VolgendeVolgende gids
Text-to-Motion and AI Character Animation
Visuele AI
Visuele AI-GIDS
Portrait Animation from a Single Photo
Portrait animation from a single photo makes a still face move, including turning, blinking, changing expression or talking.
Op deze pagina4 minuten lezen
Overzicht
It works by transferring motion from a driving video or audio track onto the source image. It powers talking-head avatars, animated family photos and presenter videos, and it raises consent concerns because one public photo is enough input.
Diepe duik
Every system has four parts: a source image that supplies appearance, a driving signal (a video of another face, or audio), a motion representation, and a generator that warps the source and fills in gaps. The First Order Motion Model (Siarohin et al., NeurIPS 2019) set the template. It learns keypoints without any labels, together with local affine transformations around each one. From the source and driving keypoints it predicts a dense motion field, which is a per-pixel flow, plus an occlusion map marking regions the source image cannot supply. It warps the source's feature maps with the flow and has a decoder paint in the occluded areas. It also uses relative motion transfer: it applies how the driving keypoints move relative to the driving video's first frame, not their absolute positions, so the source keeps its own face shape. SadTalker (CVPR 2023) animates from audio. It predicts 3D Morphable Model coefficients from speech, using ExpNet for expression and PoseVAE for head pose, then renders them through a keypoint-based face generator. LivePortrait, released by Kuaishou in 2024, uses implicit keypoints trained on a large dataset. It adds a stitching module that pastes the animated face back into the full image without misaligned shoulders, and retargeting modules for eye and lip openness. It is fast enough for near real-time use on a GPU. Diffusion-based methods such as Alibaba's EMO and the open Hallo project give more expressive results at much higher compute cost. One photo has no information about the sides of the head, the inside of the mouth or the teeth, so these regions are invented. Large head turns stretch the face, backgrounds and hair may warp along with it, and identity can drift. A common misconception is that these tools build a full 3D model of the person. Even 3DMM-based methods use only an approximate face model.
Strategische impact
Snelheid en schaal
Visuele AI kan inspectie-, detectie- en taggingtaken op schaal automatiseren.
Bouwkeuzes
Creatieve teams kunnen concepten sneller prototypen met minder handmatige revisies.
Team en workflow
Bij bewerkingen kan gebruik worden gemaakt van beeld- en videosignalen die voorheen moeilijk te verwerken waren.
The Future of Portrait Animation from a Single Photo
Real-time avatars for video calls, customer service and education are becoming practical, and diffusion priors are improving how models fill unseen regions such as teeth and profile views. Large head rotations, consistent hair and accessories, and long-duration identity stability remain difficult. Because a single photo is enough input, consent and disclosure are central concerns. Detection tools and provenance labeling are developing alongside the generators, but no detection method is reliable enough on its own.
Implementatie in de echte wereld
A genealogy app animates a scanned great-grandparent's photo with a subtle smile and blink. It looks convincing for a few seconds but breaks down if the head turns far.
An e-learning team uses a SadTalker-style pipeline to turn a presenter's headshot and narration audio into a talking-head course introduction.
A creator records themselves on a webcam to drive an illustrated character portrait with LivePortrait, using its retargeting controls to exaggerate eye opening.
A verification desk flags a supposed video of a CEO because the hair and earrings stay oddly rigid while the face moves, which suggests single-photo animation.
Risico's en vangrails
Beeldrechten en toestemming kunnen juridische risico's worden als de herkomst onduidelijk is.
De prestaties van modellen kunnen variëren afhankelijk van de belichting, demografische gegevens en omgevingen.
Valse positieve resultaten kunnen onopgemerkt blijven, tenzij de vertrouwensdrempels worden gecontroleerd.
Implementatie routekaart
Definieer acceptatiecriteria voor precisie-, terugroep- en foutkosten.
Test met gegevens die overeenkomen met echte productieomstandigheden.
Voeg menselijke beoordeling toe voor voorspellingen met weinig vertrouwen of hoge impact.
Volg modelafwijkingen en valideer opnieuw na wijzigingen in de camera of dataset.
Blijf verkennen
Free newsletter
Get the daily AI briefing
Three verified AI stories every weekday morning, written in plain English. Free forever, no ads.
One email each weekday. Unsubscribe in one click. We never sell or share your address.
Test yourself
Take the Portrait Animation from a Single Photo quiz
Instant feedback on every answer, and a shareable certificate with a verifiable ID once you pass a course.
Support free AI education. AI Understanding is a 501(c)(3) nonprofit — no ads, no paywall, ever. Make a donation
Veelgestelde vragen
What is Portrait Animation from a Single Photo?
Portrait animation from a single photo makes a still face move, including turning, blinking, changing expression or talking. It works by transferring motion from a driving video or audio track onto the source image. It powers talking-head avatars, animated family photos and presenter videos, and it raises consent concerns because one public photo is enough input.
How does the First Order Motion Model obtain its keypoints?
FOMM discovers keypoints in an unsupervised way. That is why it can work on faces and on other object categories.
What is the purpose of the occlusion map?
The occlusion map tells the generator which areas cannot come from warping the source and must be filled in.
What is the benefit of relative motion transfer?
Applying how the driving keypoints change, rather than where they are, stops the driver's facial geometry from replacing the source's.
What does SadTalker predict from the audio?
SadTalker's ExpNet and PoseVAE produce 3DMM coefficients, which a face renderer then turns into frames.
What does LivePortrait's stitching module do?
The stitching module keeps the animated crop aligned with the rest of the original image when pasting it back.
Blijf leren
Gerelateerde gidsen
Er zijn meer handleidingen voor dit onderwerp geselecteerd