Adding data to images in order to improve accuracy of CNN model

Part 1 (2020)
1

Hello,

I’m relatively experienced in deep learning and have used FastAI’s Fastbook quite a bit to train my models.

I’m at the stage where I would like to further my understanding and I’m looking to apply data (GPS and survey data) to my images so that the model is more accurate.

Whilst I’ve read about this in several papers, I don’t have any idea how to actually implement this. Searches online have proved fruitless.

Does anybody have any information and guidance, please?

Many thanks.

1 Like
2

Are you talking about multi-modal data, where each data point is both an image but some extra information associated with that image?