Tanvi Banerjee

Day or Night Activity Recognition From Video Using Fuzzy Clustering Techniques

We present an approach for activity state recognition implemented on data collected from various sensors-standard web cameras under normal illumination, web cameras using infrared lighting, and the inexpensive Microsoft Kinect camera system. Sensors such as the Kinect ensure that activity segmentation is possible during the daytime as well as night. This is especially useful for activity monitoring of older adults since falls are more prevalent at night than during the day. This paper is an application of fuzzy set techniques to a new domain. The approach described herein is capable of accurately detecting several different activity states related to fall detection and fall risk assessment including sitting, being upright, and being on the floor to ensure that elderly residents get the help they need quickly in case of emergencies and ultimately to help prevent such emergencies.

Recognizing complex instrumental activities of daily living using scene information and fuzzy logic

Provides a unique and robust solution to the extremely challenging task of ADL modeling.Incorporates scene information to build ADL models.In the absence of manually labeled surfaces, can still generate high-level activity state summaries.We provide a dataset for the computer vision community that is described in this manuscript. We describe a novel technique to combine motion data with scene information to capture activity characteristics of older adults using a single Microsoft Kinect depth sensor. Specifically, we describe a method to learn activities of daily living (ADLs) and instrumental ADLs (IADLs) in order to study the behavior patterns of older adults to detect health changes. To learn the ADLs, we incorporate scene information to provide contextual information to build our activity model. The strength of our algorithm lies in its generalizability to model different ADLs while adding more information to the model as we instantiate ADLs from learned activity states. We validate our results in a controlled environment and compare it with another widely accepted classifier, the hidden Markov model (HMM) and its variations. We also test our system on depth data collected in a dynamic unstructured environment at TigerPlace, an independent living facility for older adults. An in-home activity monitoring system would benefit from our algorithm to alert healthcare providers of significant temporal changes in ADL behavior patterns of frail older adults for fall risk, cognitive impairment, and other health changes.

Testing an in-home gait assessment tool for older adults

In this paper, we present results of an automatic vision-based gait assessment tool, using two cameras. Elderly residents from TigerPlace, a retirement community, were recruited to participate in the validation and test of the system in scripted scenarios representing everyday activities. The residents were first tested on a GAITRite mat, an electronic walkway that captures footfalls, and with inexpensive web cameras recording images. The extracted gait parameters from the camera system were compared with the GAITRite; excellent agreement was achieved. The residents then participated in the scenarios, with only the cameras recording. We found that the residents displayed different gait patterns during the realistic scenarios compared to the GAITRite runs. This finding provides support of the importance and advantage of continuous gait assessment in a daily living environment. Results on 4 elderly participants are included in the paper.

Sit-to-stand detection using fuzzy clustering techniques

The ability to rise from a chair is an important parameter to assess the balance deficits of a person. In particular, this can be an indication of risk for falling in elderly persons. Our goal is automated assessment of fall risk using video data. Towards this goal, we present a simple yet effective method of detecting transition, i.e. sit-to-stand and stand-to-sit, from image frames using fuzzy clustering methods on image moments. The technique described in this paper is shown to be robust even in the presence of noise and has been tested on several data sequences using different subjects yielding promising results.

Improvement of acoustic fall detection using Kinect depth sensing

The latest acoustic fall detection system (acoustic FADE) has achieved encouraging results on real-world dataset. However, the acoustic FADE device is difficult to be deployed in real environment due to its large size. In addition, the estimation accuracy of sound source localization (SSL) and direction of arrival (DOA) becomes much lower in multi-interference environment, which will potentially result in the distortion of the source signal using beamforming (BF). Microsoft Kinect is used in this paper to address these issues by measuring source position using the depth sensor. We employ robust minimum variance distortionless response (MVDR) adaptive BF (ABF) to take advantage of well-estimated source position for acoustic FADE. A significant reduction of false alarms and improvement of detection rate are both achieved using the proposed fusion strategy on real-world data.

What Are People Tweeting About Zika? An Exploratory Study Concerning Its Symptoms, Treatment, Transmission, and Prevention

Background In order to harness what people are tweeting about Zika, there needs to be a computational framework that leverages machine learning techniques to recognize relevant Zika tweets and, further, categorize these into disease-specific categories to address specific societal concerns related to the prevention, transmission, symptoms, and treatment of Zika virus. Objective The purpose of this study was to determine the relevancy of the tweets and what people were tweeting about the 4 disease characteristics of Zika: symptoms, transmission, prevention, and treatment. Methods A combination of natural language processing and machine learning techniques was used to determine what people were tweeting about Zika. Specifically, a two-stage classifier system was built to find relevant tweets about Zika, and then the tweets were categorized into 4 disease categories. Tweets in each disease category were then examined using latent Dirichlet allocation (LDA) to determine the 5 main tweet topics for each disease characteristic. Results Over 4 months, 1,234,605 tweets were collected. The number of tweets by males and females was similar (28.47% [351,453/1,234,605] and 23.02% [284,207/1,234,605], respectively). The classifier performed well on the training and test data for relevancy (F1 score=0.87 and 0.99, respectively) and disease characteristics (F1 score=0.79 and 0.90, respectively). Five topics for each category were found and discussed, with a focus on the symptoms category. Conclusions We demonstrate how categories of discussion on Twitter about an epidemic can be discovered so that public health officials can understand specific societal concerns within the disease-specific categories. Our two-stage classifier was able to identify relevant tweets to enable more specific analysis, including the specific aspects of Zika that were being discussed as well as misinformation being expressed. Future studies can capture sentiments and opinions on epidemic outbreaks like Zika virus in real time, which will likely inform efforts to educate the public at large.

Monitoring Patients in Hospital Beds Using Unobtrusive Depth Sensors

We present an approach for patient activity recognition in hospital rooms using depth data collected using a Kinect sensor. Depth sensors such as the Kinect ensure that activity segmentation is possible during day time as well as night while addressing the privacy concerns of patients. It also provides a technique to remotely monitor patients in a non-intrusive manner. An existing fall detection algorithm is currently generating fall alerts in several rooms in the University of Missouri Hospital (MUH). In this paper we describe a technique to reduce false alerts such as pillows falling off the bed or equipment movement. We do so by detecting the presence of the patient in the bed for the times when the fall alert is generated. We test our algorithm on 96 hours obtained in two hospital rooms from MUH.

Investigation of an Indoor Air Quality Sensor for Asthma Management in Children

Monitoring indoor air quality is critical because Americans spend 93 of their life indoors, and around 6.3 million children suffer from asthma. We want to passively and unobtrusively monitor the asthma patients environment to detect the presence of two asthma-exacerbating activities: smoking and cooking using the Foobot sensor. We propose a data-driven approach to develop a continuous monitoring-activity detection system aimed at understanding and improving indoor air quality in asthma management. In this study, we were successfully able to detect a high concentration of particulate matter, volatile organic compounds, and carbon dioxide during cooking and smoking activities. We detected 1) smoking with an error rate of 1; 2) cooking with an error rate of 11; and 3) obtained an overall 95.7 percent accuracy classification across all events (control, cooking and smoking). Such a system will allow doctors and clinicians to correlate potential asthma symptoms and exacerbation reports from patients with environmental factors without having to personally be present.

Knowledge-Driven Personalized Contextual mHealth Service for Asthma Management in Children

Wide adoption of smartphones and availability of low-cost sensors has resulted in seamless and continuous monitoring of physiology, environment, and public health notifications. However, personalized digital health and patient empowerment can become a reality only if the complex multisensory and multimodal data is processed within the patient context. Contextual processing of patient data along with personalized medical knowledge can lead to actionable information for better and timely decisions. We present a system called kHealth capable of aggregating multisensory and multimodal data from sensors (passive sensing) and answers to questionnaire (active sensing) from patients with asthma. We present our preliminary data analysis comprising data collected from real patients highlighting the challenges in deploying such an application. The results show strong promise to derive actionable information using a combination of physiological indicators from active and passive sensors that can help doctors determine more precisely the cause, severity, and control level of asthma. Information synthesized from kHealth can be used to alert patients and caregivers for seeking timely clinical assistance to better manage asthma and improve their quality of life.

Sit-to-Stand Measurement for In-Home Monitoring Using Voxel Analysis

We present algorithms to segment the activities of sitting and standing, and identify the regions of sit-to-stand (STS) transitions in a given image sequence. As a means of fall risk assessment, we propose methods to measure STS time using the 3-D modeling of a human body in voxel space as well as ellipse fitting algorithms and image features to capture orientation of the body. The proposed algorithms were tested on ten older adults with ages ranging from 83 to 97. Two techniques in combination yielded the best results, namely the voxel height in conjunction with the ellipse fit. Accurate STS time was computed on various STSs and verified using a marker-based motion capture system. This application can be used as part of a continuous video monitoring system in the homes of older adults and can provide valuable information to help detect fall risk and enable early interventions.