This Practical Work was made with the objective of attempt a work for the MsC of Electronics Enginheering in Polytechnic Institute of Cávado and Ave for the Digital Signal Processing subject
February 2022 Barcelos, Portugal
Professor of Digital Signal Processing (Nuno Dias)
Specific Objectives: Development of an algorithm to analyze a student's voice while saying the access/registration password in the classroom. This algorithm has 2 objectives:
(1) to identify the student saying the password.
(2) identify the password said by the student.
Programming language: Python. The various topics covered in class will be assessed in the algorithm developed by each group of students:
(1) Spectral representation of digital signals.
(2) Digital filters.
(3) Voice signal analysis for speaker identification and keywords.
Minimum requirements: Although the digital audio processing algorithm must run simultaneously (in online mode) with the data acquisition (see diagram below), for assessment purposes of the Digital Signal Processing of Digital Signal Processing, each group of students must demonstrate competencies in student identification and password in offline mode, on a dataset previously created by each group for this purpose. The solution to be developed must be able to identify words never heard before before heard, provided they are composed of the same sounds (phonemes) of a set of words to be provided. Each group of students should prepare a jupyter notebook with the demonstration of the developed algorithm, including:
● Code in Python
● Identification and explanation of the options taken, framing them in the various topics taught in class.
● Result of the transformations of the audio signals, in each step of the algorithm (graphical representations should be given preference to graphical representations).
Delivery:
- Jupyter notebook
- Dataset for testing the developed algorithm