Stable Diffusion- Applying ControlNet to Character Design - Part 1

Overview

Save Big on Coursera Plus. 7,000+ courses at $160 off. Limited Time Only!

Grab it

Explore the application of ControlNet to character design in this comprehensive tutorial video. Learn about various ControlNet models, including OpenPose, Depth, Segmentation, Normal pass, Canny edge detect, HED, MLSD, and Scribble. Dive into pose editing techniques using OpenPose Editor and 3D posing in Blender. Master the art of combining multiple ControlNets, brainstorming designs, and refining character concepts through inpainting, photobashing, and compositional enhancements. Discover advanced techniques for matching color tones, adjusting values, and adding finishing touches like lighting, bloom, and god rays. Perfect for digital artists and concept designers looking to leverage AI-assisted tools in their creative workflow.

Syllabus

Intro
ControlNet Overview
List of necessary tools
Part 1 ControlNet models
OpenPose model
Depth, Segmentation, Normal pass explanation
Depth pass
Segmentation pass
Normal pass
Canny edge detect
HED
MLSD
Scribble
Part 2 Pose Editors
OpenPose Editor
Detect pose from image
Testing out the open pose skeleton
Creating the second template skeleton
Testing out the second skeleton
Creating the third template skeleton
Posing the fourth template skeleton
Combining all templates into a reference sheet
Batch sourcing concepts
Posing in 3D Blender
Blender Introduction
How to pose the rig
Pose fundamentals
Starting to pose
Saving pose to pose library
Balancing the pelvis with the torso
Positioning the head
Moving the legs and feet into position
Positioning the arms
Basic gesture of the hand
Applying gesture in 3D
Setting up the camera
Rendering passes
Multiple ControlNets
Brainstorming base designs
Selecting the base design
Double-checking the effect of multiple controlnets
Cleaning the base image first pass
Inpainting the face first pass
Rendering new hand passes
Generating new hands using blender passes
Photobashing hands
Sanity check to see if limbs match up with image
Content aware fill
Technique to match color/skin tones on photobash assets
Inpainting the hands
Creating an inpainting model
Starting final touches
Adding compositional elements
Generating first compositional element
Generating second compositional element
Upscaling/inpainting the face second pass
Final face adjustments
Adjusting values
Lighting the scene
Adding bloom
Camera raw filter
Light rays/god rays
Wrapping up