstrate the effectiveness of our method, and provide insight into epidemic control in terms of geo-social data sharing.

Session S6-R1

Medical, Biomedical, and Health Informatics

Conference
4:00 PM — 5:30 PM HKT
Local
Mar 2 Tue, 3:00 AM — 4:30 AM EST

Risk Analysis Based Security Compliance Assessment and Management for Sensitive Health Data Environment

Umar Ozeer (Euris Health Cloud, France); Badara Pouye (Euris Health Cloud & Nuagérence SAS, France)

0
The digitisation of personal health information (PHI) through electronic health record (EHR) has become widespread due to their efficiency in terms of cost, storage, processing, and the subsequent quality of delivering patient care. However, security concerns remain one of its major setback. In order to handle EHR, institutions need to comply with their local government security regulations. These regulations control to which extent health data can be processed, transmitted, and stored as well as define how misuses are addressed. This paper proposes ϕcomp, a solution for monitoring, assessing, and evaluating the compliance of health applications with respect to defined security regulations. ϕcomp is able to assess the level of security risk of an application at runtime and automatically perform the required mitigation actions to recover a compliant environment. ϕcomp was implemented in an industrial context and evaluated on a medical appointment application. The results of the experiments showed that it manages the security compliance of third party applications with low performance overhead and attenuate unacceptable levels of risk to restore compliance.

Protein Structure Prediction Based on Multi-Level Information Fusion

Ethan Chen (Hunan University, China); Shaoliang Peng (National University of Defense Technology & State Key Laboratory of High Performance Computing, China); Xiaoqi Wang (Hunan University, USA); Sheng Xiao, Jiawei Luo and Zhuo Tang (Hunan University, China)

0
The magnitude of the conformational space and the inaccuracy of the energy model have challenged the prediction of protein structure from scratch for decades. In succession, many researches put forward various optimization methods to find the optimal solution and achieved good results. In the partial sampling method based on the local structure characteristics of proteins, the commonly used features include residual interaction and secondary structure. In recent years, the prediction of residue contact has been used to predict the structure of supplementary proteins with remarkable results. In this paper, a conformation selection model based on multilevel information fusion with different local structure features MLI is designed, and conformation sampling is carried out based on the framework of genetic algorithm. From the evaluation results of typical test proteins, the experimental results are competitive with the most advanced methods.

Healthcare and data privacy requirements for e-health cloud: a qualitative analysis of clinician perspectives

Taridzo Chomutare and Kassaye Yitbarek Yigzaw (Norwegian Centre for E-health Research, Norway); Silvia Olabarriaga (University of Amsterdam, The Netherlands); Alexandra Makhlysheva (Norwegian Centre for E-health Research, Norway); Marcela Tuler de Oliveira (University of Amsterdam & KEEB/BMEP, The Netherlands); Line Silsand (Norwegian Centre for E-health Research, Norway); Dagmar Krefting (HTW Berlin - University of Applied Sciences, Germany); Thomas Penzel (Charité - Universitätsmedizin Berlin, Germany); Christiaan Hillen (Secura, The Netherlands); Johan Bellika (Norwegian Centre for E-health Research, Norway)

0
Cloud computing has many benefits relevant to the healthcare industry. Although the adoption of cloud services for healthcare systems is increasing, employment of cloud services raises many security and privacy concerns for patients and healthcare providers. We still lack a clear set of requirements consented by the different stakeholders; here in particular IT and healthcare professionals. In this study, we examine whether user perspectives on requirements for e-health on the cloud are consistent with best practice guidelines and regulatory requirements. This work contributes to the requirements engineering phase for a secure e-health cloud framework developed in a European project (ASCLEPIOS, https://www.asclepios-project.eu/). We used qualitative analysis, based on in-depth interviews, to describe and characterize clinicians' perspectives on the requirements of cloud services for healthcare data security and privacy. We examined whether these user perspectives were in harmony with the regulatory framework of the General Data Protection Regulation (GDPR), and best practice guidelines of a relevant standard, ISO 18308:2011. Ten clinicians were identified and interviewed at six healthcare organizations in Norway, the Netherlands and Germany. While user perspectives were largely consistent with both GDPR and ISO, some concerning differences in access control were noted between large and small healthcare institutions.

Modeling and Closed-loop Control of Ferromagnetic Nanoparticles Microrobots

Zhiming Hao (Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences); Tiantian Xu, Chenyang Huang and Zhengyu Lai (Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, China); Xinyu Wu (Shenzhen Institutes of Advanced Technology, Chinese Academy of Science, China)

0
Microrobots capable of performing minimally invasive surgery, targeted drug delivery, and manipulation of micro-objects have shown great potentials in multiple application areas. However, microrobots assembled by nanoparticles with stable configuration and good performance on closed-loop control are still to be further studied. In this paper, ferromagnetic nanoparticles are used as experimental materials, rather than paramagnetic nanoparticles that require complex synthesis processes, and the advantage is demonstrated, for example, high pattern stability. The locomotion velocity as a function of the magnetic field frequency is modeled, analyzed, and verified by experiment. Moreover, a path following experiment based on the arbitrary planar path following algorithm is performed. The nanoparticle microrobots are of great significance for biomedical applications. In the future works, experiments in bio-fluids, multi-modal locomotion, and targeted drug delivery tasks will be investigated.

Session S6-R2

Communications and Networking

Conference
4:00 PM — 5:30 PM HKT
Local
Mar 2 Tue, 3:00 AM — 4:30 AM EST

Link Fault Repair Algorithm of Wearable Wireless Sensor Networks based on Polygon Fermat Point

Lincong Zhang and Ce Zhang (Shenyang Ligong University, China); Kefeng Wei (Northeastern University, China); Qieshi Zhang (Shenzhen Institutes of Advanced Technology, Chinese Academy of Science, China)

0
With the development of wearable wireless sensor network technology, the technology has been applied in more and more fields, such as military, medical rescue, transportation and so on. Especially in the field of medical rescue, because of the advantages of wearable wireless sensor network, such as portability, fast and flexible networking, it has a very high application prospect. However, due to the particularity of medical rescue scene, the communication quality of wearable wireless sensor network is facing a huge challenge. When there is a fault, it needs to repair the fault link timely and accurately. In this paper, a link fault repair algorithm based on polygon Fermat point (LFRA) is proposed. Firstly, the algorithm can repair the fault link by inserting the relay node after the fault occurs. Secondly, the algorithm considers the energy consumption of the nodes involved in the repair process, eliminates the nodes that do not meet the energy requirements, and avoids multiple repairs due to lack of energy. The simulation results show that the proposed algorithm has the advantages of fewer inserted relay nodes and longer maximum communication time between nodes.

Unbalanced Data Set Processing Method for Colorectal Cancer Prediction in TCM Diagnosis

Wenbin Bi (Dalian Neusoft University of Information, China); Rong Ma (Lanzhou University, China)

0
Aiming at the imbalance in the data of colorectal cancer outpatients in the diagnosis of traditional Chinese medicine, an unbalanced data set processing framework (UDSPF) is proposed. The framework improves the generalization ability of unbalanced data sample prediction, strengthens the proportion of minority class samples with high correlation with related features, and eliminates some unrelated majority class samples, further improving the balance of the overall data samples. After completing the comparative simulation experiment with other common unbalanced data prediction methods, the framework was tested and verified on the real TCM outpatient medical record data set. The experimental results show that the unbalanced data set processing method based on UDSPF can be effectively applied to the prediction of colorectal cancer, which saves time for patients to carry out corresponding treatment as early as possible.

ISDB: An Effective Ciphertext Retrieval Method for Electronic Health Records Based on SDB

Weifeng Sun, Yiming Wang and Zerui Ding (Dalian University of Technology, China); Nan Li (Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, China)

0
The central component of E-Health is electronic health records, however, the number of electronic records that need to be stored is increasing day by day. The electronic records can be stored with the help of DBaaS in the Cloud. Electronic health records contain a lot of private information, so it is unreasonable to store the plaintext records in the cloud. There will be additional computing costs if using traditional encrypted storage to retrieve all the records in the cloud database and retrieve them after decryption. In order to improve the retrieval efficiency of electronic health records as much as possible while ensuring user privacy, ISDB expands on the model of the secure query model SDB, so that the improved model supports the data types of varchar, date and date time in SQL Encryption. In ISDB, the sting-like types of exact matching function is designed, so that the model can be more effectively used for encrypted storage of electronic health records and ciphertext retrieval. By analysis and experiments, the ISDB can meet the security ciphertext retrieval of commonly used data types in electronic health records, and it is a light model and is effective.

Made with in Toronto · Privacy Policy · © 2022 Duetone Corp.