Phillip Ha
Created September 9, 2020 © Apache-2.0

ML Model for Detecting & Classifying Elephant-Human Conflict

This machine learning model for detecting and classifying potential elephant-human conflict is great for the Elephant-Edge Tracker

AdvancedFull instructions provided2 days108
ML Model for Detecting & Classifying Elephant-Human Conflict

Things used in this project

Hardware components

Samsung smart mobile phone
For audio sound capture and running model classification test
×1

Software apps and online services

Edge Impulse Studio
Edge Impulse Studio
For importing train and test datasets For developing, training and build a new ML model For evaluating the model and test the audio classification
Windows 10
Microsoft Windows 10
To run the Edge Impulse studio using a web browser on a Windows PC To deploy and export the ML models for use in any embedded device To test the model and audio streams
Snappy Ubuntu Core
Snappy Ubuntu Core
Alternative Linux system PC or embedded device to test the model

Story

Read more

Schematics

NN Classification Model Training Parameter Settings

Necessary parameter settings for Neural Network Classification training block

ML Model Creating Diagram using Edge Impulse Studio

Setting diagram to create the DL Model

MFCC Parameter Settings

Necessary parameter settings for MFCC block

Code

Tensorflow Lite ML Model for Elephant Edge Tracker to classify elephant-normal-behavior and panic

Assembly x86
The TinyML Model for use in embedded device and with Tensorflow Lite framework, float32
No preview (download only).

Tensorflow Lite ML Model for Elephant Edge Tracker to classify elephant-normal-behavior and panic

Assembly x86
The TinyML Model for use in embedded device and with Tensorflow Lite framework, int8.
No preview (download only).

Tensorflow Lite ML Model for Elephant Edge Tracker to classify elephant-normal-behavior and panic

Assembly x86
The TinyML Model for use in embedded device and with Tensorflow Lite framework, int8-quantized-with-float32-input-and-output-mode
No preview (download only).

Tensorflow ML Model for Elephant Edge Tracker to classify elephant-normal-behavior and panic

Assembly x86
The TinyML Model for use in embedded device and with Tensorflow framework, saved_model.pb
No preview (download only).

Elephant-normal-behavior test audio file

Assembly x86
Test audio file for detection and classification of the elephant-normal-behavior
No preview (download only).

Elephant-panic test audio file

Assembly x86
Test audio file for detection and classification of the elephant-panic signature. It also shows the transition from normal behavior to panic when elephants go on defending situation including potential elephant-human conflict
No preview (download only).

Credits

Phillip Ha

Phillip Ha

3 projects • 3 followers
An Electronic and Computer Science Engineer with additional DL/AI Researcher. I have designed product hardware and test software.

Comments