IdentityDP: Differential private identification protection for face images
نویسندگان
چکیده
• We propose a general framework that is suitable for the de-identification of people in face images. are first to introduce rigorously formulated DP theory into face-anonymous task. Our method does not require special training or targeted adjustments many unauthorized identity verification systems datasets never seen before. show images anonymized by our still usable identity-agnostic computer vision tasks (such as monitoring and tracking). de-identified significantly little computationally complex consumes small amount computing resources. Because explosive growth photos well their widespread dissemination easy accessibility social media, security privacy personal information become an unprecedented challenge. Meanwhile, convenience brought advanced technologies attractive. Therefore, it important use while taking careful consideration protecting people’s identities. Given image, de-identification, also known anonymization, refers generating another image with similar appearance same background, real hidden. Although extensive efforts have been made, existing techniques either insufficient photo-reality incapable well-balancing utility. In this paper, we focus on tackling these challenges improve de-identification. IdentityDP, anonymization combines data-driven deep neural network differential (DP) mechanism. This encompasses three stages: facial representations disentanglement, ? -IdentityDP perturbation reconstruction. model can effectively obfuscate identity-related faces, preserve significant visual similarity, generate high-quality be used tasks, such detection, tracking, etc. Different from previous methods, adjust balance utility through budget according practical demands provide diversity results without pre-annotations. Extensive experiments demonstrate effectiveness generalization ability proposed framework.
منابع مشابه
Ethnicity Identification from Face Images
Human facial images provide the demographic information, such as ethnicity and gender. Conversely, ethnicity and gender also play an important role in face-related applications. Image-based ethnicity identification problem is addressed in a machine learning framework. The Linear Discriminant Analysis (LDA) based scheme is presented for the two-class (Asian vs. non-Asian) ethnicity classificatio...
متن کاملRobust Identification of Face Landmarks in Profile Images
Identifying landmark points in facial images is a very important task in computer vision and has several applications. Here we present a fast and robust algorithm capable of identifying a specific set of landmarks on face profile images. The algorithm is based on the local curvature of the profile contour and on the local analysis of the face features. In order to validate the approach, the alg...
متن کاملSupervised Feature Extraction of Face Images for Improvement of Recognition Accuracy
Dimensionality reduction methods transform or select a low dimensional feature space to efficiently represent the original high dimensional feature space of data. Feature reduction techniques are an important step in many pattern recognition problems in different fields especially in analyzing of high dimensional data. Hyperspectral images are acquired by remote sensors and human face images ar...
متن کاملPower Differential based Wide Area Protection
Current differential based wide area protection (WAP) has recently been proposed as a technique to increase the reliability of protection systems. It increases system stability and can prevent large contingencies such as cascading outages and blackouts. This paper describes how power differential protection (PDP) can be used within a WAP and shows that the algorithm operates correctly for a...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: Neurocomputing
سال: 2022
ISSN: ['0925-2312', '1872-8286']
DOI: https://doi.org/10.1016/j.neucom.2022.06.039