The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →Disney Research did publish work on high-resolution neural face swapping, a visual-effects technique that could transfer one person’s facial appearance onto another performer. But the paper does not show that Disney used the system to resurrect deceased actors in released films. Calling it “facial recognition” is also misleading: the research is about changing a face’s appearance, not primarily identifying who a person is.
Where the “bring back dead actors” claim came from
A July 2, 2020, Tech Times article framed Disney’s research as the company using deepfakes and facial-recognition technology to bring back dead actors. The article discusses hypothetical uses, but it does not document a particular Disney production or name a deceased performer recreated with the system. Read the Tech Times article.
| # | Preview | Product | Price | |
|---|---|---|---|---|
| 1 |
|
Deepfakes (In the News: Need to Know Set Two) | $8.99 | Buy on Amazon |
| 2 |
|
Deepfakes: The Coming Infocalypse | $7.73 | Buy on Amazon |
| 3 |
|
Deepfake | $18.13 | Buy on Amazon |
| 4 |
|
DeepFake Technology: Complete Guide to Deepfakes, Politics and Social Media | $3.95 | Buy on Amazon |
| 5 |
|
The New Age of Sexism: How AI and Gender Bias Are Reinventing Misogyny | $14.49 | Buy on Amazon |
The underlying work was a Disney Research Studios paper, “High-Resolution Neural Face Swapping for Visual Effects,” published June 29, 2020, and presented at the 2020 Eurographics Symposium on Rendering. Its authors proposed a visual-effects method; a research paper describing a capability is not, by itself, evidence that a studio used it in a released film. Disney Research Studios’ paper summary.
What Disney’s face-swapping research does
The paper describes an automated neural system for transferring a person’s facial appearance onto another person in still images and video. The intended result is not just a pasted-on still face: it aims to follow the target performer’s expressions and remain visually consistent from frame to frame.
#1 Best Overall
- High-resolution output: The authors describe megapixel-resolution results, a useful target for detailed visual-effects work.
- Temporal stability: The system is designed to reduce flicker and instability as a face changes across video frames.
- Multi-way training: Its network is intended to handle more than a simple, fixed two-person swap.
- Compositing: The approach includes methods intended to preserve lighting, contrast, and the broader, low-frequency appearance of the face.
- Landmark refinement: Facial landmarks are stabilized to help the result track the target face more consistently.
The paper reports a proposed method and research results, not a guarantee that every shot will look convincing or a description of a complete production pipeline. Read the full paper.
Is it a deepfake or facial recognition?
“Deepfake” is a broad, understandable label
In ordinary usage, “deepfake” can refer broadly to media altered or generated with machine-learning methods. Disney’s work is a neural face-swapping technique, so the label is understandable at a high level. The paper, however, presents it as visual-effects research, not as a system for fraud, impersonation, or disinformation.
“Facial recognition” describes a different task
Facial recognition usually means identifying or verifying a person from their face. This paper is principally about transferring facial appearance and compositing it into another performance. It uses facial information to produce an image, not chiefly to answer “Who is this person?” Calling it facial recognition blurs that distinction.
What “bringing back” a performer could mean
A face swap is only one possible part of a screen performance. It does not independently recreate a performer’s body, voice, personality, or acting choices. The distinction matters because a new image of a familiar face can represent several very different production approaches.
Free tools Windows power users keep installed
One-click scans. No signup required.
| Approach | What it involves | What the Disney paper establishes |
|---|---|---|
| Face replacement | A living performer supplies movement and timing; software transfers another person’s facial appearance onto that performance. | This is the closest match to the paper’s described face-swapping method. |
| Digital double or synthetic performance | Computer-generated work may recreate more of a person’s face, body, movement, or expression. | The paper does not establish a complete synthetic actor or autonomous performance. |
| Archival compositing | Existing footage of a performer is edited into a new production. | The paper is not evidence that Disney used archival footage in this way. |
Even in a face replacement, the living performer may supply body language, head position, eye direction, timing, and emotional movement. The result can be a new performance wearing a familiar face, rather than the deceased performer’s original performance continued unchanged.
How a face-swap shot can be made
The following is a plain-language explanation of a typical face-swapping workflow, not a claim that Disney publicly disclosed every production step it used. The paper’s stated research concerns the neural swap and its visual stability.
Rank #3
- Choose the target performance. A performer is filmed, or suitable footage is selected, to provide the movement, framing, and expressions for the shot.
- Align the face. Facial features and landmarks are located so the generated appearance can follow the target face’s position and movement.
- Gather identity references. Images or footage of the face to be transferred provide visual information about its appearance. The amount and quality of suitable material can vary.
- Generate and transfer the appearance. The system produces the source identity’s facial appearance while adapting it to the target performance.
- Blend and stabilize. The result is composited into the image and refined across frames to reduce visible edges, changes in appearance, and jitter.
- Finish and review the shot. Artists may need to address color, lighting, hair, occlusions, and other details, then repair frames or shots where the result does not hold up.
Possible uses are not proof of deployment
A technique that transfers identity while retaining a target performance could, in principle, be useful for more than posthumous likeness work. Potential applications include de-aging, stunt doubles, reshoots, continuity adjustments, character transformations, or historical and archival reconstructions. Those are possible uses of the general capability, not evidence that Disney used this particular system in any named production.
Disney Research’s public listings include later work related to video face swapping, facial animation, re-aging, facial deformation transfer, geometry and texture, and performance capture. That points to continuing research in related areas; it does not establish that a technique became a commercial product or entered a specific Disney Studios production pipeline. See the listings for visual computing and capture.
Windows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallOutdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchWhy a convincing result is difficult
Lighting, motion, and camera angle
A generated face has to sit naturally in footage whose lighting, movement, and viewpoint may differ from the reference material. Disney supplemental material for later face-swapping work notes that mismatched lighting can make photorealistic swaps more difficult. Read the supplemental material.
Occlusion and difficult shots
Hair, hands, props, helmets, or microphones can cover parts of a face. Sharp turns, partial framing, motion blur, smoke, rain, and darkness can also obscure the features the system needs to follow. Large differences in facial structure between the source identity and target performer can make a natural-looking match harder.
Frame-to-frame artifacts and resolution
A face may flicker, drift with head movement, look too soft, or show conspicuous edges where it meets the surrounding image. A result that appears plausible in one frame can fail when watched in motion or inspected closely. Disney’s paper specifically addresses resolution, lighting and contrast, landmarks, and temporal consistency; those design goals should not be mistaken for a claim that all visual artifacts are solved.
Reference material and performance
Reference material may not cover all the expressions, angles, lighting conditions, ages, mouth shapes, or skin details a scene requires. And even a strong likeness cannot make the target performer reproduce the deceased actor’s exact mannerisms, voice, or choices. The technique transfers appearance; it does not recover an original performance from an image alone.
Digital likeness requires permission as well as technology
Technical feasibility does not settle whether a production may or should use someone’s likeness. Depending on the jurisdiction and agreements involved, a project may need to address estate or successor approval, contractual rights, publicity rights, union agreements, rights in archival footage, and approval of the portrayal’s context. The relevant permissions and rules vary; the existence of face-swapping software does not mean a deceased performer’s identity is free for any use.
Consent also raises a creative question that legal clearance alone may not answer: would the performer have accepted this role, portrayal, or context? A posthumous digital performance can affect the performer’s reputation and creative legacy, and it may change the work available to living performers. A face swap does not resolve those questions or provide a voice performance; voice work, dialogue rights, and lip synchronization would require separate treatment.
What the public evidence establishes
Disney Research published a real neural face-swapping method intended for visual effects. The cited paper and the 2020 Tech Times article do not identify a deceased actor, a film or episode, or a confirmed production deployment of that system. The most precise description is therefore that Disney researched a capability that could have applications involving digital likenesses—not that the available sources show Disney bringing dead actors back in released movies.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
Recommended Free Tools
