Implement Video Detection

[blake-espeland]

If we're to make the 90+% accuracy, we'll need to utilize the information available in the temporal dimension. This is likely a large, sweeping modification to the repository that should only be carried out once we're very familiar with how the repo works.

[blake-espeland]

Sequential data loading implemented in 53ea7c56d0d562ab406a83816cbca3e7fbf55ce6

[blake-espeland]

The following still needs to be completed (for nt > 1):

• [ ] Displaying sequences (plot.py)
• [x] Performing the same data augmentation on each image of a sequence
• [ ] Sequential mosaic

These changes will be put on the backlog, as the necessary changes for TSM are already implemented. Will tag as a future item for now.

[blake-espeland]

Leaving this open until future issues/feature requests come up.

[blake-espeland]

13