Impact of Conventional and AI-based Image Coding on AI-based Face Recognition Performance

AI-based face recognition has become increasingly appealing for daily life applications due to its high performance. At the same time, image coding is very commonly used and, thus, many applications perform face recognition with decoded images. However, using decoded images, which may suffer from compression artifacts, may impact the final decision-making process of AI-based face recognition systems and, thus, its overall recognition performance. This paper studies the impact of image coding on the overall face verification performance of a popular and high performing face recognition solution, ArcFace. Face recognition using both original and decoded images, with several compression rates and qualities, is considered. Tests were performed using the Labeled Faces in the Wild (LFW) face dataset, with its images coded using conventional image coding standards, notably JPEG, JPEG 2000, and JPEG XL, as well as three emerging AI-based image codecs. As expected, the experimental results show that coding can have a significant impact on face recognition performance, with its impact becoming increasingly relevant as the coding rate is reduced. It is also observed that the recent AI-based image codecs appear to offer slightly better recognition performance for the same coding rates as a consequence of their better RD performance.

I. Introduction

Face recognition has become a fundamental technology in multiple real-world applications, using pattern recognition-based tools to offer an increased level of security for many use cases. Moreover, face recognition solutions use increasingly more sophisticated artificial intelligence (AI) techniques, which have shown significantly improved performance in recent years. However, these performance gains are often associated with higher complexity and harder to understand and explain systems – consequently modern AI-based face recognition system are sometimes referred to as ‘black-box’ systems. This raises the risk of not trusting AI-based facial recognition technology if its results cannot be minimally explained, especially when this technology also brings risks in terms of privacy, which is a very sensitive societal issue. Thus, facial recognition explainability has become a pivotal step to understand AI-based facial recognition systems’ behavior and increase trust on this type of technology. Explainability can be the decisive factor to enable the usage of face recognition systems that conform with the European initiative entitled "Artificial Intelligence Act" [1], where most biometric recognition systems are considered "high-risk AI", and the requirements for their authorization will include the need for transparency and information to the user.