SpikingMOT: A Spike-Driven Multi-Object Tracker

2026-07-22

AI Summary

This work addresses the computational redundancy and low energy efficiency of dense neural networks in multi-object tracking by introducing spiking neural networks (SNNs) to this task for the first time. The authors propose a sparsity-driven trajectory prediction method that decomposes trajectory states into pseudo-trajectory bases and incorporates a prediction-error-driven posterior calibration mechanism, establishing a brain-inspired sparse dynamic modeling loop. Additionally, an activation sparsity bias mechanism is designed to further enhance efficiency. Evaluated on SportsMOT and DanceTrack, the approach achieves HOTA scores of 74.9 and 56.5, respectively, while reducing model parameters by 72% and energy consumption by 86.7%, thereby significantly improving energy efficiency without compromising—and even surpassing—state-of-the-art performance.

Abstract

Multi-object tracking (MOT) plays a fundamental role in visual perception, where accurate trajectory prediction is essential for reliable target association under complex motion patterns. Recent trackers have improved motion modeling with densely activated artificial neural networks, yet they largely overlook whether such dense responses are necessary for trajectory prediction. In this paper, we formulate activation sparsity preference (ASP) by tackling two key questions: 1. How can we identify a model architecture that appropriately and formally explains ASP, and 2. How can we translate this explanation into competitive tracking performance. Theoretical analysis shows that sparse gating is no worse than state-independent dropout under the same activation rate. Based on this insight, SpikingMOT is proposed as a spike-driven tracker that adaptively models sparse trajectory dynamics with spiking neural networks (SNNs). Specifically, SpikingMOT decomposes each trajectory state into pseudo-trajectory bases and uses the current prediction error to calibrate the posterior for next-frame prediction. With this brain-inspired loop, SpikingMOT achieves state-of-the-art performance in extensive experiments, 74.9 HOTA on SportsMOT and 56.5 HOTA on DanceTrack, while reducing the parameters and energy by 72% and 86.7%, respectively. These results bring SNNs into MOT, opening a promising direction for efficient tracking.

Problem

• multi-object tracking
• activation sparsity
• trajectory prediction
• spiking neural networks
• motion modeling

Innovation

• Spiking Neural Networks
• Activation Sparsity
• Multi-Object Tracking
• Trajectory Prediction
• Energy Efficiency