Tag
该论文提出CARPM-FIQA策略，通过在整个训练轨迹上累积相对点边际分数（类内紧凑性与类间分离度之比）来解决FR集成式人脸图像质量评估中的训练不稳定问题，并在多个基准测试上验证了其有效性。
Parents and children in Illinois and California are suing Meta for allegedly using their Facebook and Instagram photos without consent to train AI models and develop an unreleased face-recognition system, violating privacy laws.
Dating apps like Tinder are adopting face recognition to verify user identities and combat AI-generated fakes and scams, raising privacy and security concerns.
EXPL-FR is a method that explains face recognition models by aligning vision-language embeddings to the recognition space, enabling label-free auditing of semantic attributes and model comparison without requiring architecture access.
A WIRED feature examining the privacy concerns around camera-equipped smart glasses, especially Meta's Ray-Ban glasses, and how companies like Google, Samsung, and Apple are approaching the issue.
Meta's NameTag face recognition feature for Ray-Ban glasses exists in code but is inactive; executives claim it doesn't exist yet, while CTO described it in detail on a podcast, creating contradiction.
The article argues that age verification laws requiring facial recognition and government ID are actually forced identity tracking that threatens privacy, as biometric data cannot be changed if leaked.
A fully free and open-source HTTP API service based on macOS native capabilities, offering image OCR, multi-language translation, web content retrieval, face recognition, QR/barcode recognition, and text-to-speech.
Meta is working with Rank One Computing, a Pentagon supplier known for surveillance tools, to prototype face recognition for its smart glasses, raising concerns about the convergence of military-grade surveillance and consumer technology.
A Florida man was wrongfully arrested after police relied on a face-recognition match from the FACES system that was inaccurate, leading to a lawsuit by the ACLU highlighting systemic flaws in one of the oldest police face-recognition tools in the US.
Meta removed face-recognition code from its Meta AI app after WIRED exposed the unreleased NameTag system, which was designed to identify people using smart glasses.
The Electronic Frontier Foundation outlines its approach to AI, emphasizing civil liberties, surveillance risks, regulation of use rather than technology, and the need for competition and privacy protections.
WIRED analysis reveals Meta secretly embedded face-recognition code called 'NameTag' into its AI app on millions of phones, capable of identifying people captured by Ray-Ban and Oakley smart glasses cameras, despite the company publicly claiming it was still 'thinking through' such a feature. The dormant system converts faces into biometric faceprints and checks them against an on-device database, reviving technology Meta claimed to have abandoned in 2021.
YouTube is expanding its AI likeness detection tool to all users aged 18 and older, enabling anyone to monitor for deepfakes of their face and request removal through YouTube's privacy policy.