README.md



Instructions for Training ResNet Inpainting Model, Saving Predictions, and Evaluating Accuracy
We assume you have activated the dav2 Conda environment.
Or, you can activate the terra-torch3d Conda environment, and install requirements for pytorch-msssim and lpips.
You should also have generated delayed reprojections using a given depth model by following instructions here.

Train Model

Update file names and valid timesteps to load for train, validation, and test splits in the dataset script here

Update path to pretrained VisionNavNet in the model file here

Run python train_resnet_inpaint.py with desired arguments to train an inpainting model


Collect Predictions

Save predictions from pretrained ResNet inpainting model:

Update path to trained checkpoint to load and names of videos to evaluate in save_resnet_inpaint.py

Run python save_resnet_inpaint.py for each video in dataset that you want to inpaint


Save predictions from classic Telea inpainting method:

Update names of videos to evaluate in save_telea_inpaint.py

Run python save_telea_inpaint.py for each video in dataset that you want to inpaint


Evaluate Accuracy

To compute quantitative accuracy of inpainted images

Run python eval_inpaint.py for each video in dataset that you want to evaluate