Eureka AIR delivers breakthrough ideas for toughest innovation challenges, trusted by R&D personnel around the world.

A Crowd Trajectory Prediction Method Based on Multi-precision Interaction

A trajectory prediction, multi-precision technology, applied in the field of computer vision, can solve the problems of poor modeling accuracy, high complexity of pedestrian interaction, lack of heterogeneity in pedestrian interaction, etc., and achieve the effect of accurate pedestrian trajectory and accurate prediction.

Active Publication Date: 2021-12-14
BEIJING UNIV OF POSTS & TELECOMM
View PDF7 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0011] In view of the large number of pedestrians in complex scenes, the complexity of calculating pedestrian interactions based on the attention mechanism is high, and the modeling accuracy of the overall interaction modeling for the global scene is poor, and the interactions of different pedestrians lack heterogeneity. The present invention proposes A crowd trajectory prediction method based on multi-precision interaction, which makes pedestrian interaction modeling more detailed and reasonable through the interactive modeling methods of Global Interaction Module (GIM) and Local Interaction Module (LIM) , to achieve the effect of improving the accuracy of crowd trajectory prediction, and has the advantage of reducing the number of calculations while ensuring the personalized interaction of pedestrians

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • A Crowd Trajectory Prediction Method Based on Multi-precision Interaction
  • A Crowd Trajectory Prediction Method Based on Multi-precision Interaction
  • A Crowd Trajectory Prediction Method Based on Multi-precision Interaction

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0033] The implementation of the technical solutions of the present invention will be described below in conjunction with the drawings and embodiments, and the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without creative efforts fall within the protection scope of the present invention.

[0034] Existing methods generally calculate the interaction between each pair of pedestrians, which has high computational complexity, or use methods such as pooling layers to model the entire scene as a whole. This method leads to the same interaction of different pedestrians and lacks diversity. The present invention proposes a crowd trajectory prediction method based on multi-precision interaction by means of deep learning technology, adopts a multi-precision pedestrian interaction modeling method, that is, a global-l...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention proposes a crowd trajectory prediction method based on multi-precision interaction, which belongs to the technical field of computer vision. The method of the present invention includes: using LSTM network to encode the historical state of pedestrians; adopting global interactive modeling and local interactive modeling methods for pedestrian motion interaction, firstly dividing the scene when global interactive modeling, and performing regional analysis on each divided sub-area Modeling, and then integrating to obtain global interaction information, and finally splicing local interaction information and global interaction information to obtain complete interaction information; at the decoding end, an LSTM network is used to decode and predict pedestrian trajectories. The trajectory of pedestrians can be fine-tuned through the global interaction information, and the detailed motion information of pedestrians can be captured through the local interaction information. Pedestrians can make timely avoidance of recent movements and get closer to the real trajectory. By adopting the method of the invention, the trajectory of pedestrians can be predicted more accurately, and the computational complexity is reduced while ensuring the heterogeneity of mutual information.

Description

technical field [0001] The invention belongs to the technical field of computer vision, and relates to crowd trajectory prediction and crowd interaction modeling technology, in particular to a crowd trajectory prediction method based on multi-precision interaction. Background technique [0002] The trajectory prediction algorithm realizes the prediction of the target's future trajectory information based on the historical position information of pedestrians, which is one of the research hotspots in the field of computer vision at present. With the development of 5G network commercialization and urban modernization, the country attaches great importance to the construction of key directions such as smart transportation and public safety. As a necessary basic research technology, trajectory prediction algorithms are used in real-life scenarios such as autonomous driving, robot navigation, and pedestrian intention analysis. play a key role in. For example, in a robot navigatio...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G06T7/207G06T7/246G06N3/04
CPCG06T7/207G06T7/246G06T2207/10016G06T2207/30196G06T2207/30241G06N3/045
Inventor 刘绍华孙靖凯
Owner BEIJING UNIV OF POSTS & TELECOMM
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Eureka Blog
Learn More
PatSnap group products