Multi-DPP

Multi-stream Dynamic Video Summarization with an extended Determinantal point process in supervised and unsupervised settings https://arxiv.org/abs/1812.00108

Data

Uniform Sequences: Sequences that are extracted by downsampling the original framerate of the cameras from 30fps to 1 fps, for all the sequences.

Mean Sequences: Sequences that are extracted by calculating the mean of 30 frame every second.

Uniform Sequences- GoogLeNet Features: The featrures are calculated by extracting the GoogLeNet features of the Uniform Sequences.

Videos: Original Videos of the dataset for three collections

To-Do

Clean the code and upload it.