“Deep-Live-Cam” is an open-source software project that enables a real-time face swap in video contents using just a single image. It has already gathered the attention of both hobbyists and professionals in the industry of AI-generated media, all while providing an easy-to-use solution for video deepfake creation and animation of custom characters.
GIF Source: GitHub / hacksider / Deep-Live-Cam
The project, published on GitHub with the name “hacksider,” has already received more than 17,000 stars, showing strong interest and support from the community. For this purpose, “Deep-Live-Cam” utilizes advanced AI models like GFPGAN and inswapper. This software is designed to run on many hardware configurations, including CPUs, GPUs, special chips such as Apple Silicon, and Intel OpenVINO.
Although the software is incredibly useful, the project authors have taken steps to avoid any negative uses. This includes blocking out inappropriate and illegal content as much as possible, by settings which restrict the processing of data into the software.
Image Source: GitHub / hacksider / Deep-Live-Cam
The program prompts users to receive direct permissions from people involved, i.e., before editing a video to insert a face, to ensure everyone is in the loop on what is being done. Furthermore, deepfake content must be watermarked explicitly before being uploaded.
“Deep-Live-Cam” is free software under the license AGPL-3.0, freely usable and modifiable under the terms and conditions. The project is currently in a stage of active development, and all new improvements can be tracked and contributed to by means of experimental branches available on GitHub.
More detailed information can be found on the project page on the official GitHub site, including installation and usage instructions.
Leave A Comment