Building the dataset for indoor scene classification
Now that we can capture frames from the camera, it is time to create the dataset for classifying indoor environments.
In this recipe, we will construct the dataset by collecting the kitchen and bathroom images with the OV7670 camera.
The following Python script contains the code referred to in this recipe:
04_build_dataset.py:
Getting ready
Training a deep neural network from scratch for image classification commonly requires a dataset with 1,000 images per class. As you might guess, this solution is impractical for us since collecting thousands of pictures takes a lot of time.
Therefore, we will consider an alternative ML technique: transfer learning.
Transfer learning is a popular method that uses a pre-trained model to train a deep neural network with a small dataset. This ML technique will...
