Skip to content

Collection of scripts for creating synthetic mouth and eye training images.

Notifications You must be signed in to change notification settings

RamesTheGeneric/Datagen-Babble

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

ProjectBabble: Data generation

A Mouth tracking system for VR, 100% Opensource!

Currently the most of the generated blenshapes are from ARKit! The model outputs 45 blenshapes of which most cross over with SRanipals blendshapes.

Description

Contains all of the tools and source files for creating the synthetic face dataset for training. The current scripts in the V4.1 folder and separated in two different folders. The "Blender" folder contains scripts intended to be run in the provided Blender scenes. The "Processing" folder contains scripts for extracing shape values from the images into a csv file or in the console for inspection.

One thing to note

  1. At the moment, this only contains scripts and scenes for generaing lower face blendshape datasets. Support for the full face and landmarks is planned for the future.

Usage

DatasetGeneratorV4.1:

Generates the randomized shapes to be rendered.

  1. Open one of the blender scenes and import the script from the scripting tab.
  2. Adjust the amount of desired images with the frame range in the Output Properties tab on the right
  3. Change the "mesh" variable with the name of the corrisponding avatar (ex. "Babble_CA_M").
  4. (Optional) Toggle the Blender system console by going Window ==> Toggle System Console. Useful for seeing the generation progress
  5. Press the play button at the top of the blender text editor and wait for the script to finish.

Note: The script can be stopped at any time by pressing Ctrl + C in the console. Once saved, the file can be rendered at any time by rendering as an animation.

csv_from_images:

Extracts blendshape values encoded in the rendered images and write them to a csv file.

  1. Install dependences: pip install Pillow
  2. Edit the "path_to_files" variable to the directory with the rendered images.
  3. Run the script

Outputs

Blendshapes
cheekPuffLeft
cheekPuffRight
cheekSuckLeft
cheekSuckRight
jawOpen
jawForward
jawLeft
jawRight
noseSneerLeft
noseSnerrRight
mouthFunnel
mouthPucker
mouthLeft
mouthRight
mouthRollUpper
mouthRollLower
mouthShrugUpper
mouthShrugLower
mouthClose
mouthSmileLeft
mouthSmileRight
mouthFrownLeft
mouthFrownRight
mouthDimpleLeft
mouthDimpleRight
mouthUpperUpLeft
mouthUpperUpRight
mouthLowerDownLeft
mouthLowerDownRight
mouthPressLeft
mouthPressRight
mouthStretchLeft
mouthStretchRight
tongueOut
tongueUp
tongueDown
tongueLeft
tongueRight
tongueRoll
tongueBendDown
tongueCurlUp
tongueSquish
tongueFlat
tongueTwistLeft
tognueTwistRight

About

Collection of scripts for creating synthetic mouth and eye training images.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages