A Simple Recipe for Contrastively Pre-training Video-First Encoders Beyond 16 Frames.
Pinelopi PapalampidiSkanda KoppulaShreya PathakJustin ChiuJoseph HeywardViorica PatrauceanJiajun ShenAntoine MiechAndrew ZissermanAida NematzadehPublished in: CoRR (2023)
Keyphrases
- video frames
- key frames
- video content
- video data
- video sequences
- real time
- computer vision
- training set
- spatio temporal
- compressed video
- successive frames
- space time
- neural network
- real time video
- temporal coherence
- single frame
- video images
- temporal filtering
- video processing
- video shots
- reference frame
- training process
- inter frame
- frame rate
- spatial and temporal
- training samples