Human Motion Analysis Based on Sequential Modeling of Radar Signal and Stereo Image Features

Falls are one of the greatest threats to elderly health in their daily living routines and activities. Therefore, it is very important to detect falls of an elderly in a timely and accurate manner, so that immediate response and proper care can be provided, by sending fall alarms to caregivers. Radar is an effective non-intrusive sensing modality which is well suited for this purpose, which can detect human motions in all types of environments, penetrate walls and fabrics, preserve privacy, and is insensitive to lighting conditions. Micro-Doppler features are utilized in radar signal corresponding to human body motions and gait to detect falls using a narrowband pulse-Doppler radar. Human motions cause time-varying Doppler signatures, which are analyzed using time-frequency representations and matching pursuit decomposition (MPD) for feature extraction and fall detection. The extracted features include MPD features and the principal components of the time-frequency signal representations. To analyze the sequential characteristics of typical falls, the extracted features are used for training and testing hidden Markov models (HMM) in different falling scenarios. Experimental results demonstrate that the proposed algorithm and method achieve fast and accurate fall detections. The risk of falls increases sharply when the elderly or patients try to exit beds. Thus, if a bed exit can be detected at an early stage of this motion, the related injuries can be prevented with a high probability. To detect bed exit for fall prevention, the trajectory of head movements is used for recognize such human motion. A head detector is trained using the histogram of oriented gradient (HOG) features of the head and shoulder areas from recorded bed exit images. A data association algorithm is applied on the head detection results to eliminate head detection false alarms. Then the three dimensional (3D) head trajectories are constructed by matching scale-invariant feature transform (SIFT) keypoints in the detected head areas from both the left and right stereo images. The extracted 3D head trajectories are used for training and testing an HMM based classifier for recognizing bed exit activities. The results of the classifier are presented and discussed in the thesis, which demonstrates the effectiveness of the proposed stereo vision based bed exit detection approach.

Designing for User-facing Uncertainty in Everyday Sensing and Prediction

Thesis (Ph.D.)--University of Washington, 2016-07

A Mobile High Sensitivity On-field Organophosphorous Compounds Detecting System for IoT Based Food Safety Tracking

Food safety has always been a social issue that draws great public attention. With the rapid development of wireless communication technologies and intelligent devices, more and more Internet of Things (IoT) systems are applied in the food safety tracking field. However, connection between things and information system is usually established by pre-storing information of things into RFID Tag, which is inapplicable for on-field food safety detection. Therefore, considering pesticide residue is one of the severe threaten to food safety, a new portable, high-sensitivity, low-power, on-field organophosphorus (OP) compounds detection system is proposed in this thesis to realize the on-field food safety detection. The system is designed based on optical detection method by using a customized photo-detection sensor. A Micro Controller Unit (MCU) and a Bluetooth Low Energy (BLE) module are used to quantize and transmit detection result. An Android Application (APP) is also developed for the system to processing and display detection result as well as control the detection process. Besides, a quartzose sample container and black system box are also designed and made for the system demonstration. Several optimizations are made in wireless communication, circuit layout, Android APP and industrial design to realize the mobility, low power and intelligence.

Methods for the Quantification of Motor Stability for the Assessment of Fall Risk

The research field of the Thesis is the evaluation of motor variability and the analysis of motor stability for the assessment of fall risk. Since many falls occur during walking, a better understanding of motor stability could lead to the definition of a reliable fall risk index aiming at measuring and assessing the risk of fall in the elderly, in the attempt to prevent traumatic events. Several motor variability and stability measures are proposed in the literature, but still a proper methodological characterization is lacking. Moreover, the relationship between many of these measures and fall history or fall risk is still unknown, or not completely clear. The aim of this thesis is hence to: i) analyze the influence of experimental implementation parameters on variability/stability measures and understand how variations in these parameters affect the outputs; ii) assess the relationship between variability/stability measures and long- short-term fall history. Several implementation issues have been addressed. Following the need for a methodological standardization of gait variability/stability measures, highlighted in particular for orbital stability analysis through a systematic review, general indications about implementation of orbital stability analysis have been showed, together with an analysis of the number of strides and the test-retest reliability of several variability/stability numbers. Indications about the influence of directional changes on measures have been provided. The association between measures and long/short-term fall history has also been assessed. Of all the analyzed variability/stability measures, Multiscale entropy and Recurrence quantification analysis demonstrated particularly good results in terms of reliability, applicability and association with fall history. Therefore, these measures should be taken in consideration for the definition of a fall risk index.

Towards automated Android app collusion detection

Android OS supports multiple communication methods between apps. This opens the possibility to carry out threats in a collaborative fashion, c.f. the Soundcomber example from 2011. In this paper we provide a concise definition of collusion and report on a number of automated detection approaches, developed in co-operation with Intel Security.

IL PROTOCOLLO LiFE. COMPARAZIONE TRA LA VERSIONE INDIVIDUALE E DI GRUPPO PER PREVENIRE LE CADUTE NELL’ANZIANO. UNA REVISIONE SISTEMATICA DELLA LETTERATURA

Background: Negli ultimi anni si è verificato un significativo aumento dell’aspettativa di vita che ha portato ad un incremento dell’età media della popolazione e di conseguenza un maggior rischio di cadere. Ricercare una strategia per promuovere il mantenimento della propria salute e ridurre le cadute risulta quindi di fondamentale importanza. Nasce così il programma LiFE basato sull’integrazione di attività di forza, equilibrio e PA all’interno della quotidianità. Obiettivo: Ricercare una strategia efficace a lungo termine ed economicamente sostenibile per ridurre le cadute nell’anziano e le conseguenze ad esse correlate. Materiali e metodi: Per la realizzazione di questa revisione sistematica è stato seguito il PRISMA Statement. Sono stati ricercati Trial Clinici Randomizzati che confrontassero la versione originale di LiFE a quella di gruppo gLiFE all’interno del sito dello studio e alcune banche dati. Il rischio di Bias di ogni articolo incluso è stato valutato tramite la PEDro scale. Risultati: Sono stati inclusi 3 articoli tramite i quali si è studiata l’efficacia, l’aderenza ed il costo dei due programmi analizzati. Non sono state dimostrate differenze statisticamente significative né per efficacia né per aderenza. Il costo economico totale per ogni partecipante al gruppo gLiFE è risultato inferiore in entrambi gli studi che lo hanno analizzato rispetto ai partecipanti a LiFE. Dividendo il costo totale in spesa a carico del partecipante e spesa a carico della società, risulta essere conveniente l’applicazione di gruppo per il singolo ma non per la società. Conclusioni: Il protocollo gLiFE è comparabile al protocollo LiFE in termini di efficacia, aderenza e costo. Si dovrebbe dare la possibilità agli anziani di poter scegliere il programma più adatto a loro in base alle loro caratteristiche ed esigenze. Non essendoci valutazioni a lungo termine sono necessari ulteriori studi per confrontare il costo-efficacia fra LiFE e gLiFE ma anche fra altri protocolli.

Sistema de apoio a rondas de segurança

Dissertação de mestrado integrado em Engenharia Eletrónica Industrial e Computadores

The rise and fall in menopausal hormone therapy and breast cancer incidence.

Studies conducted in different areas of North America and Europe showed a 5-10% decline in the incidence of breast cancer following reductions up to 70% in menopause hormone therapy (HT) use after 2002. The observation that the decline was larger in (or limited to) women aged > or =50 years weighs in favour of an effect of reduced HT use on breast cancer incidence. However, changes in screening are also likely to play a role in the decreasing incidence of breast cancer observed in several countries. In particular, the technical improvements and the increased effectiveness of breast cancer screening and detection during the 1990s led to a decreased number of pre-clinical cases found by screening in subsequent years. Further, disentangling the effects of HT use and screening is difficult, as women who stop using HT may also undergo mammography screening less frequently. Thus, the reasons of the falls in incidence remain open to discussion.

Detection of recurrence of large-bowel carcinoma by radioimmunoassay of circulating carcinoembryonic antigen (C.E.A.).

The usefulness and limitations of the carcinoembryonic antigen (C.E.A.) radioimmunoassay for the evaluation of tumour resection and for the detection of tumour relapse were studied in patients with large-bowel carcinoma. The level of plasma-C.E.A. was determined before any treatment in a group of 101 patients with histologically proven adenocarcinoma of the colon and rectum. 71% of all patients and 63% of cases with localised tumour (Dukes A and B) had a preoperative C.E.A. value of 5 ng. per ml. or higher. This limit was reached by only 1 of 90 apparently healthy, non-smoking blood-donors. Among 45 patients for whom a complete tumour resection was reported, all patients except 5 showed a drop of C.E.A. to normal values after surgery. The 5 patients whose C.E.A. did not fall to below 5 ng. per ml. showed a subsequent rise in C.E.A. level and were all found later to have a tumour relapse. The results indicate that an incomplete drop of circulating C.E.A. level one month after surgery has a bad prognostic significance. 22 of these patients were followed up by repeated C.E.A. radioimmunoassay for several months after surgery. 8 showed a progressive increase in C.E.A. levels preceding clinical diagnosis of tumour relapse by two to ten months. 6 other patients showed a moderate increase in C.E.A. levels, suggesting a tumour relapse not yet clinically detectable. The remaining 8 patients showed no increase in C.E.A. level above 5 ng. per ml. and no clinical symptoms of relapse. The results demonstrate that relapses of colon and rectum carcinoma can be detected by increased C.E.A. levels months before the appearance of any clinical evidence.

Combining clinical factors and quantitative ultrasound improves the detection of women both at low and high risk for hip fracture.

We hypothesized that combining clinical risk factors (CRF) with the heel stiffness index (SI) measured via quantitative ultrasound (QUS) would improve the detection of women both at low and high risk for hip fracture. Categorizing women by risk score improved the specificity of detection to 42.4%, versus 33.8% using CRF alone and 38.4% using the SI alone. This combined CRF-SI score could be used wherever and whenever DXA is not readily accessible. INTRODUCTION AND HYPOTHESIS: Several strategies have been proposed to identify women at high risk for osteoporosis-related fractures; we wanted to investigate whether combining clinical risk factors (CRF) and heel QUS parameters could provide a more accurate tool to identify women at both low and high risk for hip fracture than either CRF or QUS alone. METHODS: We pooled two Caucasian cohorts, EPIDOS and SEMOF, into a large database named "EPISEM", in which 12,064 women, 70 to 100 years old, were analyzed. Amongst all the CRF available in EPISEM, we used only the ones which were statistically significant in a Cox multivariate model. Then, we constructed a risk score, by combining the QUS-derived heel stiffness index (SI) and the following seven CRF: patient age, body mass index (BMI), fracture history, fall history, diabetes history, chair-test results, and past estrogen treatment. RESULTS: Using the composite SI-CRF score, 42% of the women who did not report a hip fracture were found to be at low risk at baseline, and 57% of those who subsequently sustained a fracture were at high risk. Using the SI alone, corresponding percentages were 38% and 52%; using CRF alone, 34% and 53%. The number of subjects in the intermediate group was reduced from 5,400 (including 112 hip fractures) and 5,032 (including 111 hip fractures) to 4,549 (including 100 including fractures) for the CRF and QUS alone versus the combination score. CONCLUSIONS: Combining clinical risk factors to heel bone ultrasound appears to correctly identify more women at low risk for hip fracture than either the stiffness index or the CRF alone; it improves the detection of women both at low and high risk.

White mold detection in common beans through leaf reflectance spectroscopy

ABSTRACT This study aimed to identify wavelengths based on leaf reflectance (400-1050 nm) to estimate white mold severity in common beans at different seasons. Two experiments were carried out, one during fall and another in winter. Partial Least Squares (PLS) regression was used to establish a set of wavelengths that better estimates the disease severity at a specific date. Therefore, observations were previously divided in two sub-groups. The first one (calibration) was used for model building and the second subgroup for model testing. Error measurements and correlation between measured and predicted values of disease severity index were employed to provide the best wavelengths in both seasons. The average indexes of each experiment were of 5.8% and 7.4%, which is considered low. Spectral bands ranged between blue and green, green and red, and red and infrared, being most sensitive for disease estimation. Beyond the transition ranges, other spectral regions also presented wavelengths with potential to determine the disease severity, such as red, green, and near infrared.

Detection of enterotoxins produced by B. cereus through PCR analysis of ground and roasted coffee samples in Rio de Janeiro, Brazil

Coffee is one of the most appreciated drinks in the world. Coffee ground is obtained from the fruit of a small plant that belongs to the genus Coffea. Coffea arabica and Coffea canephora robusta are the two most commercially important species. They are more commonly known as arabica and robusta, respectively. Two-thirds of Coffea arabica plants are grown in South and Central America, and Eastern Africa - the place of origin for this coffee species. Contamination by microorganisms has been a major matter affecting coffee quality in Brazil, mainly due to the harvesting method adopted. Brazilian harvests are based on fruits collected from the ground mixed with those that fall on collection cloths. As the Bacillus cereus bacterium frequently uses the soil as its environmental reservoir, it is easily capable of becoming a contaminant. This study aimed to evaluate the contamination and potential of B. cereus enterotoxin genes encoding the HBL and NHE complexes, which were observed in strains of ground and roasted coffee samples sold in Rio de Janeiro. The PCR (Polymerase Chain Reaction) results revealed high potential of enterotoxin production in the samples. The method described by Speck (1984) was used for the isolation of contaminants. The investigation of the potential production of enterotoxins through isolates of the microorganism was performed using the B. cereus enterotoxin Reverse Passive Latex Agglutination test-kit (BCET-RPLA, Oxoid), according to the manufacturer's instructions. The potential of enterotoxin production was investigated using polymerase chain reaction (PCR) methods for hblA, hblD and hblC genes (encoding hemolysin HBL) and for nheA, nheB and nheC genes (encoding non-hemolytic enterotoxin - NHE). Of all the 17 strains, 100% were positive for at least 1 enterotoxin gene; 52.9% (9/17) were positive for the 3 genes encoding the HBL complex; 35.3% (6/17) were positive for the three NHE encoding genes; and 29.4% (5/17) were positive for all enterotoxic genes.

Evaluation of Machine Learning Classifiers for Mobile Network Intrusion Detection Systems

Mobile malwares are increasing with the growing number of Mobile users. Mobile malwares can perform several operations which lead to cybersecurity threats such as, stealing financial or personal information, installing malicious applications, sending premium SMS, creating backdoors, keylogging and crypto-ransomware attacks. Knowing the fact that there are many illegitimate Applications available on the App stores, most of the mobile users remain careless about the security of their Mobile devices and become the potential victim of these threats. Previous studies have shown that not every antivirus is capable of detecting all the threats; due to the fact that Mobile malwares use advance techniques to avoid detection. A Network-based IDS at the operator side will bring an extra layer of security to the subscribers and can detect many advanced threats by analyzing their traffic patterns. Machine Learning(ML) will provide the ability to these systems to detect unknown threats for which signatures are not yet known. This research is focused on the evaluation of Machine Learning classifiers in Network-based Intrusion detection systems for Mobile Networks. In this study, different techniques of Network-based intrusion detection with their advantages, disadvantages and state of the art in Hybrid solutions are discussed. Finally, a ML based NIDS is proposed which will work as a subsystem, to Network-based IDS deployed by Mobile Operators, that can help in detecting unknown threats and reducing false positives. In this research, several ML classifiers were implemented and evaluated. This study is focused on Android-based malwares, as Android is the most popular OS among users, hence most targeted by cyber criminals. Supervised ML algorithms based classifiers were built using the dataset which contained the labeled instances of relevant features. These features were extracted from the traffic generated by samples of several malware families and benign applications. These classifiers were able to detect malicious traffic patterns with the TPR upto 99.6% during Cross-validation test. Also, several experiments were conducted to detect unknown malware traffic and to detect false positives. These classifiers were able to detect unknown threats with the Accuracy of 97.5%. These classifiers could be integrated with current NIDS', which use signatures, statistical or knowledge-based techniques to detect malicious traffic. Technique to integrate the output from ML classifier with traditional NIDS is discussed and proposed for future work.

Buried Treasure: The Economics of Leak Detection and Water Loss Prevention in Ontario

On average approximately 13% of the water that is withdrawn by Canadian municipal water suppliers is lost before it reaches final users. This is an important topic for several reasons: water losses cost money, losses force water agencies to draw more water from lakes and streams thereby putting more stress on aquatic ecosystems, leaks reduce system reliability, leaks may contribute to future pipe failures, and leaks may allow contaminants to enter water systems thereby reducing water quality and threatening the health of water users. Some benefits of leak detection fall outside water agencies’ accounting purview (e.g. reduced health risks to households connected to public water supply systems) and, as a result, may not be considered adequately in water agency decision-making. Because of the regulatory environment in which Canadian water agencies operate, some of these benefits-especially those external to the agency or those that may accrue to the agency in future time periods- may not be fully counted when agencies decide on leak detection efforts. Our analysis suggests potential reforms to promote increased efforts for leak detection: adoption of a Canada-wide goal of universal water metering; development of full-cost accounting and, pricing for water supplies; and co-operation amongst the provinces to promulgate standards for leak detection efforts and provide incentives to promote improved efficiency and rational investment decision-making.

Fully electrochemical hyphenated flow system for preconcentration, cleanup, stripping, capillary electrophoresis with stacking and contactless conductivity detection of trace heavy metals

Successful coupling of electrochemical preconcentration (EPC) to capillary electrophoresis (CE) with contactless conductivity detection (C(4)D) is reported for the first time. The EPC-CE interface comprises a dual glassy carbon electrode (GCE) block, a spacer and an upper block with flow inlet and outlet, pseudo-reference electrode and a fitting for the CE silica column, consisting of an orifice perpendicular to the surface of a glassy carbon electrode with a bushing inside to ensure a tight press fit. The end of the capillary in contact with the GCE is slant polished, thus defining a reproducible distance from the electrode surface to the column bore. First results with EPC-CE-C(4)D are very promising, as revealed by enrichment factors of two orders of magnitude for Tl, Cu, Pb and Cd ion peak area signals. Detection limits for 10 min deposition time fall around 20 nmol L(-1) with linear calibration curves over a wide range. Besides preconcentration, easy matrix exchange between accumulation and stripping/injection favors procedures like sample cleanup and optimization of pH, ionic strength and complexing power. This was demonstrated for highly saline samples by using a low conductivity buffer for stripping/injection to improve separation and promote field-enhanced sample stacking during electromigration along the capillary. (C) 2010 Elsevier B.V. All rights reserved.