Human Action Anticipation: A Survey

Bolin Lai,Sam Toyer,Tushar Nagarajan,Rohit Girdhar,Shengxin Zha,James M. Rehg,Kris Kitani,Kristen Grauman,Ruta Desai,Miao Liu

2024-10-18

Abstract:Predicting future human behavior is an increasingly popular topic in computer vision, driven by the interest in applications such as autonomous vehicles, digital assistants and human-robot interactions. The literature on behavior prediction spans various tasks, including action anticipation, activity forecasting, intent prediction, goal prediction, and so on. Our survey aims to tie together this fragmented literature, covering recent technical innovations as well as the development of new large-scale datasets for model training and evaluation. We also summarize the widely-used metrics for different tasks and provide a comprehensive performance comparison of existing approaches on eleven action anticipation datasets. This survey serves as not only a reference for contemporary methodologies in action anticipation, but also a guideline for future research direction of this evolving landscape.

Computer Vision and Pattern Recognition,Machine Learning

What problem does this paper attempt to address?

The problem this paper attempts to address is the prediction of human future behavior, which is a widely focused issue in the field of computer vision. Specifically, the paper focuses on different tasks of behavior prediction, including action anticipation, activity forecasting, intent prediction, and goal prediction.

The main objectives of the paper are:
1. **Integrate fragmented literature**: To consolidate research on human behavior prediction scattered across different fields and provide a comprehensive review.
2. **Summarize existing technologies**: To cover recent technological innovations and the development of new large-scale datasets used for model training and evaluation.
3. **Evaluate existing methods**: To summarize widely used evaluation metrics and comprehensively compare the performance of existing methods on 11 action anticipation datasets.
4. **Guide future research directions**: To serve not only as a reference for contemporary methods but also to provide guidelines for future research directions in this evolving field.

Through these objectives, the paper aims to provide researchers and practitioners with a systematic framework to understand and advance the research on human behavior prediction.