Autoencoder Based Inter-Vehicle Generalization for In-Cabin Occupant Classification

TL;DR

This paper introduces an autoencoder-based method for occupant classification that enhances model transferability across different vehicle interiors without needing target domain data during training.

Contribution

It proposes a novel autoencoder approach that improves inter-vehicle occupant classification generalization, outperforming some pre-trained models and transforming images from unknown vehicles.

Findings

Autoencoder matches classification models trained from scratch.

Autoencoder outperforms some pre-trained models.

Transforms images from unknown vehicles into the trained vehicle domain.

Abstract

Common domain shift problem formulations consider the integration of multiple source domains, or the target domain during training. Regarding the generalization of machine learning models between different car interiors, we formulate the criterion of training in a single vehicle: without access to the target distribution of the vehicle the model would be deployed to, neither with access to multiple vehicles during training. We performed an investigation on the SVIRO dataset for occupant classification on the rear bench and propose an autoencoder based approach to improve the transferability. The autoencoder is on par with commonly used classification models when trained from scratch and sometimes out-performs models pre-trained on a large amount of data. Moreover, the autoencoder can transform images from unknown vehicles into the vehicle it was trained on. These results are corroborated by an evaluation on real infrared images from two vehicle interiors.