面容通常被视作读解个体内在本质的窗口，但这一窗口远非透明无碍，而更像是“揭示—隐藏”的冲突地带。进入算法时代，面容被日益兴起的机器视觉记录、识别和分析，由此成为“摄取—隐匿”的对抗场所。机器视觉犹如《哈利·波特》魔法世界中的摄魂怪，以公共安全等名义摄取面容，识别目标对象的身份、情感和价值。为了抵抗机器视觉无休止的面容摄取，个体采取蒙面、易容等“隐匿术”，试图令其摄取过程遭遇挫败或发生故障。然而，当面容监控超越道德界限并与其所谓的公共安全目标相矛盾时，个体的“隐匿术”就会暴露其固有的局限性。除了良好的制度保障，我们还需要彼此间的守望相助，用“守护咒”驱散部分居心叵测的“摄魂怪”，从而更好地保护面部隐私和数字自主权。

面容摄取, 机器视觉, 算法治理, 生物识别, 隐私保护

Conventionally framed as a window to the inner essence of individuals, the human face proves far from transparent. Rather, it operates as a contested terrain of revelation and concealment. In the algorithmic  era, the face is increasingly recorded, identified, and analyzed by machine vision technologies, thereby becoming a battlefield of extraction and concealment. Machine vision, akin to the soul-sucking Dementors  in the“Harry Potter”universe, extracts facial data in the name of public security, decoding it to identify  individuals, their emotions, and value orientations. To counteract this relentless facial extraction, individuals  resort to concealment techniques such as masking and disguise, aiming to frustrate or disrupt the extraction  process. However, when facial surveillance transcends moral boundaries and contradicts its purported public  safety goals, individual concealment techniques reveal their “Arts of Concealing”. Beyond robust institutional safeguards, we also need mutual support and vigilance, employing a “guardian charm” to dispel some of the scheming “Dementors”, thereby better protecting facial privacy and digital autonomy.

facial extraction, machine vision, algorithmic governance, biometrics, privacy protection