You are on page 1of 10

GUITAR CHORDS

SOUND
CLASSIFICATION 1
OBJECTIVE
To Classify Guitar Chords, Human Speech, Clap and Background Noise, which will
be the first step to understand the currently using Algorithms like song
recommendation.
To learn opensource software like Google Teachable Machine, p5.js, ml5.js.

2
MOTIVATION
Although GTZAN Dataset is available for Music Classification, it will involve transfer
learning and feature extraction for model training.
I wanted to work on Google’s Teachable Machine who takes care of above mentioned
part and just allow you to create model by giving input datasets.
I also have Guitar and wanted to use it for any project so I think this the best
opportunity to make best use of it. It will also help to look into how Song
recommendation Algorithm works.
As for JavaScript, currently node.js is in trending but to learn the commands for that, I
need to first master the JavaScript.
And p5.js web browser is beginners friendly and also works as a server for all of your
codes.
3
LITERATURE REVIEW
Speechcommands18w
Pretrained model for Sound Classification used in Google’s Teachable Machine.
Focuses on Key Word Spotting. [Ex. Ok Google!].
General Sound Classification
Used for computing similarity metrics between a target sound and other sounds in a database.
Use of hidden Markov models for representing audio spectra.
Environmental Sound Classification
3 Datasets were used - ESC-50, ESC-10 and UrbanSound8K
Feature Extraction using Transfer learning is used.
CNN Architecture is involved
4
LITERATURE REVIEW
Detecting Illegal Deforestation Activities
Detecting Sound Of Humpback Whales

5
OPENSOURCE SOFTWARE
p5.js Web browser
Google Teachable Machine
ml5.js library

6
APPLICATION
Music Classification
Classifying music based on factors such as genre or instruments played.
Key role in organizing audio libraries by genre, improving recommendation
algorithms, and discovering trends and listener preferences through data analysis.
The GTZAN genre collection dataset is available. It consists of 1000 audio files each
having 30 seconds duration. There are 10 classes ( 10 music genres) each containing
100 audio tracks.

7
OUTCOME
To learn JavaScript and how to use Teachable Machine.
I got to know how “Call back” function & “Arrow” functions are used.
Also got to know about the pretrained model that is used for model training in Audio
Classification.

8
REFERENCES
https://ml5js.org/reference/api-soundClassifier/
https://editor.p5js.org/
https://apple.github.io/turicreate/docs/userguide/sound_classifier/
Speech Commands: A Dataset for Limited-Vocabulary Speech Recognition
General Sound Classification - Michael Casey, MERL Cambridge Research
Laboratory
Environmental Sound Classification With Convolutional Neural Networks - Karol J.
Piczak, Institute of Electronic Systems Warsaw University of Technology

9
THANK YOU!!

10

You might also like