Made the first commercial deepfake in Russia
Generated the face of the young
George Miloslavsky
for the "Sber"
Objective: To recreate a realistic face of young Leonid Kuravlev from the movie "Ivan Vasilievich Changes his profession" in a commercial for Sberbank
The client's goal: to create a wow-effect in the audience, to form an association with Sber as a high-tech corporation.
Our task: to create a high-quality generation in large resolution, using only archival film footage.
24 days
7,000 frames
Took us 2.5 minutes to generate the face for the video
Assembled from the original film to generate a face
What was done in the project
1. support for the production team in the selection of the actor.

2. organized presence on set: control of actor's turns, light, frame dynamics.

3. processed Soviet newsreel data for possible use in face generation.
4. Color correction of the materials was performed.

5. A pretrain model was created to increase the speed and quality of generation production.

6. Accelerated generation algorithm using intelligent equipment capacity allocation.
Made the world's first educational deepfake
We generated an expert head for the Gazprom Neft educational course
Task: For an educational course you need to generate the head of a recognizable expert. He cannot take part in the shooting. The understudy speaks for him
The client's task: to prepare educational content without the possibility of organizing it with an expert
Our task: to generate the expert's head in different angles and states
14 days
60,000 frames
Took to collect data with an expert from a 10-minute video
were processed by a neural network
Told about the process of making a deepfake video
Generated Travolta's head on the presenter's head and told how to do it
We did a head generation of a famous actor on the host, answered questions about deepfakes and revealed some of the secrets of video production using neural networks for the YouTube channel Technicolor
Told how you can make money on deepfaeks and brought the portrait to life
We made several generations of famous people's faces on the host, showed their new technology and talked about the future of deepfakes for the YouTube channel
Made on