A simple example of a four-class image classifier using a small dataset (320 images of flowers: 80 sample x 4 categories) and a very simple CNN, with and without data augmentation.
The main goal of this example is to demonstrate the use of the MATLAB functionality for data augmentation in image classification solutions: the augmentedImageDatastore and the imageDataAugmenter.
This example should be easy to modify and expand to the user's needs.
- The validation accuracy improves -- from ~79% (Part 1 in the code) to ~83% (Part 2) -- using a very simple CNN, as a result of data augmentation alone.
- Interestingly enough, using a pretrained AlexNet, the validation accuracy drops -- from 100% (Part 3) to ~98% (Part 4) -- which shows that data augmentation wouldn't be necessary in this case.
Oge Marques (2022). Image classification using data augmentation (https://www.mathworks.com/matlabcentral/fileexchange/68728-image-classification-using-data-augmentation), MATLAB Central File Exchange. Retrieved .
MATLAB Release Compatibility
Platform CompatibilityWindows macOS Linux
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!Start Hunting!