MissFormer: (In-)attention-based handling of missing observations for trajectory filtering and prediction

TL;DR

MissFormer introduces a transformer-based model capable of handling missing observations in trajectory data, improving object tracking and prediction in scenarios with incomplete or noisy data.

Contribution

This paper presents a novel transformer-based approach that explicitly models missing data in trajectory sequences, enabling robust filtering and prediction under incomplete observations.

Findings

Effective handling of missing data in synthetic and real-world tracking scenarios

Improved trajectory inference from noisy and incomplete inputs

Model acts as both filter and predictor depending on missing data pattern

Abstract

In applications such as object tracking, time-series data inevitably carry missing observations. Following the success of deep learning-based models for various sequence learning tasks, these models increasingly replace classic approaches in object tracking applications for inferring the objects' motion states. While traditional tracking approaches can deal with missing observations, most of their deep counterparts are, by default, not suited for this. Towards this end, this paper introduces a transformer-based approach for handling missing observations in variable input length trajectory data. The model is formed indirectly by successively increasing the complexity of the demanded inference tasks. Starting from reproducing noise-free trajectories, the model then learns to infer trajectories from noisy inputs. By providing missing tokens, binary-encoded missing events, the model learns to in-attend to missing data and infers a complete trajectory conditioned on the remaining inputs. In the case of a sequence of successive missing events, the model then acts as a pure prediction model. The abilities of the approach are demonstrated on synthetic data and real-world data reflecting prototypical object tracking scenarios.