Projects
The Technology Behind the Virtual Human "Drip": Achieving Speed & Flexibility
ZOZO Technologies announced their virtual human unit "Drip." Previously, we shared a discussion about the project's history and vision with Hao Li (Pinscreen, Inc.), a leader in the field of computer vision.
This time, we focus on the unique method of creating virtual humans that was mentioned in the discussion. We interviewed Yudai Tamamura and Yuki Ikegami from ZOZO Technologies MATRIX about their production method, which achieves video content production at around half the speed of traditional methods.
PROFILE
Yudai Tamamura
After working in game-related video production and 3DCG at a major IT venture, Yudai Tamamura joined ZOZO Technologies in 2020. He currently leads projects in the virtual fashion domain within the XR/AI team. His main tools are Houdini, Substance, ZBrush, and UE4.
PROFILE
Yuki Ikegami
After experiences as a planner and pattern maker at an apparel brand, Yuki Ikegami joined ZOZO Technologies in 2020. He is responsible for virtual fashion modeling in the XR/AI team. His main tools are CLO, Marvelous Designer, Substance, and ZBrush.
From Concept to Image
The creation of a post image begins with deciding on the content concept. Discussions include the virtual model's pose, the type of live-action background, and the clothes to be worn. Then, artists with specialized skills in various fields (such as modeling, animation, compositing) work simultaneously.① Concept Planning for Post Content
- Model's posing and expressions
- Live-action background
- Clothing, shoes, accessories, etc.
② Shooting Live-Action Background & HDRI Images
- Shooting live-action backgrounds to be combined with the virtual model
- Capturing 360-degree images to replicate the lighting environment of the shooting location in CG, converting them to HDRI data to handle in a CG environment
③ 3D Modeling of Clothes, Shoes, Accessories
- Using 3DCG production software like Marvelous Designer and Maya
- Specialized 3D modelers handle the production
④ Compositing CG (Model) & Live-Action (Background)
In current virtual model/influencer production, it is common to save labor by using live-action for everything except the face, which is CG. However, Drip's model is entirely CG. To achieve quick production of such full-body CG virtual models, the distinctive technology used in step ④—the compositing of CG (model) and live-action (background)—is crucial.
This article is for members only.
Please register to read the rest of the article.
What you can do with a membership
- Read members-only articles
and use text-to-speech. - Unlimited article favourites
and browsing history. - Attend members-only events.
- Get the latest information
with our email newsletter.
How was the article?
Share
Related Articles
.jpg?w=600&fm=webp&dpr=2)
