Gesture/ Posture Recognition Using Spectrogram
Problem Statement:
Using millimeter-wave frequency modulated continuous wave radars (FMCW), we can monitor different activities, gestures, and postures. FMCW radar use both transmitted and received signals to let us see through the electromagnetic waves. Using the generated signal by radar, we can form a spectrogram specifically for each activity. Using them, we can estimate the ongoing posture, gesture, or activity.
Task:
1. Raw data collection ( the student will be assisted in this phase by the supervisor).
2. Forming the spectrogram.
3. Using Machin Learning and Deep Learning methods for gesture/ posture recognition.
Note: You are supposed to know how to write codes in Python language. Moreover, you should be acquainted with Machine Learning and Deep Learning methods.
Work:
%20 Theory, 40% Data collection, 30% Simulations, %10 Writing
Contact:
Ehsan Sam Sadeghi (E.sadeghi@utwente.nl)