Amir Ghodrati

8

Papers

21

Total Citations

Papers (8)

Mobile Video Diffusion

Clockwork Diffusion: Efficient Generation With Model-Step Distillation

Modeling Video Evolution for Action Recognition

Actor and Action Video Segmentation From a Sentence

ActionBytes: Learning From Trimmed Videos to Localize Actions

FrameExit: Conditional Early Exiting for Efficient Video Recognition

DeepProposal: Hunting Objects by Cascading Deep Convolutional Layers

SALISA: Saliency-Based Input Sampling for Efficient Video Object Detection