Transformers for Multi-Object Tracking on Point Clouds

TL;DR

This paper introduces TransMOT, a transformer-based end-to-end online tracker and detector for point cloud data, leveraging attention mechanisms to improve multi-object tracking in automotive sensor data.

Contribution

The paper presents a novel transformer architecture that unifies detection and tracking in point cloud data, using a feature-space approach and a new module for track prediction.

Findings

Outperforms Kalman filter-based baseline on nuScenes dataset

Handles sensor input at arbitrary timesteps and frame skips

Utilizes rich latent space for improved tracking accuracy

Abstract

We present TransMOT, a novel transformer-based end-to-end trainable online tracker and detector for point cloud data. The model utilizes a cross- and a self-attention mechanism and is applicable to lidar data in an automotive context, as well as other data types, such as radar. Both track management and the detection of new tracks are performed by the same transformer decoder module and the tracker state is encoded in feature space. With this approach, we make use of the rich latent space of the detector for tracking rather than relying on low-dimensional bounding boxes. Still, we are able to retain some of the desirable properties of traditional Kalman-filter based approaches, such as an ability to handle sensor input at arbitrary timesteps or to compensate frame skips. This is possible due to a novel module that transforms the track information from one frame to the next on feature-level and thereby fulfills a similar task as the prediction step of a Kalman filter. Results are presented on the challenging real-world dataset nuScenes, where the proposed model outperforms its Kalman filter-based tracking baseline.