This free AI can control anyone's face

How to Use Facepook: A Guide to AI Photo Manipulation

Introduction to Facepook

• The video introduces two free and open-source AI tools for manipulating facial photos, including head movement, eye control, and lip editing.

• The first tool discussed is called Facepook, which operates online via a Hugging Face space. A link will be provided in the description.

Features of Facepook

• Users can upload images in JPEG or WEBP formats; PNG files may not work effectively.

• The interface allows users to hover over facial features (e.g., eyebrows, eyes) to highlight them for manipulation.

• Users can drag their mouse to move the subject's head based on cursor position.

Manipulating Facial Expressions

• Eye control is demonstrated by dragging up or down to open or close the eyes.

• Eyebrows can also be adjusted by dragging up or down for raising or lowering expressions.

• Mouth movements are controlled similarly; users can create various expressions by dragging left/right and up/down.

Additional Controls

• Users cannot change gaze direction (e.g., cross-eyed), nor manipulate features like tongues; only smiles are adjustable.

• To tilt the face, users must hover over a circular area around the face and drag accordingly.

Downloading and Local Installation

• After editing, images can be downloaded directly from the platform.

• Instructions for local installation of Facepook are available on GitHub; links will be provided in the description.

Installing Anaconda for Local Use

Setting Up Your Environment

• To install locally, users need Anaconda. The video guides through installing Miniconda as a lightweight alternative.

• Miniconda avoids unnecessary packages that come with full Anaconda installations, saving disk space and time.

Installation Steps

• Users should download the latest Miniconda installer suitable for their operating system (Windows recommended).

• It’s advised to use Python 3.11 instead of 3.12 due to compatibility issues with many open-source AI tools.

How to Install and Set Up Anaconda and Git

Installing Anaconda

• After completing the installation of Anaconda, it's important to check if it is recognized in the command prompt by typing cond --version. If not recognized, it indicates that Anaconda hasn't been added to the system path yet.

• To add Anaconda to the system path, search for "Edit the system environment variables," click on it, then navigate to Environment Variables. Locate 'Path', click edit, and add the path where Anaconda is installed (e.g., C:ProgramDataMiniconda).

• Once you have copied the correct path from your installation directory (like C:ProgramDataMinicondaScripts), paste it into the Path variable in Environment Variables.

• After updating the path, reopen Command Prompt and verify again with cond --version, which should now display a version number indicating successful installation.

Creating a Virtual Environment

• Create a virtual environment named "face pooke" using Python 3.10 by executing: conda create -n face pooke python=3.10. Confirm with 'y' when prompted.

• Activate your newly created environment with: conda activate face pooke. The command prompt will show (face pooke) indicating that you are now working within this environment.

Installing Git

• If Git is not already installed, download it from its official site. Choose the appropriate installer for your operating system (Windows in this case).

• Follow through with default settings during installation; clicking 'Next' until completion ensures all necessary components are installed without customization.

Cloning a Repository

• Use Git to clone a repository by typing: git clone <repository-url>. This action copies all files from the specified repository into your active virtual environment ("face pooke").

• Navigate into the cloned folder using: cd Facebook, ensuring you respect capitalization as it's case-sensitive.

Installing Dependencies

• Install required packages listed in requirements.txt by running: pip install -r requirements.txt. This installs all dependencies needed for proper functionality of your project.

• Acknowledge that installing these packages can take time depending on internet speed; some packages like PyTorch may be large (around 2.5 GB).

Additional Tools Mentioned

• The video introduces an AI tool called Chat LLM by Abacus AI, which integrates various AI models including GPT and Claudsonit 3.5 for enhanced functionalities such as generating images or videos based on prompts.

How to Install Face Pooke Locally

Installation Overview

• The installation process for Face Pooke involves downloading various packages listed in the requirements file, which may take some time due to the number of dependencies.

• After installing dependencies without errors, the next step is to install front-end dependencies, indicating that the installation process is lengthy and complex.

Setting Up Environment

• Users need to change directories to the client folder using CD client before proceeding with further installations.

• The installation of Bun is required; users can find instructions on bun.sh. This package works across Windows, Mac OS, and Linux platforms.

Activating Virtual Environment

• To activate the virtual environment, users must navigate back to their CMD in the client folder and execute cond activate face pooke.

• Once activated, users should run bun install to complete front-end setup before building it.

Running the Application

• After building the front end successfully, users need to run Python app.py, ensuring they are in the correct directory (face pooke).

• If it's a first-time run of app.py, additional large dependencies will be installed which may take considerable time.

Accessing Interface

• Upon successful execution, a message indicates that the application is running on HTTP at Local Host:8080. Users can then access this URL in their browser.

Exploring Expression Editor

Introduction to Expression Editor

• Expression Editor is introduced as a similar tool to Face Pooke but utilizes sliders instead of mouse dragging for adjusting facial expressions.

Functionality Demonstration

• Users can upload images (like Elon Musk's), and manipulate settings such as head rotation using sliders for precise control over expressions.

Slider Adjustments

• Various sliders allow adjustments for head tilt and eye movements. For instance:

• Head rotation up/down and left/right.

• Eyebrow adjustments impact expression significantly.

Eye Control Features

• Sliders control blinking and winking effects:

• Adjusting blink slider closes or opens eyes wider.

• Winking effect varies based on slider position.

Mouth Expressions

• The mouth expression can also be manipulated through sliders:

• Different positions create distinct expressions like smiles or frowns.

Expression Editing Tutorial

Introduction to Expression Editor

• The tutorial begins with a demonstration of the "woo slider," showcasing how dragging it left or right alters facial expressions minimally, creating a "woo-shaped" mouth when moved to the right.

Adjusting Facial Features

• The presenter resets settings and adjusts eye features, including eyebrows and blinking. However, limitations are noted as certain expressions like cross-eyed or tongue-out cannot be achieved.

Exploring Mouth Expressions

• Further adjustments are made to the mouth settings, illustrating various expressions by moving sliders. This highlights the flexibility of expression editing tools similar to Face Pooke.

Tool Capabilities and Recommendations

• The tutorial emphasizes that both tools discussed are based on Live Portrait technology, which offers more advanced capabilities for controlling expressions using reference videos.

Conclusion and Engagement