Learning Domain and Pose Invariance for Thermal-to-Visible Face Recognition

TL;DR

This paper introduces a novel framework that learns domain and pose invariant representations to improve thermal-to-visible face recognition, especially for off-pose thermal images, outperforming existing methods.

Contribution

The proposed framework simultaneously learns domain and pose invariant features using modified networks and a joint-loss function, addressing pose variations in thermal-to-visible face recognition.

Findings

Effective on three thermal-visible datasets

Improves off-pose thermal to frontal visible face matching

Enhances performance even for frontal thermal to frontal visible matching

Abstract

Interest in thermal to visible face recognition has grown significantly over the last decade due to advancements in thermal infrared cameras and analytics beyond the visible spectrum. Despite large discrepancies between thermal and visible spectra, existing approaches bridge domain gaps by either synthesizing visible faces from thermal faces or by learning the cross-spectrum image representations. These approaches typically work well with frontal facial imagery collected at varying ranges and expressions, but exhibit significantly reduced performance when matching thermal faces with varying poses to frontal visible faces. We propose a novel Domain and Pose Invariant Framework that simultaneously learns domain and pose invariant representations. Our proposed framework is composed of modified networks for extracting the most correlated intermediate representations from off-pose thermal and frontal visible face imagery, a sub-network to jointly bridge domain and pose gaps, and a joint-loss function comprised of cross-spectrum and pose-correction losses. We demonstrate efficacy and advantages of the proposed method by evaluating on three thermal-visible datasets: ARL Visible-to-Thermal Face, ARL Multimodal Face, and Tufts Face. Although DPIF focuses on learning to match off-pose thermal to frontal visible faces, we also show that DPIF enhances performance when matching frontal thermal face images to frontal visible face images.