HeadPosr: End-to-end Trainable Head Pose Estimation using Transformer Encoders

TL;DR

HeadPosr introduces an end-to-end trainable head pose estimation method using transformer encoders, demonstrating superior performance and establishing new benchmarks across multiple datasets.

Contribution

The paper presents a novel transformer-based architecture for head pose estimation, including extensive ablation studies and outperforming existing state-of-the-art methods.

Findings

Outperforms all compared methods on AFLW2000 and BIWI datasets.

Demonstrates the effectiveness of transformer encoders in HPE.

Sets new benchmarks for head pose estimation accuracy.

Abstract

In this paper, HeadPosr is proposed to predict the head poses using a single RGB image. \textit{HeadPosr} uses a novel architecture which includes a transformer encoder. In concrete, it consists of: (1) backbone; (2) connector; (3) transformer encoder; (4) prediction head. The significance of using a transformer encoder for HPE is studied. An extensive ablation study is performed on varying the (1) number of encoders; (2) number of heads; (3) different position embeddings; (4) different activations; (5) input channel size, in a transformer used in HeadPosr. Further studies on using: (1) different backbones, (2) using different learning rates are also shown. The elaborated experiments and ablations studies are conducted using three different open-source widely used datasets for HPE, i.e., 300W-LP, AFLW2000, and BIWI datasets. Experiments illustrate that \textit{HeadPosr} outperforms all the state-of-art methods including both the landmark-free and the others based on using landmark or depth estimation on the AFLW2000 dataset and BIWI datasets when trained with 300W-LP. It also outperforms when averaging the results from the compared datasets, hence setting a benchmark for the problem of HPE, also demonstrating the effectiveness of using transformers over the state-of-the-art.