- Context-Aware Activity Recognition Systems
- Interactive and Immersive Displays
- Embedded Systems Design Techniques
- Green IT and Sustainability
- Stroke Rehabilitation and Recovery
- Personal Information Management and User Behavior
- Usability and User Interface Design
- Gaze Tracking and Assistive Technology
- VLSI and FPGA Design Techniques
- IoT and Edge/Fog Computing
- Retinal Imaging and Analysis
- Augmented Reality Applications
- Tactile and Sensory Interactions
- Parallel Computing and Optimization Techniques
- Speech and Audio Processing
- COVID-19 diagnosis using AI
- Human Pose and Action Recognition
- Innovative Human-Technology Interaction
- Mobile Health and mHealth Applications
- Manufacturing Process and Optimization
- Hand Gesture Recognition Systems
- EEG and Brain-Computer Interfaces
- Human Mobility and Location-Based Analysis
- Innovative Teaching and Learning Methods
- Design Education and Practice
Carnegie Mellon University
2014-2023
Madeira Tecnopolo
2020
University of Geneva
2020
Agencia Regional da Energia e Ambiente da Regiao Autonoma da Madeira
2020
Universidade da Madeira
2020
University of Lisbon
2020
Active Technologies (Italy)
2020
Instituto de Tecnologias Interativas
2020
Northwestern University
2017
National Institutes of Health
2011
The design of an activity recognition and monitoring system based on the eWatch, multi-sensor platform worn different body positions, is presented in this paper. identifies user's realtime using multiple sensors records classification results during a day. We compare time domain feature sets sampling rates, analyze tradeoff between accuracy computational complexity. positions used for wearing electronic devices was evaluated
Ecological momentary assessment (EMA) assesses individuals' current experiences, behaviors, and moods as they occur in real time their natural environment. EMA studies, particularly those of longer duration, are complex require an infrastructure to support the data flow monitoring completion.Our objective is provide a practical guide developing implementing study, with focus on methods logistics conducting such study.The EMPOWER study was 12-month that used examine triggers lapses relapse...
SenSay is a context-aware mobile phone that adapts to dynamically changing environmental and physiological states. In addition manipulating ringer volume, vibration, alerts, can provide remote callers with the ability communicate urgency of their calls, make call suggestions users when they are idle, caller feedback on current status user. A number sensors including accelerometers, light, microphones mounted at various points body data about user’s context. decision module uses set rules...
Context-aware computing describes the situation where a wearable/mobile computer is aware of its user's state and surroundings modifies behavior based on this information. We designed, implemented, evaluated wearable system which can learn context-dependent personal preferences by identifying individual user states observing how interacts with in these states. This learning occurs online does not require external supervision. The relies techniques from machine statistical analysis. A case...
Advances in artificial intelligence (AI) have made it increasingly applicable to supplement expert's decision-making the form of a decision support system on various tasks. For instance, an AI-based can provide therapists quantitative analysis patient's status improve practices rehabilitation assessment. However, there is limited knowledge potential these systems. In this paper, we present development and evaluation interactive that supports collaborative making with for This automatically...
Diabetic retinopathy (DR) detection is a critical retinal image analysis task in the context of early blindness prevention. Unfortunately, order to train model accurately detect DR based on presence different lesions, typically dataset with medical expert's annotations at pixel level needed. In this paper, new methodology multiple instance learning (MIL) framework developed overcome necessity by leveraging implicit information present made level. Contrary previous MIL-based systems, main...
Context-aware computing describes the situationwhere a wearable / mobile computer is aware of itsuser's state and surroundings modifies its behaviorbased on this information. We designed, implemented andevaluated system which can determine typicaluser context transition probabilities onlineand without external supervision. The relies ontechniques from machine learning, statistical analysisand graph algorithms. It be used for onlineclassification prediction. Our results indicate thepower our...
The eWatch is a wearable sensing, notification, and computing platform built into wrist watch form factor making it highly available, instantly viewable, ideally located for sensors, unobtrusive to users. Bluetooth communication provides wireless link cellular phone or stationary computer. senses light, motion, audio, temperature visual, tactile notification. system ample processing capabilities with multiple day battery life enabling realistic user studies. This paper the motivation...
Context-aware mobile computing requires wearable sensors to acquire information about the user. Continuous sensing rapidly depletes -wearable system's energy, which is a critically constrained resource. In this paper, we analyze trade-off between power consumption and prediction accuracy of context classifiers working on dual-axis accelerometer data collected from eWaich notification platform. We improve techniques by providing competitive classification performance even in low frequency...
Self-monitoring (SM) of food intake is central to weight loss treatment. Technology makes it possible reinforce this behavior change strategy by providing real-time feedback (FB) tailored the diary entry. To test feasibility 1-4 daily FB messages dietary recordings via a smartphone, we conducted 12-week pilot randomized clinical trial in Pittsburgh, PA US 2015. We compared 3 groups: SM using Lose It! smartphone app (Group 1); + 2); and attending three in-person group sessions 3). The sample...
Though tactile maps have been shown to be useful tools for visually impaired individuals, their availability has limited by manufacturing and design costs. In this paper, we present a system that uses 3D printing (1) make more affordable produce, (2) allow individuals independently customize maps, (3) provide interactivity using widely available mobile devices. Our consists of three parts: web interface, modeling algorithm, an interactive touchscreen application. hosted at...
Artificial intelligence (AI) and robotic coaches promise the improved engagement of patients on rehabilitation exercises through social interaction. While previous work explored potential automatically monitoring for AI coaches, deployment these systems remains a challenge. Previous described lack involving stakeholders to design such functionalities as one major causes. In this paper, we present our efforts eliciting detailed specifications how could interact with guide patient's in an...
Deep learning models have been successfully used in medical image analysis problems but they require a large amount of labeled images to obtain good performance. However, such datasets are costly acquire. Active techniques can be minimize the number required training labels while maximizing model's In this work, we propose novel sampling method that queries unlabeled examples maximize average distance all set learned feature space. We then extend our define better initial set, without need...
Due to the limited number of therapists, task-oriented exercises are often prescribed for post-stroke survivors as in-home rehabilitation. During rehabilitation, a patient may become unmotivated or confused comply prescriptions without feedback therapist. To address this challenge, paper proposes an automated method that can achieve not only qualitative, but also quantitative assessment stroke rehabilitation exercises. Specifically, we explored threshold model utilizes outputs binary...
Article Free Access Share on The design of a wearable computer Authors: Len Bass Carnegie Mellon University, Pittsburgh, Pa PaView Profile , Chris Kasabach Richard Martin Dan Siewiorek Asim Smailagic John Stivoric Authors Info & Claims CHI '97: Proceedings the ACM SIGCHI Conference Human factors in computing systemsMarch 1997 Pages 139–146https://doi.org/10.1145/258549.258634Online:27 March 1997Publication History 44citation1,332DownloadsMetricsTotal Citations44Total Downloads1,332Last 12...
Embedded systems encompass a wide range of applications, technologies, and disciplines, necessitating broad approach to education. We describe embedded system coursework during the first 4 years university education (the U.S. undergraduate level). application curriculum areas include: small single-microcontroller control systems, distributed control, system-on-chip, networking, PCs, critical robotics, computer peripherals, wireless data signal processing, command control. Additional...
Clinical decision support systems have the potential to improve work flows of experts in practice (e.g. therapist's evidence-based rehabilitation assessment). However, adoption these is challenging, and gains not fully demonstrated yet. In this paper, we identified needs therapists assess patient's functional abilities alternative perspectives with quantitative information on exercise motions). As a result, co-designed developed an intelligent system that automatically identifies salient...
article Free AccessRapid design and manufacture of wearable computers Authors: S. Finger Department Civil Environmental Engineering, Carnegie Mellon MellonView Profile , M. Terk Engineering Design Research Center, E. Subrahamanian C. Kasabach F. Prinz Departments Mechanical Material Science Stanford University UniversityView D. P. Siewiorek Electrical Computer Director the A. Smailagic J. Stivoric L. Weiss Robotics Institute, Authors Info & Claims Communications ACMVolume 39Issue 201...
The paper describes the mobile information and communication aspects of a next generation train maintenance diagnosis system, discusses working prototype features, research results. Wearable/Mobile computers combined with wireless technology improve efficiency accuracy work. This enables personnel at site to communicate remote helpdesk/expertise center through digital data, audio, image.
We propose DiaWear, a novel assistive mobile phone-based calorie monitoring system to improve the quality of life diabetes patients and individuals with unique nutrition management needs. Our goal is achieve improved daily semi-automatic food recognition using wearable cell phone. DiaWear currently uses neural network classification scheme identify items from captured image. It difficult account for varying implicit nature certain foods traditional image techniques. To overcome these...
Diabetic Retinopathy (DR) is one of the leading causes preventable blindness in developed world. With increasing number diabetic patients there a growing need an automated system for DR detection. We propose Eye WeS, method that not only detects eye fundus images but also pinpoints regions image contain lesions, while being trained with labels only. show it possible to convert any pre-trained convolutional neural network into weakly-supervised model their performance and efficiency. EyeWeS...