Session S1-R1

Medical, Biomedical, and Health Informatics

Conference
11:00 AM — 12:30 PM HKT
Local
Feb 28 Sun, 7:00 PM — 8:30 PM PST

MABEL: An AI-Powered Mammographic Breast Lesion Diagnostic System

Zhicheng Yang, Zhenjie Cao, Yanbo Zhang and Peng Chang (PAII Inc., USA); Shibin WU, Lingyun Huang and Wei Xu (Ping An Technology, China); Mei Han (PAII Inc., USA); Jing Xiao (Ping An Insurance Company of China,Ltd., China); Mingxiang Wu (Shenzhen People’s Hospital, China); Jie Ma (Shenzhen People's Hospital, China)

4
Mammography plays an essential role in early detection of breast cancer. Interpreting mammography is a professional task that requires well-trained radiologists with long-time clinical experience. In this paper, we present MABEL, an artificial intelligence-powered system to assist doctors for breast cancer screening and diagnosis in mammograms, in order to reduce their workloads and accelerate the diagnostic process. Our system smoothly integrates our upgraded lesion identification models, provides a doctor-oriented annotation tool and web interface, and can communicate with Picture Archiving and Communication System (PACS) in our collaborative hospital. Our lesion identification performance is evaluated on both public and in-house datasets, in which mass detection has achieved state-of-the-art accuracy in the single-view manner. The overall high satisfaction from doctors of our system is also demonstrated.

HealthCam: Machine Learning Models on Mobile Devices for Unhealthy Packaged Food Detection and Classification

David Cantillo and Brandon Cervantes (Universidad del Norte, Colombia); Jairo Cardona (Norte University, Colombia)

2
We live in a world where most foods and beverages are unhealthy due to their high content of sugars, sodium, fat, among others [1]. Although many of these foods, especially packaged foods (snacks) for children have nutritional information, this is not very clear, and its interpretation is not easy for a normal person. For this reason, HealthCAM is developed. HealthCAM is a mobile application that uses augmented reality and machine learning techniques to detect the type of snack and indicate to the user through a visual warning on the phone's camera in real time the degree of sugar, sodium and fat that the snack has.

A novel smartphone App-based assessment of standing postural control: Demonstration of reliability and sensitivity to aging and task constraints

Junhong Zhou (Harvard Medical School, USA); Wanting Yu, Hao Zhu, On-Yee Lo, Natalia Gouskova, Thomas Travison, Lewis Lipsitz, Alvaro Pascual-Leone and Brad Manor (Hinda and Arthur Marcus Institute for Aging Research, Hebrew SeniorLife, USA)

1
Laboratory-based assessments of standing posture provide important insights into balance, mobility, and many other factors. We created a smartphone-based assessment of standing posture-completed with the phone placed in the user's pocket-for use in remote, non-laboratory settings. We here tested the reliability and sensitivity of App-derived metrics of postural sway to participant age and standing condition. Fifteen healthy younger and 15 older adults completed two laboratory visits separated by one week. On each visit, they followed multi-media instructions provided by the App to complete three 30-second trials each of standing with eyes open (EO), eyes closed (EC), and eyes open while performing a serial subtraction dual task (DT). Sway data were simultaneously collected with the App and a gold-standard force plate. Participants also completed App-based tests within their own homes on three separate days. To characterize sway, path length and root-mean-square (RMS) were derived from the acceleration and angular velocity signal acquired from the phone's internal motion sensor, and from the center of pressure (COP) signal acquired by the force plate. Across repeated trials conducted on the same day within the laboratory, App-derived path length and RMS of acceleration and angular velocity across conditions demonstrated moderate to excellent test-retest reliability (Intraclass Correlation Coefficients (ICCs)=0.69-0.97) for both the younger and older adults. Force plate-derived metrics demonstrated low to excellent test-retest reliability (ICCs=0.44-0.93). Compared to within-day laboratory testing, App-derived sway outcomes generally exhibited greater variability when tested across days within the home (ICCs=0.14-0.79). All sway metrics derived from the App were sensitive to age group (F>43.8, p<0.001) and task condition (F>31.8, p<0.001) in both laboratory and home settings. The smartphone App we created enabled reliable and sensitive assessment of standing postural sway within different task conditions, in both relatively healthy younger and older adults.

Operationally-Informed Hospital-Wide Discharge Prediction Using Machine Learning

Andrew Ward, Ariana J Mann and Jacqueline Vallon (Stanford University, USA); Gabriel Escobar (Kaiser Permanente, USA); Nicholas Bambos (Stanford University, USA); Alejandro Schuler (Kaiser Permanente, USA)

2
Accurate patient discharge time estimates are invaluable for hospital operations management. They are vital for efficient and effective scheduling of hospital resources including beds and staff. Unexpected discharges place strain on the patient families and care providers, in addition to causing hospital inefficiencies.Due to the increasing availability of electronic health record data, predictive models can be leveraged to not only offer clinical decision support, but also to optimize hospital operations.In this work, we incorporate clinical knowledge from operational leaders at Kaiser Permanente Northern California to design a predictive model for patient discharge using a novel dataset that contains hourly data from the electronic health records of 14 different Kaiser Permanente hospitals. We train and test several algorithms with varying complexity to predict patient-level discharges for the following day at operationally relevant times on the hospital-centric timescale.The highest AUC we achieve is 0.729 with a gradient boosted model, which significantly outperforms both the current estimates deployed in these 14 facilities and the baseline model without hourly data. A feature permutation importance assessment is performed and we conclude that the majority of the improvement is due to the inclusion of the detailed, hourly data.

Frameworks, Methodologies and Specification Tools for the Enterprise Architecture Application in Healthcare Systems: A Systematic Literature Review

Silvano Júnior, Francisco Silva and Gustavo Galisa (IFPB, Brazil); Francisco Petronio Alencar de Medeiros (Federal Institute of Paraiba & IFPB, Brazil); Heremita Lira (Federal Institute of Paraiba - IFPB, Brazil)

2
Although there are many works on the Enterprise Architecture (EA) application in healthcare systems, the literature lacks studies that provide a systematic approach to this topic specifically. This work presents a deep and broad Systematic Literature Review (SLR) to select studies demonstrating current EA practices in healthcare systems. The researchers established an SLR protocol returning 280 primary studies after the first step of the Data Selection and a consolidated inclusion of 46 articles after the second step. It was assessed the level of disagreement during the team's evaluations using Cohen's Kappa. This SLR revealed essential aspects of state-of-the-art EA application in healthcare systems, such as the most used frameworks, methodologies, and specification tools, and the criteria for their choice.

Session S1-R2

E-Health Services and Applications

Conference
11:00 AM — 12:30 PM HKT
Local
Feb 28 Sun, 7:00 PM — 8:30 PM PST

Flow-Based Situation-Aware Approach for eHealth Data Processing

Jordano Ribeiro Celestrini and Alessandro Baldi (Federal University of Espirito Santo, Brazil); Celso Alberto Saibel Santos (Federal University of Espírito Santo, Brazil); Rodrigo Varejão Andreão (Federal Institute of Espírito Santo, Brazil); José Gonçalves Pereira Filho (Federal University of Espírito Santo, Brazil)

1
A major challenge for remote patient monitoring (RPM) systems is processing data collected from a large number of healthcare devices and developing suitable algorithms and approaches to react accordingly to a wide spectrum of situations of interest, which must be properly detected. Flow-based programming, whose operation is based on state changes, is a promising approach to overcome this challenge, facilitating the personalization of data processing according to patients' health conditions. However, this approach is not suited to the detection of complex contextual situations, thereby hindering its adoption in RPM systems for data processing. In this regard, we propose an approach that combines situation awareness and flow-based programming to widen the capability of RPM systems to handle many-sided scenarios, which requires the monitoring of complex patient healthcare conditions. An evaluation of the proposed approach was conducted to demonstrate the flexibility of the solution for processing heterogeneous health information.

Intelligent Epidemiological Surveillance in the Brazilian Semiarid

Raimundo Valter Filho (IFCE, Brazil); José Neuman de Souza (Federal University of Ceará, Brazil); Antonio Oliveira (Federal Institute of Ceara, Brazil); Samuel Brasil Albuquerque (École Polytechnique Palaiseau, France); Luis Odorico Monteiro Andrade and Ivana Cristina de Barreto (FIOCRUZ, Brazil); Daniel de Andrade, Luzia Lucélia Saraiva Ribeiro and Flávio Cardoso (AVICENA, Brazil); William Vitorino (IFCE, Brazil); Francisco Gabriel da Silva (AVICENA, Brazil)

1
Right after the Chinese example in conducting COVID-19 epidemic originated in Wuhan, the readiness to detect and respond by health authorities to local (sometimes global) epidemics has become central lately. Within the idea of health 4:0, information about the individual is essential in supporting public community health policies. This paper presents a proposal for an epidemiological surveillance system applied to arboviruses. Data mining techniques and Machine Learning (ML) are used to design mathematical models for detecting epidemics enhanced by Aedes Aegypti (vector for dengue, chikungunaya, yellow fever and zica). A Prove of Concept (PoC) is presented for dengue epidemics detection, a common endemic disease in the semiarid region of Brazil.

DocPal: A Voice-based EHR Assistant for Health Practitioners

Vidisha Bhatt, Juan Li and Bikesh Maharjan (North Dakota State University, USA)

1
Electronic health record (EHR) systems have been widely adopted across healthcare organizations. While there are many benefits of using EHR such as improved accessibility and secure sharing of patient data, a shortcoming is that its manual data input is time-consuming and error-prone. Physicians spend as much as 49.2% of their office time on EHR. In this paper, we present the design, development, and evaluation of a voice-based assistant, DocPal, to assist healthcare practitioners to access and update EHR through their voice. User survey and experimental evaluation illustrate that DocPal has good usability, time efficiency, and accuracy. When applied in the healthcare industry, we expect it to reduce data entry time and provide better patient care.

Development and Evaluation of ADCareOnto - an Ontology for Personalized Home Care for Persons with Alzheimer's Disease

Juan Li, Rasha Hendawi, Vikram Pandey, Rafa Alenezi and Xin Wang (North Dakota State University, USA); Bo Xie (The University of Texas at Austin, USA); Cui Tao (University of Texas Health Science Center at Houston, USA)

1
Alzheimer's disease (AD) poses serious challenges for both patients and their family caregivers. In this paper we present the design, development, and evaluation of an ontology model, ADCareOnto, to assist family caregivers providing personalized care for persons living with AD. ADCareOnto includes top-level categories, concepts, and relations about informal care for persons with AD. To enable personalization in care, ADCareOnto also includes a comprehensive user profile modeling that includes various characteristics of both AD patients and caregivers. AD care thus can be tailored based on the user's unique concerns, preferences, and needs. We verified and validated the design of ADCareOnto and evaluated it using a real use case. The results support the quality of its content and techniques.

Machine Learning Based Autism Spectrum Disorder Detection from Videos

Chongruo Wu (University of California, Davis, USA); Sidrah Liaqat, Halil Helvaci and Sen-ching Samson Cheung (University of Kentucky, USA); Chen-Nee Chuah, Sally Ozonoff and Gregory Young (University of California, Davis, USA)

1
Early diagnosis of Autism Spectrum Disorder (ASD) is crucial for best outcomes to interventions. In this paper, we present a machine learning (ML) approach to ASD diagnosis based on identifying specific behaviors from videos of infants of ages 6 through 36 months. The behaviors of interest include directed gaze towards faces or objects of interest, positive affect, and vocalization. The dataset consists of 2000 videos of 3-minute duration with these behaviors manually coded by expert raters. Moreover, the dataset has statistical features including duration and frequency of the above mentioned behaviors in the video collection as well as independent ASD diagnosis by clinicians. We tackle the ML problem in a two-stage approach. Firstly, we develop deep learning models for automatic identification of clinically relevant behaviors exhibited by infants in a one-on-one interaction setting with parents or expert clinicians. We report baseline results of behavior classification using two methods: (1) image based model (2) facial behavior features based model. We achieve 70% accuracy for smile, 68% accuracy for look face, 67% for look object and 53% accuracy for vocalization. Secondly, we focus on ASD diagnosis prediction by applying a feature selection process to identify the most significant statistical behavioral features and a over and under sampling process to mitigate the class imbalance, followed by developing a baseline ML classifier to achieve an accuracy of 82% for ASD diagnosis.

Session S1-R3

Communications and Networking

Conference
11:00 AM — 12:30 PM HKT
Local
Feb 28 Sun, 7:00 PM — 8:30 PM PST

A Knowledge-Based Decision Support System for In Vitro Fertilization Treatment

Xizhe Wang (University of Massachusetts Lowell, USA); Ning Zhang (University of Massachusetts, Lowell, USA); Jia Wang, Jing Ni and Xinzi Sun (University of Massachusetts Lowell, USA); John Zhang and Zitao Liu (New Hope Fertility Center, USA); Cao Yu (The University of Massachusetts Lowell, USA); Benyuan Liu (University of Massachusetts Lowell, USA)

1
In Vitro Fertilization (IVF) is the most widely used Assisted Reproductive Technology (ART). IVF usually involves controlled ovarian stimulation, oocyte retrieval, fertilization in the laboratory with subsequent embryo transfer. The first two steps correspond with females' follicular phase and ovulation in their menstrual cycle. Therefore, we refer to it as the treatment cycle in our paper. The treatment cycle is crucial because the stimulation medications in IVF treatment are applied directly on patients. In order to optimize the stimulation effects and lower the side effects of the stimulation medications, prompt treatment adjustments are in need. In addition, the quality and quantity of the retrieved oocytes have a significant effect on the outcome of the following procedures. To improve the IVF success rate, we propose a knowledge-based decision support system that can provide medical advice on the treatment protocol and medication adjustment for each patient visit during IVF treatment cycle. Our system is efficient in data processing and light-weighted which can be easily embedded into electronic medical record systems. Moreover, an oocyte retrieval oriented evaluation demonstrates that our system performs well in terms of accuracy of advice for the protocols and medications.

TinyDL: Edge Computing and Deep Learning Based Real-time Hand Gesture Recognition Using Wearable Sensor

Brian Coffen and Md Shaad Mahmud (University of New Hampshire, USA)

1
Offloading data analysis to edge devices by decentralizing processing can be used to decrease bandwidth requirements, latency, and can decrease the total transmission time required in wireless devices. This can be especially useful for compact wearable devices used for health monitoring, human activity recognition, and gesture recognition, where sending raw data over wireless protocols such as Bluetooth can be both time and power consuming. By performing analysis on the wearable device, wireless radio usage can be greatly decreased, reducing a main power consumer on the device. Deep learning (DL) methods, specifically using Tensorflow (TF) and Keras were evaluated for their usage in such a case, in this example gesture recognition. A multilayer long short-term memory (LSTM) model was trained and evaluated off of data (10 gestures, 1000 trials total, balanced) from a finger-worn ring profile device that collected acceleration data, and was found to perform with accuracy from 75-95% per gesture. The attempted conversion of the model into a compressed TF Lite format, to allow for analysis on-device did not succeed, due to current incompatibilities between the different frameworks. Future work may improve the accuracy, and potentially expand the use of neural networks on wearables for health diagnostics or as inputs devices.

Design and research of intelligent message platform system in hospital

Xi Luo (Dalian Neusoft University of Information, China); Jianbo Zheng (Chinese Academy of Sciences, China)

1
Aiming at the problems of existing in large-scale organizations, such as the disunity of message sending and receiving and management, the complexity and low efficiency of interconnection between departments and healthcare systems, this paper proposes the design of intelligent message collaboration system integrating multiple message interaction modes and questionnaire publishing. It realizes the message push mechanism which integrates SMS, email and instant communication, the feedback statistics mechanism based on questionnaire, and the performance and stability of the system are tested by providing a standard interface to connect with the business system. The experimental results show that the system is open, easy to expand and easy to maintain. It provides a solution for the unified release and supervision of information within large organizations such as hospital.

Data Fusion Enabled Approach for Sleep-aware Applications

Fan Yang (South China University of Technology & ShenZhen Institutes of Advanced Technology Chinese Academy of Sciences, China); Xiping Hu and Jianbo Zheng (Chinese Academy of Sciences, China)

1
In the big data era, thousands of hundreds of devices play the role of data producer as well as data consumer. However, wireless devices bear the power exhausted problem in various situations. How to balance the trade-off between data processing speed and power efficiency is meaningful to be researched. In this study, we propose a Sleep Data Fusion Networks module (SFDN) which has a star topology Bluetooth network to fuse data of sleep-aware applications basing on our designed application protocol. In the network, a center Bluetooth device fuses data generated from target node Bluetooth devices. Due to the low power consumption of Bluetooth as well as connection safety, our method is a better choice for a long lifetime and non-real time sleep data processing and fusion tasks then the Wi-Fi-based approach used in EAST and Smart-Alarm sleep-aware applications.

Research on fault-tolerant algorithm for emergency medical rescue UAV formation

Fei Tai Zhao and Yuan Shi Song (Xi'an University of Technology, China); Jianbo Zheng (Chinese Academy of Sciences, China)

1
With the maturing of Unmanned Aerial Vehicle (UAV) technology and the further expansion of aerial photography technology, China's civil unmanned aerial vehicle (UAV) application field is increasingly extensive. With disasters occurring suddenly, rescue forces are often unable to timely understand the situation and quickly arrive at the scene due to emergency rescue operations and poor rescue environment. UAV has the advantages of small size, low cost, flexibility, easy to use, low environmental requirements, which is suitable for war and disaster relief and other dangerous emergency environment. In this paper, a topology control algorithm based on k-hop local topology node mobility is proposed to realize the two-connected fault-tolerant network. The FTLMF algorithm proposed in this paper improves the success rate of implementing the two-connected fault-tolerant network under the constraint of node mobility, has a shorter adjustment period and offset distance, and maintain the connectivity of the original link. The topology control algorithm has a potential app-lication prospect in network fault-tolerant control.

Session S2-R1

Medical, Biomedical, and Health Informatics

Conference
2:00 PM — 3:30 PM HKT
Local
Feb 28 Sun, 10:00 PM — 11:30 PM PST

Healthcare Optimization

Jeff I Jones and Jim Jones (Collective Social Intelligence Pty Ltd, Australia)

1
Collective Social Intelligence (CSI) is an approach to new product development research that centers on creating an effective capability for big data analysis, the internet of things (IoT), online social networks and project portfolio management to improve the ways organizations and communities relate to each other in order to function collaboratively as a viable system. This has implications for all facets of society, most pertinently the institutions underpinning education, healthcare and government. In this paper we examine people, process and technology in relation to the need for optimizing healthcare to take full advantage of design thinking and systems thinking to produce change and continuous improvement through collaboration. We show healthcare policy and implementation as one example of why this approach is vitally important and how this approach creates change and improvement. We also describe CSI and our software R&D as solutions to mitigate the negative effects of social media on our institutions. This approach is predicated on a convergence of existing and highly regarded strategies for management and information flow with mechanisms and cultures driving modes of communication and visualization.

Efficient-CovidNet: Deep Learning Based COVID-19 Detection From Chest X-Ray Images

Yash Chaudhary and Manan Mehta (Maharaja Agrasen Institute of Technology, India); Raghav Sharma (University School of Information, Communication and Technology, India); Deepak Gupta (Maharaja Agrasen Institute of Technology, GGSIP University, India); Ashish Khanna (GGSIP University, India); Joel J. P. C. Rodrigues (Federal University of Piauí (UFPI), Brazil & Instituto de Telecomunicações, Portugal)

1
The COVID-19 pandemic has wreaked havoc all over the world. The rising number of cases have overburdened healthcare systems even in the most developed countries. To ease the burden on healthcare systems a quick and efficient testing technique is needed. Currently, the RT-PCR testing is done with time consuming and laborious an alternative is a detection from Chest X-Ray images. It has been discovered in published studies that Chest X-Rays of COVID-19 patients have specific malformations that can be used to identify a positive case. Inspired by the work done on "COVID-Net" by Linda Wang, Zhong Qiu Lin and Alexander Wong, a Deep Learning approach to detect coronavirus from Chest X-Ray images is used in this study. To surpass previous results the EfficientNet Convolutional Neural Network (CNN) model is proposed. This model not only achieves +2% accuracy, but it also attains higher sensitivity and Positive Predictive Values. The study uses the open source COVIDx dataset. It has approximately 14,000 X-Ray images. To the best of authors' knowledge, this dataset contains the largest number of COVID-19 positive cases. The study offers a Deep Learning approach contributing to create an efficient COVID-19 detector that can be used in the real world.

Attention Bias in Emotional Conflict in Major Depression Disorder: An Eye Tracking Study

Jing Zhu, Tao Gong, Zihan Wang and Chen Xia (Lanzhou University, China); Zhijie Ding (The Third People's Hospital of Tianshui City, China); Xiaowei Li (School of Information Science and Engineering, Lanzhou University, China)

1
Major depression disorder (MDD) has been proved to have difficulty in emotional conflict processing. The objective of this study is to investigate the different attention deployment patterns in emotional conflict processing between MDDs and Healthy Controls (HCs) using eye tracking data. A face-word Stroop task was used, and 49 MDDs and 50 healthy controls (HCs) were recruited in the experiment. Finally, our results indicated that MDDs demonstrated lower accuracy (ACC) compared with HCs during the process of emotional conflict. Moreover, we found attention bias in the process of attention maintenance but not vigilance, and it may be one of the possible reasons for different ability of emotional conflict processing between MDDs and HCs.

A New Skeletal Representation Based on Gait for Depression Detection

Haifeng Lu and Wei Shao (Lanzhou University, China); Edith C.-H. Ngai (The University of Hong Kong & Uppsala University, Hong Kong); Xiping Hu and Bin Hu (Lanzhou University, China)

1
As the challenge of depression problems increases today, it is important to effectively and timely detect for patients' treatments and depression prevention. Current methods of automated depression diagnosis depend almost entirely on audio, video, and Electroencephalogram (EEG) etc. In this paper, we propose a novel method to detect depression using gait data that is collected by Kinect camera. Its key components include a camera rectification method and a rigid-body representation of the human body. The camera rectification method achieves the purpose of improving data processing accuracy by changing the camera angle. The rigid-body representation can not only improve the robustness of detecting depression patients with noisy input, but also can reduce the classification time. We evaluate our method on the depression gait dataset in postgraduate students. The proposed method has a good performance in classic machine learning algorithms, and the best accuracy can achieve 88.89%. Our solution provides a new method for automatic depression detection (ADD) that has exciting implications in clinical theory and practice, and has the advantages of high accuracy, inexpensive, low time cost, and no-contact.

Session S2-R2

E-Health Services and Applications

Conference
2:00 PM — 3:30 PM HKT
Local
Feb 28 Sun, 10:00 PM — 11:30 PM PST

F-score Based EEG Channel Selection Methods for Emotion Recognition

Chengjian Zhao, Cancheng Li and Jinlong Chao (Lanzhou Univsersity, China); Tao Wang (Lanzhou University, China); Chang Lei (Lanzhou Univsersity, China); Juan Liu (Air Force Medical Center of FMMU, China); Hong Peng (Lanzhou University, China)

1
Emotion, as an advanced function of the human brain, affects kinds of human behaviors. Electroencephalographs (EEG) are widely used in the field of emotion classification owing to their low cost and portability. In this work, we study the effects of a non-linear EEG feature and a channel selection method on emotion recognition. First, the fractal dimension(FD) which could reflect the state of the brain is extracted with a sliding window. The top seven channels are screened out by calculating the F-score from the whole samples. Then, based on the signals from forehead channels, filtered channels and associated channels, emotions on valence and arousal are classified by Support Vector Machine(SVM) and K Nearest Neighbours(KNN). The result shows that the forehead channels Fp2, AF8, Fpz play an important role in valence classification. When combining the forehead channels with other channels that have higher F-score, the SVM classifier has a better accuracy on the whole set with 89.37% on valence and 87.07% on arousal. Besides, the overall accuracy calculated on each participants with associated channels get significant improvement. Especially, the KNN classifier has a much better result on every subject. This phenomenon indicates that by combining the higher F-score channels with the forehead channels, the associated channels can not only take advantage of the forehead channels' ability to categorize emotions but also consider individual differences.

Multimodal Surface Material Classification Based on Ensemble Learning with Optimized Features

Xiang Liu (Shandong University of Science and Technology, China); Hancheng Wu, Senlin Fang and Zhengkun Yi (Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, China); Xinyu Wu (Shenzhen Institutes of Advanced Technology, Chinese Academy of Science, China)

1
In this paper, we propose a novel method for multimodal material classification based on ensemble learning and optimized features. The proposed method consists of three key steps. Firstly, we extract a set of features for each modality. Compared to existing methods, the extracted features are relatively simple but more effective when they are incorporated into the classifiers. Then, the feature selection algorithms, including Multi-Cluster Feature Selection (MCFS) and Laplacian Score (LS) are employed to reduce the feature dimension due to the curse of dimensionality. Finally, an ensemble learning method is proposed to integrate the merits of different feature selection methods. The effectiveness of the proposed method is demonstrated on the LMT-108 surface material dataset which includes multiple modalities such as sound, acceleration, and image. The experimental results have shown that our approach performs better than the competing methods.

An Adaptive Attention Regulation Method Based on Biocybernetic Loop

Yi Zhang, Han Xiao, Jian Zhang and Hanshu Cai (Lanzhou University, China)

1
This study aims to establish an adaptive attention regulation method by combining neurofeedback and classical feedback control theory. It combined open-loop control and closed-loop negative feedback to achieve a universal attention regulation model. The EEG data under attention stimulation was collected, and the baseline threshold range was selected using the optimal threshold selection method, and each individual was calibrated individually. The sliding window method is used to evaluate the individual's attention state and make the feedback training scene to follow its changes. The experiment found that through adaptive feedback training, the individual can adjust the attention state to the individual's optimal baseline threshold. This method can cope with the differences between different individuals and provide each individual with the same level of feedback regulation. Similarly, this study may provide a universal treatment method for other mental diseases.

Exercise Intervention Framework of Emotion Regulation Based on Heart Rate Variability

Han Xiao and Yi Zhang (Lanzhou University, China); Xinchen Lin (LanZhou University, China); Hanshu Cai (Lanzhou University, China)

1
In this paper, we propose a personalized exercise intervention framework based on heart rate variability (HRV) to improve emotion regulation. Firstly, we deeply study the mechanism of exercise improving emotion regulation and introduce an "Emotion Regulation-ANS-Exercise" framework. From this connection, we use heart rate variability as the parameter of biofeedback. Secondly, we quantify a kind of movement state (represented by HRV) which is more beneficial to emotion regulation for each subject through the emotion regulation experiment, and then we design an exercise feedback system by combining biofeedback and PID controller. In this process, we calculate the target speed by the deviation between the target HRV and the current HRV. According to the target speed and real-time speed, the subject will adjust the exercise intensity to make their current HRV close to the target HRV. Finally, through the long-term comparison with the control group, we conclude that the personalized exercise intervention framework designed in this paper is more conducive to emotion regulation.

Session S2-R3

Communications and Networking

Conference
2:00 PM — 3:30 PM HKT
Local
Feb 28 Sun, 10:00 PM — 11:30 PM PST

Enhancement of Upper Limb Movement Classification based on Wiener Filtering Technique

Yazan Jarrah (SIAT-UCAS, China); Mojisola Asogbon (Shenzhen Institute of Advanced Technology, China); Oluwarotimi Samuel (Chinese Academy of Sciences, Beijing, China); Mingxing Zhu (University of Chinese Academy of Sciences & the CAS Key Laboratory of Human-Machine Intelligence-Synergy Systems, Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, China); Xin Wang (Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, China); Alberto López Delis (University of Brasília, Brazil); Weimin Huang (Shenzhen Children's Hospital, China); Shixiong Chen (Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, China); Guanglin Li (SIAT, China)

1
Electromyogram pattern recognition (EMG-PR) is considered a potential method for upper limb prosthesis control. In principle, the feature extraction technique has been ranked the most influential factor that affect the EMG-PR method's performance. Despite the progress made thus far, there are inevitable interferences that could not be handled by the usual signal filtering approaches that are applied to enhance the extracted features. To address this issue, this study proposed a technique based on Wiener filtering for the preprocessing of EMG signals towards improving the classification performance of EMG-PR systems. The performance of the proposed approach was investigated with high-density surface EMG recordings obtained from four transhumeral amputees who performed five classes of limb movements. Then, five time-domain features were analyzed in terms of their decoding accuracy, sensitive, and F1-score, with and without the application of the proposed technique for linear discriminant analysis and support vector machine classifiers. Experimental results showed that by applying the proposed technique to the different feature sets, significant improvements in classification accuracy, sensitivity, and F1-score were observed across all subjects and classifiers. The proposed method improved the average classification accuracy by an increase of approximately 6.24% compared with the conventional method, while an increment as high as 16.77% was recorded for individual classes of movements. The outcomes of this study indicate that Wiener filtering may potentially enhance the performance of EMG-PR systems in practical applications.

Blockchain-based Multi-role Healthcare Data Sharing System

Yao Yu and Qirui Li (Northeastern University, China); Qieshi Zhang (Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, China); Wenjian Hu and Shumei Liu (Northeastern University, China)

1
Blockchain has unique advantages in data privacy protection and data integrity. We can solve many security problems, such as the single point of failure and data sharing in the current centralized system through blockchain approach. Existing studies have demonstrated that the application of blockchain in the medical system could improve the patient's medical experience. However, we found that the blockchain-based medical systems didn't consider the problems of long insurance claim cycle and complicated procedures. There are also very few healthcare systems that provide targeted sharing protocols for medical data and personal health data. In this paper, we propose a multi-role healthcare data sharing system framework based on blockchain. In this system, we reduce the storage cost through the collaborative storage of blockchain and IPFS. Then we design a smart contract on insurance to help patients achieve automatic insurance claim. In addition, we design two different sharing protocols to realize the fine management of personal data. The system analysis shows that our proposed blockchain-based multi-role healthcare data sharing system can effectively address the actual needs of users and has perfect performance in data storage, privacy protection, insurance claims and personal data management.

Cooperative Wireless Edges with Composite Resource Allocation in Hierarchical Networks

Nan Li (Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, China); Xiping Hu (Chinese Academy of Sciences, China); Edith C.-H. Ngai (The University of Hong Kong & Uppsala University, Hong Kong); Erol Gelenbe (Institute of Theoretical and Applied Informatics, Polish Academy of Sciences, Poland & University of Cote d'Azur, France)

1
With the expansion of the IoT, it is important to optimize available bandwidth to reliably support edge to device communications. Thus we propose a wireless network where each edge server communicates with its end devices using its wireless band as a primary channel, assisted by a secondary edge server that can relay communications via its own wireless band as a secondary channel. The network can optimize capacity by balancing load between primary and secondary wireless bands, and we analyze the geometry of achievable rate regions, depending on the state of bands modeled as Rayleigh fading channels. The allocation of a connection to the primary or secondary band is formulated as an optimization problem which is then solved, and illustrated with numerical examples.

Supply Distribution Center Planning in UAV-based Logistics Networks for Post-Disaster Supply Delivery

Yang Huang (Army Logistics University, China); Han Han and Bo Zhang (National Innovation Institute of Defense Technology, China); Zhansheng Gong and Xisheng Su (Army Logistics University, China)

1
Using UAVs to transport emergency supplies has the potential to greatly reduce the time and costs in post disaster rescue missions, and the prerequisite is the efficient planning of a UAV delivery networks. This paper considers the supply distribution center (SDC) planning issue in UAV-based logistics networks (UAV-LNs) and formulate an optimization problem, which aims at minimizing the average timeliness of UAV-LNs. By considering both the service capability of heterogenous nodes and the robustness of different demand points, a genetic algorithm is devised to tackle the high-dimensional optimization problem. Numerical results are conducted to compare and analyze the scaling performance of the algorithm in network, resource and area dimensions. It is shown that the proposed genetic algorithm is robust and of low complexity across various scenarios.

Session S3-R1

Medical, Biomedical, and Health Informatics

Conference
4:00 PM — 5:30 PM HKT
Local
Mar 1 Mon, 12:00 AM — 1:30 AM PST

A Novel Gait Analysis Method Based on The Pseudo-velocity Model for Depression Detection

Tao Wang and Jieqiong Sun (Lanzhou University, China); Jinlong Chao (Lanzhou Univsersity, China); Shuzhen Zheng (Lanzhou University, China); Chengjian Zhao (Lanzhou Univsersity, China); Chunyun Wu and Hong Peng (Lanzhou University, China)

1
As the occurrence of depression in society becomes increasingly more common, it is an urgent task to find more objective and effective tools for real-time depression assessment. Gait analysis offers a new low-cost and contactless method for depression diagnosis. Therefore, interest in gait-based depression detection using depth sensors, such as Kinect, has grown rapidly in recent years. In this paper, a pseudo-velocity model is built to analyze the abnormal gait related to the depression by combining the velocity and angular velocity of the joints. Subsequently, we extract some features in time and frequency domain from our model to establish the classification model for depression detection. Experimental results on depression gait data recordings from 43 scored-depressed and 52 non-depressed individuals show that the proposed method achieves a good classification accuracy of 92.35% and is superior to other existing methods. The outstanding classification performance suggests that the proposed method has potential clinical value in depression detection.

A Machine Learning Method to Track the Disease Deterioration Risk for Acute Exacerbation of Chronic Obstructive Pulmonary Diseases

Junfeng Peng (Guangdong University of Education)

1
Due to the complexity and high heterogeneity of the Acute Exacerbation of Chronic Obstructive Pulmonary Disease (AECOPD), the guidelines (global initiative for chronic obstructive, GOLD) is unable to fully guide the treatment of AECOPD. How to provide a rapid treatment in line with the development of the AECOPD after admission is still an important clinical issue. The generation time of clinical data is the key to achieve the goal. In this paper, we propose a model combining the random forest with the data generation time to track the diseases deterioration risk of the AECOPD. The experimental results show that the proposed model has the potential to track the whole process deterioration risk and assists the junior clinicians in diagnosing deterioration risk for AECOPD patients.

A Control Method With Terrain Classification and Recognition for Lower Limb Soft Exosuit

Jiangpeng Ni, Chen Chunjie and Zhuo Wang (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); Liu Yida (Shenzhen Institutes of Advanced Technology Chinese Academy of Sciences, China)

1
Soft Exosuit is a kind of Lower-limb wearable robots to augment and assist the wearer's performance. The wearer needs different assistance modes to reduce the metabolic rate when walking on different terrains. Therefore, assistance modes need to be selected according to different terrains for the wearer of soft Exosuit. Recently, our team has designed a stable terrain classification and recognition system (TCRS) for the soft Exosuit to discriminate terrains and estimating environmental features. Through this system, soft Exosuit can perceive the environment to auxiliary control of the locomotion modes. A depth sensor with an inertial measurement unit(IMU) can acquire to stabilize the point cloud of environments. Subsequently, the 2D point cloud is extracted from the origin 3D point cloud. Then, they are classified to estimate terrain environmental features, including the incline angle of the slope, the width, and the stairs' height. Finally, the TCRS was evaluated by classifying and recognizing five basic terrains in three different experimental scenarios while the subject was wearing the soft Exosuit with the TCRS module. The results show that the terrain classification accuracy rate reaches 97.74%, and the environmental features estimation error is less than 15%. The promising results indicate the robustness and the potential application of the presented TCRS to provide proper auxiliary force to reduce the metabolic rate of wearers on different terrains.

Integration of a novel attribute and classical topology metrics of hyper-networks for automatic diagnosis of Major depressive disorder

Yongchao Li, Nan Chen, Yin Wang and Lin Yang (Lanzhou University, China); Weihao Zheng (Zhejiang University, China); Zhijun Yao and Bin Hu (Lanzhou University, China)

1
Conventional hyper-network coefficients ignore the weighted hyper-edge information which could be vital in researching the specificity of brain disease. Functional hyper-networks for 64 healthy controls (HC) and 56 patients with major depressive disorder (MDD) were constructed using the least absolute shrinkage and selection operator (Lasso). Not only the classical topology metrics but also a novel hyper-edge weight (HEW) attribute were extracted as features to promote the functional-based auto-diagnosis accuracy of MDD. We compared the categorization performance of each hyper-network coefficient. A multi-feature ensemble model was applied to fuse different kinds of features. We obtained 82.15 % accuracy with the classical hyper-network clustering coefficient (HCC) and 84.08 % accuracy with the HEW attribute on the MDD dataset. The performance was further improved to 89.24% by combining all the properties of the hyper-networks. The multi-feature ensemble model combining different hyper-network coefficients provides new insights into the automatic diagnosis with diverse information of MDD.

Session S3-R2

E-Health Services and Applications

Conference
4:00 PM — 5:30 PM HKT
Local
Mar 1 Mon, 12:00 AM — 1:30 AM PST

Segmentation of pulmonary vessels based on MSFM method

Xiaowei Wang, LiYing Cheng and XuanShuang Gao (Shenyang Normal University, China); Nan Li (Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, China); Wenjun Tan (Northeastern University, China)

1
Accurate segmentation of pulmonary blood vessels from CT images is of great significance for lung disease detection and segmentation of other lung structures. Manual segmentation is difficult to accurately segment vascular tissue for various reasons. Therefore, in view of the existing problems and shortcomings of the existing lung vessel segmentation method, a more efficient lung vessel segmentation algorithm is proposed, that is, the multi-template fast marching method (MSFM algorithm). Firstly, it used the hole filling and maximum inter-class variance algorithm for preprocess. In the process, the lung parenchyma is extracted from the chest CT, and then the lung blood vessels are extracted in the lung parenchymal area using the MSFM algorithm. In the extraction process, the threshold and gradient are used to limit the progress of the process and the lung blood vessels are more accurately segmented. Through experimental verification, the accuracy of lung blood vessel segmentation based on MSFM algorithm is improved.

A novel smartphone App-based assessment of gait in Parkinson's disease

Doning Su and Zhu Liu (Beijing Tiantan Hospital, China); Xin Jiang (Shenzhen People’s Hospital, China); Wanting Yu (Hinda and Arthur Marcus Institute for Aging Research, Hebrew SeniorLife, USA); Huizi Ma, Zhan Wang, Chunxue Wang and Xuemei Wang (Beijing Tiantan Hospital, China); Wanli Hu (Beijing ChaoYang Hospital, China); Brad Manor (Hinda and Arthur Marcus Institute for Aging Research, Hebrew SeniorLife, USA); Tao Feng (Beijing Tiantan Hospital, China); Junhong Zhou (Harvard Medical School, USA)

1
The measurement of gait characteristics in patients with PD may provide insights into the locomotor control and other factors in PD, and helps the management of this disease. We developed a smartphone-based assessment of gait-completed with the phone placed in the user's pocket-for use in remote and non-laboratory settings. We here tested the validity of App-derived metrics of gait in PD and the association between the gait metrics and clinical and functional characteristics. Fifty-two patients with clinically-diagnosed PD completed this study, consisting of App walking tests, UPDRS III for disease severity and MoCA for cognitive function. During the walking tests, they followed multi-media instructions provided by the App to complete two 45-second trials each of walking normally (i.e., single task) and walking while performing a serial subtraction dual task (i.e., dual task condition, DT). Gait data were simultaneously collected with the App and a gold-standard Mobility Lab wearable sensor system. The stride times (ST) and stride time variability (STV) of gait were derived from the acceleration and angular velocity signal acquired from the phone's internal motion sensor, and from the sensors of Mobility Lab. Strong correlations between the ST and ST derived from the App and those from Mobility Lab were observed (r>0.95, p<0.0001), revealing excellent validity of the App-based gait assessment. Compared to single task walking, the ST and STV in dual task walking were significantly greater (F>6.3, p<0.01), indicating more unstable gait in this condition. The STV in both single and dual task walking was associated with the total score of UPDRS III, that is, those with greater STV had greater UPDRS III score (r>0.37, p<0.01). The STV in dual task condition and the dual task costs (i.e., percent change from single to dual task condition) to STV were also correlated with MoCA score (r>0.44, p<0.004). Participants with greater STV and dual task costs to STV had lower MoCA score (i.e., poorer cognitive function). The smartphone App we created enabled ease-of-its-use assessment of gait in different task conditions with excellent validity, in those with PD.

Multi-view Weighted Feature Fusion Using CNN for Pneumonia Detection on Chest X-Rays

Shaoliang Peng and Xiongjun Zhao (Hunan University, China); Xiaoyong Wei (Sichuan University, China); Donqing Wei (Shanghai Jiao Tong University, China); Yuehua Peng (Dalian Naval Academy, China)

1
Chest X-ray is still the most common and important method for diagnosing pneumonia. However, the analysis of chest radiographs requires professional radiologists, and over-reliance on radiologists may lead to erroneous diagnosis or missed diagnosis. Using convolutional neural networks(CNNs) for diagnosis chest diseases on chest X-ray has achieved better results, but most of the previous models are only trained by frontal-view X-ray images. Unique from past studies, in this paper, we proposed a model that can learn multi-view semantic information from chest X-rays to detect pneumonia. Our model includes two stages of feature extraction and feature fusion, and is trained on MIMIC-CXR-JPG dataset, currently the largest publicly available chest x-ray dataset, containing 377,110 JPG format images. We demonstrate that such multi-view weighted feature fusion model outperforms the models that use features only from one view. Our results are better than previous models for pneumonia detection.

Alzheimer's Disease Distinction Based On Gait Feature Analysis

ZhiYang You (LanZhou University, China); Zeng You (Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, China); Yilong Li (University of Wisconsin-Madison, China); Shipeng Zhao (Lanzhou University, China); Huixia Ren (Shenzhen People’s Hospital, China); Xiping Hu (Lanzhou University, China)

1
Alzheimer's disease(AD) is a neurodegenerative disease that progresses slowly but worsens gradually, also, the most common kinds of dementia. Clinically, the diagnosis of AD is mainly based on rating scales and neuroimaging technology which is invasive, costly and time-consuming. Other than that, the clinical pathology has become irreversible when neuroimaging characteristics appear. It is imperative to develop new noninvasive methods for early diagnosis of AD. Several studies indicated the probable association of cognitive decline with gait changes might shed light on potential features for distinction of AD. This paper aims to exploit the feasibility of gait features for early diagnosis of mild cognitive impairment(MCI) and AD by using machine learning methods. A device-free AD detection system is built, with a natural undisturbed gait collecting system and a well-performed Long Short-Term Memory(LSTM) based model, in this article. Moreover, it can serve as a simplified, non-invasive, and highly accurate clinical auxiliary tool for early diagnosis and distinction of AD. Experimental results showed a 90.48%, 92.00%, and 88.24% in accuracy, sensitivity, and specificity respectively for distinguishing AD by using the method with LSTM based model. Furthermore, the gait cycle and stride length in MCI or AD were more variable than in healthy controls through redefining and calculating the gait features with skeleton data obtained by Kinect devices.

Session S3-R3

Communications and Networking

Conference
4:00 PM — 5:30 PM HKT
Local
Mar 1 Mon, 12:00 AM — 1:30 AM PST

Multi-level Co-occurrence Graph Convolutional LSTM for Skeleton-based Action Recognition

Shihao Xu (Lanzhou University, China); Haocong Rao (Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, China); Xiping Hu and Bin Hu (Lanzhou University, China)

1
Human action recognition plays an important role in e-health applications, such as surgical skill analysis, patient monitoring, and automatic nursing systems. Recently, skeleton-based action recognition gains massive attention. It is an essential yet challenging task that requires effectively modeling the intra-frame skeleton representation and inter-frame temporal dynamics. Traditional Long Short-Term Memory (LSTM) based methods mainly capture long-term action context information from global level, yet they cannot fully model the relationship between different joints or persons to mine crucial co-occurrence features from different levels. To overcome this drawback, we propose a general end-to-end Multi-level Co-occurrence Graph Convolutional LSTM (MCGC-LSTM). By incorporating graph convolutional networks (GCN) into LSTM, our model can not only better exploit body structural information from skeletons but also enhance the multi-level co-occurrence feature learning. Specifically, we first devise multi-level co-occurrence (MC) memory units coupled with GCN to automatically model the spatial relationship between joints, and simultaneously capture the co-occurrence features from different joints, persons, and frames. Then we construct aggregated features of multi-level co-occurrences (AFMC) from MC memory units to better represent the intra-frame action context encoding, and leverage a concurrent LSTM (Co-LSTM) to further model their temporal dynamics for action recognition. Experiments show that our proposed model significantly outperforms mainstream methods on NTU RGB+D 120 dataset and Northwestern-UCLA dataset.

A Novel Segmentation-based Adaptive Feature Extraction Methodology for Discriminating Heart Sounds

Shuping Sun (4 Building, China); Dayong Huang, Biqiang Zhang and Peiguang He (Nanyang Institute of Technology, China); Long Yan (Shenzhen Univ, China); Tingting Huang (Nanyang Institute of Technology, China)

1
To utilize heart sound features that may vary according to their suitability for segmentation, automatic adaptive feature extraction methodology is proposed to discriminate heart sounds. The innovation of this methodology is primarily reflected in the automatic segmentation and extraction of the first complex heart sound (CS1) and second complex heart sound (CS2) or each cardiac sound (CS), and automatic extraction of the segmentation-based frequency feature FF1 or FF2, determination of the diagnostic features [γ11, γ12] and [γ21, γ22, γ23]. Two stages corresponding to the implementation of the novel methodology are summarized as follows. In stage 1, the time intervals between two sequential peaks are automatically calculated and statistically analyzed, and the result is used to determine whether a given heart sound can be segmented. Stage 2 involves automatic extraction of segmentation-based adaptive features for adapting the heart sound to the frequency domain. The performance evaluation was validated using the scatter diagrams of the features extracted from the heart sounds from online databases and clinical databases.

A Comprehensive Channel and Feature Selection Method for Myoelectric Pattern Recognition

Mengyao Li (Shenzhen Institutes of Advanced Technology, University of Chinese Academy of Sciences, China); Yue Ma, Liangsheng Zheng, Can Wang and Wei Feng (Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, China); Xinyu Wu (Shenzhen Institutes of Advanced Technology, Chinese Academy of Science, China)

1
The advent of myoelectric control schemes provides promising chances for locomotion empowerment and restoration of those with disabilities. Despite substantial efforts have been made into advancing sEMG-based motion recognition, it may be a little tricky to determine appropriate muscles and features for people with muscle disorders or different muscle use preferences. To mitigate it, an advantageous sEMG channel and feature selection method based on ReliefF algorithm was proposed. Related experiments were conducted on a eight able-bodied subject database to showcase the feasibility and efficiency of the proposed approach, that is, considerable high classification performance was maintained with the original feature set reduced by more than half. Ulteriorly, we also investigated the influences of different number of neighbors or features on classification accuracy for ascertaining the optimal values. The strengths of our proposed method lie in not only customized channel and feature selection for individual users with respective muscle development and use preference, but also offering preliminary insight for a general mapping mechanism between human muscles and corresponding motions.

A deep metric learning algorithm for similarity measure of the gene expression profile

Shaoliang Peng, Lei Zhang, Yaning Yang and Wei Liu (Hunan University, China); Fei Li (Computer Network Information Center, Chinese Academy of Sciences, China); Hao Hong (Beijing Institute of Radiation Medicine, China); Shulin Wang (Hunan University, China)

1
Clustering gene expression profiles is a fundamental task in the genome and biomedical research. With the development of RNA-seq and gene chip technology, mass gene expression profile data has been generated, which puts forward two requirements for related research of gene expression profile: i) accurate analysis of drug R&D requires high accuracy of similarity analysis, ii) large-scale analysis of data requires as little running time as possible. We propose a faster, more accurate method called DeepCDNet, which is based on the framework of the Siamese network. DeepCDNet uses the DenseNet structure and optimized loss function to achieve rapid convergence, and the similarity between expression spectra is calculated by a cosine function. The experiment results show that: i) our method breaks through the limitation of high dimensions of gene expression profile and can quickly and accurately learn the required gene characteristics, ii) The accuracy of our method in similarity analysis is greatly improved, iii) as the dimension of data increases, the advantage of our method on time cost gradually becomes more prominent, and time consumption is less.

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