Abstract:Multi-object tracking (MOT) in the scenario of low-frame-rate videos is a promising solution to better meet the computing, storage, and transmitting bandwidth resource constraints of edge devices. Tracking with a low frame rate poses particular challenges in the association stage as objects in two successive frames typically exhibit much quicker variations in locations, velocities, appearances, and visibilities than those in normal frame rates. In this paper, we observe severe performance degeneration of many existing association strategies caused by such variations. Though optical-flow-based methods like CenterTrack can handle the large displacement to some extent due to their large receptive field, the temporally local nature makes them fail to give reliable displacement estimations of objects that newly appear in the current frame (i.e., not visible in the previous frame). To overcome the local nature of optical-flow-based methods, we propose an online tracking method by extending the CenterTrack architecture with a new head, named APP, to recognize unreliable displacement estimations. Further, to capture the fine-grained and private unreliability of each displacement estimation, we extend the binary APP predictions to displacement uncertainties. To this end, we reformulate the displacement estimation task via Bayesian deep learning tools. With APP predictions, we propose to conduct association in a multi-stage manner where vision cues or historical motion cues are leveraged in the corresponding stage. By rethinking the commonly used bipartite matching algorithms, we equip the proposed multi-stage association policy with a hybrid matching strategy conditioned on displacement uncertainties. Our method shows robustness in preserving identities in low-frame-rate video sequences. Experimental results on public datasets in various low-frame-rate settings demonstrate the advantages of the proposed method.

Multi-person tracking-by-detection with local particle filtering and global occlusion handling

APPTracker Plus : Displacement Uncertainty for Occlusion Handling in Low-Frame-Rate Multiple Object Tracking

A Particle Filter Based Multi-person Tracking with Occlusion Handling

Multi-object Tracking Through Occlusions by Local Tracklets Filtering and Global Tracklets Association with Detection Responses

Online Multi-Camera Tracking-by-detection Approach with Particle Filter

An Efficient Multi-object Tracking Method Using Multiple Particle Filters

A Novel Particle Filter Based People Tracking Method Through Occlusion

Beyond Local Search: Tracking Objects Everywhere with Instance-Specific Proposals

Multi-Object Tracking And Occlusion Reasoning Based On Adaptive Weighing Particle Filter

Person Tracking with Partial Occlusion Handling.

Occlusion Handling Based on Particle Filter in Surveillance System

Multi-person Multi-Camera Tracking for Live Stream Videos Based on Improved Motion Model and Matching Cascade

Probabilistic Detection-based Particle Filter for Multi-target Tracking.

Person tracking based on combination of blob modeling and particle filter

Person-Tracking With Occlusion Using Appearance Filters

Deep Alignment Network Based Multi-person Tracking with Occlusion and Motion Reasoning

Memory-based State Estimation for Handling Occlusion During Object Tracking by Particle Filter

Dynamic Multi-Cue Tracking Using Particle Filter

Multi-object Tracking by Expanding Long-Tracklets

Multi-Camera Multi-Person Tracking for Intelligent Meeting Scenarios

Multi-object tracking based on behaviour and partial observation