Edge Impulse Inc. / Performance Calibration: Bird sound classifier Public

Edge Impulse Inc. / Performance Calibration: Bird sound classifier

Classifies audio as representative of either the house sparrow, rose-ringed parakeet, or background noise.

Audio

About this project

Have you ever wondered how to use your Edge Impulse project's Performance calibration feature to optimize your audio detection models? Performance calibration allows you to test, fine-tune, and simulate running your model with continuous real-world or synthetically generated audio data streams to gain an immediate understanding of how your model will perform in the field. Clone this project to build an embedded ML project to detect various bird calls in your environment from your device's microphone input!

Sensor & Block Information

  • Microphone audio data (.wav files) @ 16000Hz
  • MFCC DSP block for non-human voice audio
  • Neural Network Classifier with prediction outputs: "housesparrow", "roseringedparakeet", "noise"

bird.jpeg

Creating your first impulse (100% complete)

Acquire data

Every Machine Learning project starts with data. You can capture data from a development board or your phone, or import data you already collected.

Design an impulse

Teach the model to interpret previously unseen data, based on historical data. Use this to categorize new data, or to find anomalies in sensor readings.

Deploy

Package the complete impulse up, from signal processing code to trained model, and deploy it on your device. This ensures that the impulse runs with low latency and without requiring a network connection.

Download block output

Title Type Size
MFCC training data NPY file 4394 windows
MFCC training labels NPY file 4394 windows
MFCC testing data NPY file 10521 windows
MFCC testing labels NPY file 10521 windows
NN Classifier model (version #1) TensorFlow Lite (float32) 27 KB
NN Classifier model (version #1) TensorFlow Lite (int8 quantized) 11 KB
NN Classifier model (version #1) TensorFlow Lite (int8 quantized with float32 input and output) 11 KB
NN Classifier model (version #1) TensorFlow SavedModel 36 KB
NN Classifier model (version #2) TensorFlow Lite (float32) 13 KB
NN Classifier model (version #2) TensorFlow Lite (int8 quantized) 10 KB
NN Classifier model (version #2) TensorFlow Lite (int8 quantized with float32 input and output) 10 KB
NN Classifier model (version #2) TensorFlow SavedModel 23 KB
NN Classifier model (version #3) TensorFlow Lite (float32) 13 KB
NN Classifier model (version #3) TensorFlow Lite (int8 quantized) 10 KB
NN Classifier model (version #3) TensorFlow Lite (int8 quantized with float32 input and output) 10 KB
NN Classifier model (version #3) TensorFlow SavedModel 24 KB

Clone project

You are viewing a public Edge Impulse project. Clone this project to add data or make changes.

Summary

Data collected
2h 33m 47s

Project info

Project ID 134182
Project version 3
License No license attached