1. Social LSTM: Human Trajectory Prediction in Crowded Spaces
Alahi, A., Goel, K., Ramanathan, V., Robicquet, A., Fei-Fei, L., & Savarese, S. (2016). Social lstm: Human trajectory prediction in crowded spaces. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 961-971).
1. Introduction
- 对周围环境的语义理解对预测有很大帮助
- 建立人-人之间的相互作用模型能提高多物体追踪的鲁棒性和精确度(social forces)
影响上述工作的两个假设:
- 对于物体相互之间的作用是人为手动建模,不是通过数据驱动
- 预测时间短
LSTM的优缺点:
LSTM对不同的序列预测任务很成功:handwriting and speech generation
优:能预测长序列
缺:不能得到序列直接的相互关系
2. Model
建立模型的目的:
which can account for the behaviour of other people within a large neighbourhood, while predicting a person’s path.
2.1 Social LSTM
空间相邻的两个LSTM之间用池化层来相互分享信息。
遇到的问题:
拥挤场景中每个人相邻的人数量不一样—->解决办法,”Social“ pooling layers
评论加载中