pixel art face from a video game and a realistic human face generated from it

Face Depixelizer

PROTIP: Press 'i' to view the image gallery, 'v' to view the video gallery, or 'r' to view a random entry.

This submission is currently being researched & evaluated!

You can help confirm this entry by contributing facts, media, and other evidence of notability and mutation.

About

Face Depixelizer is an application that generates high-resolution human portraits from low-resolution input images. The application briefly gained popularity online in June 2020 when it was used to upscale pixelated portraits of characters from various video game franchises.

History

On June 19th, 2020, Russian developer Bomze (Denis Malimonov) uploaded Face Depixelizer to GitHub.[1] The application generates high-resolution human portraits based on a low-resolution input image. On the same day, Bomze tweeted[2] about the application, with the tweet gaining over 4,200 retweets and 10,600 likes (shown below).


Bomze @tg_bomze Face Depixelizer Given a low-resolution input image, model generates high-resolution images that are perceptually realistic and downscale correctly. GitHub: github.com/tg-bomze/Face-. Colab: colab.research.google.com/github/tg-bomz.. P.S. Colab is based on the github.com/adamian98/pulse input downscale GIF 5:56 PM - Jun 19, 2020 - Twitter Web App Input Image

Features

Face Depixelizer is based on "PULSE: Self-Supervised Photo Upsampling via Latent Space Exploration of Generative Models"[3] repository.[1] Given a low-resolution input image, Face Depixelizer searches the outputs of a generative model for high-resolution images that are perceivably realistic and downscaled correctly.

Highlights

Following the June 19th, 2020, announcement tweet, on June 20th, Twitter[4] user @Rob_Milliken inquired whether the application could be used on 8-bit video game characters, with the developer responding[5] with an upscaled image of the Wolfenstein main protagonist B.J. Blazkowicz (shown below). The tweet received over 2,700 retweets and 12,600 likes in two days.


Original Result 200 200 400 400 600 600 800 800 1000 1000 250 500 750 1000 250 500 750 1000

Starting on that day, multiple users posted result images of video game characters processed by Face Depixelizer. For example, an image of Cacodemon posted by @papaabar[6] gained over 300 retweets and 2,400 likes in two days (shown below, left). An image of Creeper posted by @jonathanfly[7] gained over 240 retweets and 1,700 likes in the same period (shown below, right).


Original Result 200 200 400 400 600 600 800 800 1000 1000 250 500 750 1000 250 500 750 1000

In the following days, several online outlets reported on the app, including articles by Kotaku[8] and PetaPixel.[9]

Various Examples


Original Result 200 200 400 400 - 600 600 - 800 800 1000 1000 250 500 750 1000 250 500 750 1000
Original Result 200 - 200 400 400 600 600 800 800 1000 1000 250 500 750 1000 250 500 750 1000 Original Result 200 200 400 400 600 600 800 800 1000 1000 250 500 750 1000 250 500 750 1000

Search Interest

External References

Recent Videos

There are no videos currently available.

Recent Images 24 total


Top Comments


+ Add a Comment

Comments (10)


Display Comments

Add a Comment


Word Up! You must login or signup first!