Efficient Dense Crowd Trajectory Prediction Via Dynamic Clustering

2026-03-18

AI Summary

This work addresses the challenges of trajectory prediction in dense crowd scenarios, where large data volumes, high noise levels, and substantial computational costs hinder performance. To tackle these issues, the authors propose an efficient trajectory prediction approach based on dynamic clustering, which adaptively aggregates individuals with similar motion attributes over time into cohesive groups. The trajectory of each group’s centroid is then used as input to the prediction model in place of individual trajectories. A plug-and-play dynamic clustering mechanism is designed to seamlessly integrate with existing prediction architectures. Extensive experiments on multiple dense-scene datasets demonstrate that the proposed method significantly reduces computational overhead and memory consumption while maintaining competitive prediction accuracy, thereby enabling faster inference and improved overall efficiency.

Technology Category

• Humans and AI: Human-Aware Planning and Behavior Prediction
• Machine Learning: Clustering
• Computer Vision: Motion & Tracking

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Economics, Online Markets and Human Computation: Architectures and workflows that use LLMs for crowd work
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

Crowd trajectory prediction plays a crucial role in public safety and management, where it can help prevent disasters such as stampedes. Recent works address the problem by predicting individual trajectories and considering surrounding objects based on manually annotated data. However, these approaches tend to overlook dense crowd scenarios, where the challenges of automation become more pronounced due to the massiveness, noisiness, and inaccuracy of the tracking outputs, resulting in high computational costs. To address these challenges, we propose and extensively evaluate a novel cluster-based approach that groups individuals based on similar attributes over time, enabling faster execution through accurate group summarisation. Our plug-and-play method can be combined with existing trajectory predictors by using our output centroid in place of their pedestrian input. We evaluate our proposed method on several challenging dense crowd scenes. We demonstrated that our approach leads to faster processing and lower memory usage when compared with state-of-the-art methods, while maintaining the accuracy

Problem

• dense crowd
• trajectory prediction
• computational cost
• tracking noise
• public safety

Innovation

• dynamic clustering
• dense crowd trajectory prediction
• group summarization
• plug-and-play
• computational efficiency