# Predicting 3D Human Dynamics from Video

## TL;DR

This paper introduces a novel method for predicting future 3D human mesh sequences from past video frames, applicable to various actions, using an autoregressive model trained on in-the-wild videos without 3D labels.

## Contribution

It is the first approach to predict 3D human motion sequences directly from 2D video input without requiring 3D ground truth annotations.

## Key findings

- Effective prediction of 3D human motion from 2D videos.
- Applicable to diverse actions like walking, bowling, and squatting.
- Operates without 3D labeled training data.

## Abstract

Given a video of a person in action, we can easily guess the 3D future motion of the person. In this work, we present perhaps the first approach for predicting a future 3D mesh model sequence of a person from past video input. We do this for periodic motions such as walking and also actions like bowling and squatting seen in sports or workout videos. While there has been a surge of future prediction problems in computer vision, most approaches predict 3D future from 3D past or 2D future from 2D past inputs. In this work, we focus on the problem of predicting 3D future motion from past image sequences, which has a plethora of practical applications in autonomous systems that must operate safely around people from visual inputs. Inspired by the success of autoregressive models in language modeling tasks, we learn an intermediate latent space on which we predict the future. This effectively facilitates autoregressive predictions when the input differs from the output domain. Our approach can be trained on video sequences obtained in-the-wild without 3D ground truth labels. The project website with videos can be found at https://jasonyzhang.com/phd.