VIBE: Video Inference for Human Body Pose

and Shape Estimation

CVPR 2020

June 5, 2023

Presenter, Jeongwan On

0. Introduction

• Tremendous progress has been made on estimating 3D human pose and shape from a single image.

​

• While this is useful for many applications, it is the motion of the body in the world that tells us about human behavior.

​

• While this problem has received over 30 years of study, we may ask why reliable methods are still not readily available

1. Lack of 2D-3D mapping annotations
2. Lack of in-the-wild datasets

0. Introduction

1. Architecture

2. Generator

2. Generator – Feature extractor(CNN)

2. Generator – Temporal Encdoer (GRU)

2. Generator - Regressor

3. Discriminator

3. Discriminator

4. Loss - Discriminator

4. Loss – Generator

5. Experiments

5. Experiments

5. Experiments

5. Experiments

5. Experiments

Thank you!