Lifting Multi-View Detection and Tracking to the Bird's Eye View

TL;DR

This paper enhances multi-view detection and tracking by aggregating multi-view features onto a bird's eye view, integrating appearance and motion cues, and generalizing across multiple datasets for improved performance.

Contribution

It introduces a novel architecture that combines multi-view feature aggregation with multi-step learning and cross-domain generalization for detection and tracking.

Findings

Achieves state-of-the-art results on multiple datasets.

Effectively combines appearance and motion cues.

Generalizes across pedestrian and roadside perception domains.

Abstract

Taking advantage of multi-view aggregation presents a promising solution to tackle challenges such as occlusion and missed detection in multi-object tracking and detection. Recent advancements in multi-view detection and 3D object recognition have significantly improved performance by strategically projecting all views onto the ground plane and conducting detection analysis from a Bird's Eye View. In this paper, we compare modern lifting methods, both parameter-free and parameterized, to multi-view aggregation. Additionally, we present an architecture that aggregates the features of multiple times steps to learn robust detection and combines appearance- and motion-based cues for tracking. Most current tracking approaches either focus on pedestrians or vehicles. In our work, we combine both branches and add new challenges to multi-view detection with cross-scene setups. Our method generalizes to three public datasets across two domains: (1) pedestrian: Wildtrack and MultiviewX, and (2) roadside perception: Synthehicle, achieving state-of-the-art performance in detection and tracking. https://github.com/tteepe/TrackTacular