Video Face Retouch——润饰一张静态帧，不重新生成

使用流程

Video Face Retouch applies a gentle, conservative polish to a still frame you have pulled out of a video. To be clear about what this is: it works on one uploaded image. It does not open video files and it does not follow a face from frame to frame. Each face is checked on its own, skin is separated from hair, beard, clothing and background, and the eyes, brows, nose and lips are protected before anything changes. Only three bounded areas on the forehead and both cheeks are touched, at a deliberately reduced strength, so a frame still cuts together with the ones around it. Pores, fine detail and age are kept, and everything else is copied straight into a lossless PNG. In a group shot every face is handled separately.

处理中

视频脸部修饰使用 AI 动作。任何单独的余量都会在该效果开始处理之前显示。

最适合

sharp, evenly lit still frames with a complete near-front face and an uncovered forehead, eyebrows, eyes, cheeks, nose, lips, jaw and chin — one person or a whole group.

不太适合： video files, which are not accepted, plus side profiles, tilted or cropped faces, glasses, hair or hands covering features, very small faces in a wide shot, motion-blurred frames, and skin that is already even.