generate augmented dataset by applying augmentation in 3d space

alr-internship / self-supervised-depth-denoising

Denoising YCB Objects with a self-supervised deep neural network

MIT License

9 stars 0 forks source link

generate augmented dataset by applying augmentation in 3d space #3

Closed claudius-kienle closed 2 years ago

claudius-kienle commented 2 years ago

augmentation adds new data samples, the network can train on. Therefore, augmentation will improve the generalization of a trained network.

claudius-kienle commented 2 years ago

Augmentation script added.

Script located in data_processing/augment_in_3d.py The script generates N random augmentations of each data tuple. The augmentation contains a random rotation with max rotation values (default 5 degreees) and a translation in m. The augmented data tuples get saved in the given out dir, concatenated with an index.