GAN-based Photo Video Synthesis Summary of Generating Videos with - - PowerPoint PPT Presentation

▶

Apr 02, 2024 237 likes •321 views

GAN-based Photo Video Synthesis Summary of Generating Videos with Scene Dynamics Lei Zhang CS 297 Introduction Train with unlabeled video Extend GAN to video Introduce a two-stream generative model that split foreground from the

SLIDE 1

GAN-based Photo Video Synthesis

Summary of Generating Videos with Scene Dynamics Lei Zhang CS 297

SLIDE 2

Introduction

Train with unlabeled video
Extend GAN to video
Introduce a two-stream generative model that split foreground from the

background, which to learn move and non-move objects respectively

SLIDE 3

Discriminator

Able to classify realistic scenes from synthetically generated scenes
Able to recognize realistic motion between frames
It uses five layer spatio-temporal convolutional network

SLIDE 4

Two Stream Video GAN

SLIDE 5

Future Generation

Give a static image to extrapolate possible consequent

frames

To improve

○ Generate similar but not identical scenes ○ Only generate 1-2 seconds video

SLIDE 6

Future Generation with Plausible Motions

SLIDE 7

REFERENCE [1] Vondrick, Carl, Hamed Pirsiavash, and Antonio Torralba. "Generating videos with scene dynamics." Advances In Neural Information Processing Systems. 2016.