Model-based Multi-object Visual Tracking: Identification and Standard Model Limitations

TL;DR

This paper applies multi-object tracking methods from radar to pedestrian tracking with 2D bounding boxes, using the SPO model and PMBM filter, revealing model-data mismatches and proposing future improvements.

Contribution

It adapts the SPO model and PMBM filter for pedestrian tracking and discusses parameter selection, highlighting limitations and potential model modifications.

Findings

PMBM filter yields promising results on MOT-17 dataset

Identified mismatch between SPO model and pedestrian data

Discussion on parameter selection from first principles and data

Abstract

This paper uses multi-object tracking methods known from the radar tracking community to address the problem of pedestrian tracking using 2D bounding box detections. The standard point-object (SPO) model is adopted, and the posterior density is computed using the Poisson multi-Bernoulli mixture (PMBM) filter. The selection of the model parameters rooted in continuous time is discussed, including the birth and survival probabilities. Some parameters are selected from the first principles, while others are identified from the data, which is, in this case, the publicly available MOT-17 dataset. Although the resulting PMBM algorithm yields promising results, a mismatch between the SPO model and the data is revealed. The model-based approach assumes that modifying the problematic components causing the SPO model-data mismatch will lead to better model-based algorithms in future developments.