[关闭]
@ArrowLLL 2017-10-23T03:51:51.000000Z 字数 8402 阅读 1316

learning note : Understanding Pedestrian Behaviors from Stationary Crowd Groups

OPTIMAL


原文 :Understanding Pedestrian Behaviors From Stationary Crowd Groups


文章主要有两点创新。

一是建立了一种模型 general energy map , 代表行人会走到这个地方的难易程度, 或者说一个行人会出现在场景中一个位置的概率。
表示这个位置; 是需要从data中学习到的参数.

模型主要由3个部分组成 :

通过场景中的K个观察点(observation of )建立似然函数使用梯度下降法更新参数

应用主要在于

静态人群的检测方法来源于参考文章 [39],笔记最后detecte stationary crowd group region的链接即是。
路径生成的方法主要来源于参考的两篇文章标号[17, 32],在笔记最后generate pedestrian walking path的两篇链接即是。

第二个创新点是建立了一个新的pedestrian walking path dataset,在一个video中标注了约12000人行为信息。


contribution

  1. A novel model is proposed form pedestrian behavior modeling by including stationary crowd groups as a key component.
    Through inference based on the interactions between stationary crowd groups and pedestrians, our model can be used to investigate pedestrian behaviors.

  2. A large pedestrian walking path dataset is built.
    The walking routes of more than 12, 000 pedestrians from a one-hour crowd video are annotated.

  3. The effectiveness of the proposed model is demonstrated by multiple applications on the proposed dataset. Including

    • pedestrian walking path prediction
    • pedestrian destination prediction
    • pedestrian personality estimation and classification
    • abnormal event detection
  4. compare with existing agent based models, three points :

    • the factor of stationary crowd groups is introduced for the first time to model pedestrian behavior.
    • The proposed model can be dynamically update with time to adapt the change of stationary crowd group.
    • model personality, which is a key factor that makes each individual behave differently.

Pedestrian Behavior Modeling

A pedestrian usually selects the most convenient and effcient path for reaching the destination ;
propose a general scene energy map to model the traveling diffculty of enery location of the scene .
is also a probability map shows the probability of pedestrian appearing at each location.

Personalized energy maps are generated based on the general energy map and a personality parameter P;
can be view as a different pedestrians' interpretations of the general map .

General energy map modeling

is modeled with three channel caculated based on

  • : Scene Layout
  • : Moving Pedestrians
  • : Stationary Groups

are weight paramters for different terms

Scene layout factor

  • : a set of locations occuied by scene obstacles which are unreachables

  • measures the distance from the current location to its nearest scene obstacle location
  • : a parameter indicating the influence bandwidth(importance).

Influence of moving pedestrians

  • : the th moving pedestrian
  • : the spatical location of at current time

  • measures the distance from the current location to the moving pedestrian
  • : the influence bandwidth of the moving pedestrian term.

Influence of stationary crowd groups

  • : the th stationary crowd group region automatically dected

  • measures the distance from to the stationary crowd group region
  • : the influence bandwidth of the stationary crowd group term
  • : calculated as the average distance among group members
    measure the sparsity of stationary crowd group region
  • : control the influence of group sparsity on estimation result

Personalized energy map modeling

: personality parameter, equivalent to the influence bandwidth of the terms

  • large means the energy values are small at locations near obstacles and stationary crowd group.
  • small means that the pedestrian is walking aggressively and cares less about abstacles

Path generation

: the most effcient and probable route from to according to the current energy map or

Model learning

by dividing a marginalization term, , the energy map can be transformed to a probility distribution :

where

Gaven as independent observation of , the likelihood of these observation is :

Parameter can be then be optimized as

Gradient descent is used for updating parameters

Pedestrian walking route dataset

Dataset details

  1. much longer than any existing one with ground truth on tracking
  2. a crowd surveillance dataset which is difficult and challenging for vison task.
  3. The dataset is well annotated

Statiscal analysis of the annotated data

The strong correlations between (a) and (b)-(c) indicate that stationary crowd is a key factor that decreases traffic effcienty.

Learning Result

A pedestrian is not sensitive to scene obstacles

Apedestrian might prefer to adjust walking speed rather than change predecided walking direction to avoid close contact with other moving pedestrians.

When stationary crowds emerge in front of a pedestrian, he/she has to change his/her walking route to bypass the stationary crowds.

Applications

Prediction on pedestrian walking paths

over cost value is define as

  • : walking cost of the observed route based on the current map
  • : the cost of the optimized route
  • small indicates better match

conclusion : the influence of stationary crowd groups is necessary when modeling pedestrian behaviors, and the stationary crowd groups should be modeled differently from scene obstacles.

Prediction of pedestrian destinations

Gaven and part of the walking path, we can also predict the destination of this pedestrian

  • : the first half of observated trajectory
  • : the first half of which is the optimized route ended with
  • : the distance between the two half trajectories
  • Smaller indicates that the pedestrian is more likely to go to the destination which are manually labeled.

Personality attribute estimation

  • : the observed trajectory of current pedestrian
  • : the optimal walking path calculated using personalized energy
  • : the distance between the two trajectories

All the pedestrians can be classified into three categories based on their walking behaviors: aggressive, conservative, and abnormal.

Conclusion


拓展研读

前继文章

进一步学习

添加新批注
在作者公开此批注前,只有你和作者可见。
回复批注