CahyaPutera / Sound-Source-Localization

Sound Localization project, originally from Arief Ramdhan's Jakarta Machine Learning Talk with some adjustment. The goal is to Build MLP and CNN networks for Sound Source Localization.

Sound Source Localization - Demonstration

Project Definition

Build MLP and CNN network for Sound Source Localization.

Project Details

Input: binaural audio files

Output: labels

Classify 3 azimuth angles using binaural audio files as inputs. The azimuth angles are: 0, 30, and 60 degrees

Use MFCC as input features

Metrics: classification accuracy

The link to the dataset:

bit.ly/ssl_demo

The Process:

1. Mount google drive to acquire datset
2. Locate data and copy it to google colab
3. Import libraries
4. Load data
5. Build Neural Network
6. Train model
7. Test model
8. Evaluate model

Sources :

https://github.com/seth814/Audio-Classification

https://github.com/arief25ramadhan/Sound-Source-Localization

Credit goes to :

ARIEF RAMADHAN (https://github.com/arief25ramadhan)