Wearable Sensing

In this paper, we develop an internet of medical things (IoMT)-based electrocardiogram(ECG) recorder for monitoring heart conditions in practical cases. To remove noise from signals recorded by these non-clinical devices, we propose a cloud-based denoising approach that utilizes deep neural network techniques in the time-frequency domain through the two stages. Accordingly, we exploit the fractional Stockwell transform (FrST) to transfer the ECG signal into the time-frequency domain and apply the deep robust two-stage network (DeepRTSNet) for the noise cancellation.


The data we are providing this time is a part of the dataset which was used in our previous work, titled “Integrating Activity Recognition and Nursing Care Records: The System, Deployment, and a Verification Study”. The authors of this work proposed a theory that extending of start and end times of the activities can increase the prediction rate. The reason behind the theory is that many of the nurses provided the labels before or after completing an activity. In the paper, they verified and proved this theory.

Last Updated On: 
Thu, 06/30/2022 - 01:03

The Baseline set described in the IEEE article (https://ieeexplore.ieee.org/document/10077565)   as Baseline_set  contains 1442450 rows, where the number of rows varied between 15395 and 197542 for the 16 subjects;  the average per subject being 69095 rows. The data set is filtered and standardized as described in III.C in the submission . The other data sets used in the article are derived from Baseline set.


This is the data for the paper "Fusion of Human Gaze and Machine Vision for Predicting Intended Locomotion Mode" published on IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2022. 


We collect IMU measurements under three different patterns: Fixing a smartphone in front of his chest (chest), swing a smartphone while holding it in his hand (swing), and putting a smartphone in his pocket (pocket). We use Google Pixel 3XL for the pattern of chest and Google Pixel 3a for the patterns of swing and pocket. The sampling frequency of each measurement is fixed to 15Hz. We collect the measurement of 111 paths in total, categorized into 4 types. We partition them into 84 and 27 paths, used for training and testing, respectively. It takes 10 hours to collect all datasets.


Visual perception can improve transitions between different locomotion mode controllers (e.g., level-ground walking to stair ascent) by sensing the walking environment prior to physical interactions. Here we developed the "StairNet" dataset to support the development of vision-based stair recognition systems. The dataset builds on ExoNet – the largest open-source dataset of egocentric images of real-world walking environments.


Context awareness is an emerging field in pervasive computing with applications that have started to emerge in medical systems. The present work seeks to determine which contexts are important for medical applications and what various domains of context aware applications exist in healthcare. Methods: A systematic scoping review of context aware medical systems currently being used in healthcare settings was conducted.


This dataset is related to dog activity and is sensor data.



There exist several commonly used datasets in relation to object detection that include COCO (with multiple versions) and ImageNet containing large annotations for 80 and 1000 objects (i.e. classes) respectively. However, very limited datasets are available comprising specific objects identified by visually imapeired people (VIP) such as wheel-bins, trash-Bags, e-Scooters, advertising boards, and bollard. Furthermore, the annotations for these objects are not available in existing sources.