Gaze estimation models are widely used in applications such as driver attention monitoring and human-computer interaction. While many methods for gaze estimation exist, they rely heavily on data-hungry deep learning to achieve high performance. This reliance often forces practitioners to harvest training data from unverified public datasets, outsource model training, or rely on pre-trained models. However, such practices expose gaze estimation models to backdoor attacks. In such attacks, adversaries inject backdoor triggers by poisoning the training data, creating a backdoor vulnerability: the model performs normally with benign inputs, but produces manipulated gaze directions when a specific trigger is present. This compromises the security of many gaze-based applications, such as causing the model to fail in tracking the driver's attention. To date, there is no defense that addresses backdoor attacks on gaze estimation models. In response, we introduce SecureGaze, the first solution designed to protect gaze estimation models from such attacks. Unlike classification models, defending gaze estimation poses unique challenges due to its continuous output space and globally activated backdoor behavior. By identifying distinctive characteristics of backdoored gaze estimation models, we develop a novel and effective approach to reverse-engineer the trigger function for reliable backdoor detection. Extensive evaluations in both digital and physical worlds demonstrate that SecureGaze effectively counters a range of backdoor attacks and outperforms seven state-of-the-art defenses adapted from classification models.

Eye gaze contains rich information about human attention and cognitive processes. This capability makes the underlying technology, known as gaze tracking, a critical enabler for many ubiquitous applications and has triggered the development of easy-to-use gaze estimation services. Indeed, by utilizing the ubiquitous cameras on tablets and smartphones, users can readily access many gaze estimation services. In using these services, users must provide their full-face images to the gaze estimator, which is often a black box. This poses significant privacy threats to the users, especially when a malicious service provider gathers a large collection of face images to classify sensitive user attributes. In this work, we present PrivateGaze, the first approach that can effectively preserve users’ privacy in black-box gaze tracking services without compromising gaze estimation performance. Specifically, we proposed a novel framework to train a privacy preserver that converts full-face images into obfuscated counterparts, which are effective for gaze estimation while containing no privacy information. Evaluation on four datasets shows that the obfuscated image can protect users’ private information, such as identity and gender, against unauthorized attribute classification. Meanwhile, when used directly by the black-box gaze estimator as inputs, the obfuscated images lead to comparable tracking performance to the conventional, unprotected full-face images.

Smartphones are moving towards the fullscreen design for better user experience. This trend forces front cameras to be placed under screen, leading to Under-Screen Cameras (USC). Accordingly, a small area of the screen is made translucent to allow light to reach the USC. In this paper, we utilize the translucent screen's features to inconspicuously modify its pixels, imperceptible to human eyes but inducing perturbations on USC images. These screen perturbations affect deep learning models in image classification and face recognition. They can be employed to protect user privacy, or disrupt the front camera's functionality in the malicious case. We design two methods, one-pixel perturbation and multiple-pixel perturbation, that can add screen perturbations to images captured by USC and successfully fool various deep learning models. Our evaluations, with three commercial full-screen smartphones on testbed datasets and synthesized datasets, show that screen perturbations significantly decrease the average image classification accuracy, dropping from 85% to only 14% for one-pixel perturbation and 5.5% for multiple-pixel perturbation. For face recognition, the average accuracy drops from 91% to merely 1.8% and 0.25%, respectively.