Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/mohammedyunus009/dnn_acoustic_rec
A acoustic sound or environmental sound recogniser, uses deep neural networks to train on models
https://github.com/mohammedyunus009/dnn_acoustic_rec
deep-learning deep-neural-networks extract-features keras sound sound-processing sound-synthesis soundcloud-api tensorflow
Last synced: 12 days ago
JSON representation
A acoustic sound or environmental sound recogniser, uses deep neural networks to train on models
- Host: GitHub
- URL: https://github.com/mohammedyunus009/dnn_acoustic_rec
- Owner: mohammedyunus009
- License: apache-2.0
- Created: 2018-03-30T10:45:27.000Z (almost 7 years ago)
- Default Branch: master
- Last Pushed: 2018-05-22T19:26:19.000Z (over 6 years ago)
- Last Synced: 2024-11-02T00:42:09.377Z (2 months ago)
- Topics: deep-learning, deep-neural-networks, extract-features, keras, sound, sound-processing, sound-synthesis, soundcloud-api, tensorflow
- Language: Python
- Homepage:
- Size: 27.3 KB
- Stars: 3
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Introduction (machine learning ,sound recognition)
This is a keras implementation project
This project is used to recognize environmental sounds ,with a high accuracy of 80% percent ,it has a deep neural network (LSTM can be implemented),can be used to train on the extracted features of audio files in
[https://github.com/mohammedyunus009/dev_datasets](https://github.com/mohammedyunus009/dev_datasets) and [https://github.com/mohammedyunus009/datasets](https://github.com/mohammedyunus009/datasets).It has the capability to recognize on 15 diffrent classes of sounds such as :
'bus' 'cafe/restaurant' 'car' 'city_center' 'forest_path'
'grocery_store' 'home' 'beach' 'library' 'metro_station'
'office' 'residential_area' 'train' 'tram' 'park'vtu under graduate project (visvervaraya university of technology)
# Requirements
This library runs with keras.
`pip install -r requirements.txt`
OR for conda and linux users run
`sh setup.sh`# Quickstart
STEP 1
configure the config file in `src` folderSTEP 2
*`python calculate_logmel.py` to extract features and pickle in memory
STEP 3
* `python kera_model.py --dev_train` to run in development mode (train on development dataset)
* `python kera_model.py --eva_train` to run in evaluation mode (train on evaluation dataset)
Reconfigure the configuration file in src* `python kera_model.py --dev_recognize` used to calculate the accuracy of the model in development mode
* `python kera_model.py --eva_recognize` used to calculate the accuracy of the model in evaluation mode
STEP 4
* `python session.py` used to put in production and test new filesContact me for more information