MAIN RESEARCH INTERESTS
- Gesture recognition based on sEMG signals and Deep Learning
- Signal, Image and Video Processing / Analysis
- Image and Video Coding
- HDR Image Compression
- Data Hiding in Images and Video
- Fast Transform Algorithms; Time / Space Optimisation of Computationally Intensive Algorithms
- Digital Signal Processors and Microprocessors; Real-time implementation of DSP algorithms
The MyoUP (Myo University of Patras) database contains recordings from 8 intact subjects (3 females, 5 males; 1 left handed, 7 right handed; age 22.38 ± 1.06 years). The acquisition process was divided into three parts: 5 basic finger movements, 12 isotonic and isometric hand configurations and 5 grasping hand-gestures. The recording device used was the Myo Armband by Thalmic labs (8 dry sEMG channels and sampling frequency of 200Hz).
The sEMG dataset is available for free download here.
Reference ENF Recordings for Greece
Reference Electrical Network Frequency (ENF) signals from the Greek main power grid are recorded 24/7 at the Digital Signal and Image Processing Lab, ECE Dept, University of Patras, as of July 13, 2016.
Everyday 24 recordings (wav files), i.e. one wav file per hour, are produced and zipped in one file of approx. 150 MB. In the end of the day, the zip file is uploaded to the ftp server and is ready for download.
Reference (ground-truth) ENF signals are extracted from power grid recordings, measured by a signal recorder that is connected to a power outlet using a step-down transformer. Τhe signal is routed to the sound card and is recorded using the appropriate software. A sampling rate of 1 kHz is used.
The Matlab code for the extraction of the ENF signal from the power grid recording is available for free download here.
The recordings of the main power grid of Greece are available for download from http://haradros.ece.upatras.gr/ under the directory ws-ENF-Recordigs-Greece/. In the same directory, The ENF-extract.zip file contains the MATLAB code for the extraction of the ENF signals of the power grid recordings.
To access the dataset, please send an Email to email@example.com and you will receive the login and password info. Please provide your affiliation with your Email request (for statistical purposes).