Image classification (persons, animals, other) on raspberry pi from pi-camera (in live time) using custom model tflite (output to terminal), using TensorFlow Lite
Short summary:
In this article, I will explain how to create a solution for image classification for the classes (persons, animals, others) on Raspberry Pi from Pi-camera in live time.
I will use a custom model tflite which I converted from h5.
And all results will be shown in the terminal. Code for this article is available here.
Note before you start:
So, let's start :)
Hardware preparation:
Software preparation:
1. Create a neural network model to predict 3 classes: persons, animals, others.
I have already done it in the article below. So in this article, you do not need to do this.
2 Convert h5 model to tflite model
I have already done it in the article below. So in this article, you do not need to do this.
3 So, let’s try to predict class from camera raspberry pi in live time with output to the terminal using Tensorflow Lite and custom model tflite (which I trained and saved above).
For that, you need to run the next code:
#clone my repo
git clone https://github.com/oleksandr-g-rock/image_classification_pi_camera_live_time_custom_model_tflite.git#go to folder
cd image_classification_pi_camera_live_time_custom_model_tflite#download tflite model
wget https://github.com/oleksandr-g-rock/how_to_convert_h5_model_to_tflite/raw/main/animall_person_other_v2_fine_tuned.tflite#run script
python3 classify_picamera_with_live_time_custom_tflite_model.py
4 You should see something like that.
Result:
In this article, we created a solution for image classification for the classes (persons, animals, others) on Raspberry Pi from Pi-camera in live time.
I will use a custom model tflite which I converted from h5.
And all results will be shown in the terminal. Code for this article is available here.