I’m building a model that does emotion recognition ( give me a picture, i’ll tell you what is the emotion felt by the person in that picture).
I want to do some data augmentation on my dataset using DCGAN.
I’m using code from the tutorial.
I have 7 different classes. I’ll have to go up to 22 classes eventually.
My dataset is structured like this:
dataset/
├── happy/
│ ├── image1.png
│ ├── image2.png
│ ├── image3.png
│ └── image4.png
├── sad/
│ ├── image1.png
│ ├── image2.png
│ ├── image3.png
│ └── image4.png
…
It properly imports all the images from the different folders.
I just can’t find a way to retrieve the labels of the generated images.
The only alternative I have for now is running the model for each emotion separately.
If anyone knows the proper way, please let me know.
Thanks.