Welcome to
Mnemosis is a 4D solution for a markerless surface motion capture transfer. Record yourself and use a finished photorealistic scan within minutes, complete with automatically generated skeleton and a RIG, able to deform, greatly cutting the time and cost of production. Our solutions are executed by utilizing state-of-the-art optical measurement methods, algorithms of 2D, 3D and 4D data processing, as well as neural networks and machine learning.
4D Body
4D Face
4D
Body
4D model is a scan of a person in motion. By utilizing up to a hundred 4K cameras simultaneously, we are creating point cloud with a resolution of 4 points per square millimeter with accuracy below 0.3 mm. Automatic processing software registers photorealistic geometry and texture that includes skin, behaviour of clothing, while generating RIG, consisting of mesh, blendshapes and a skeleton.
RIG
Real actor
This is a recording of a person in motion taken during the process of scanning.
Cloud of points
Cloud of points is the first step of acquisition. The higher the resolution the more points are in the cloud.
Skeleton
Calculated automatically from CoP skeleton is a simplyfied representation of a human motion, calculated from every frame separately.
RIG
Mesh
Automatically generated geometry provides a sufficient level of details on the surface.
Blendshapes
Automatically generated key blend-shapes allowing for accurate representation of body geometry deformations.
Every next layer is based on the previous one and each of them can make a workflow of animator, modeller or rigger significantly faster and easier. Consider at which stage we can be helpful to you.
Interact with our 4D Body model .
4D
Face
We are using a separate system for a marker or markerless face scanning in order to capture every facial expression in detail. As a result, we achieve life-like 4D face model that includes automatically generated RIG consisting of mesh and blendshapes.
RIG
Real actor
This is the real actor’s face recorded during the process of scanning.
Cloud of points
Cloud of points is the first step of acquisition. The higher the resolution the more points are in the cloud.
Mesh
Mesh gives a certain level of geometry details that can later be enriched with textures and blend-shapes.
RIG
Blendshapes
Automatically generated key blendshapes allow for smooth transitions between face expressions.
High precision of scanning combined with a structured light and our algorithms enables us to capture every micro-expression with a photorealistic quality.
Interact with our 4D Face model
As one of the few on the market, we have direct access to a professional database of over 300,000 actors and models through our sister agency, Stars Impresariat. This guarantees not only the highest quality and repeatability of shots but also full compliance with copyright law and personal data protection. We provide data that is ethically sourced and free from legal encumbrances.
Use Cases:
Computer Vision: Training models for character, gesture, and emotion recognition.
Metaverse & VR/AR: Creating realistic avatars and digital humans.
VFX & Gaming: Generating digital doubles and NPC characters.
Robotics: Simulating human-robot interactions.
Research and Development: Analyzing biomechanics and human behaviors.
What makes us unique?
High-Fidelity: Data in 4K/8K resolution, precise 3D scans, and motion capture.
Diversity: A wide spectrum of age, gender, and ethnicity to reduce bias in AI models.
Structure and Annotations: Data is fully described and ready for implementation in machine learning pipelines.
Full Legal Control: All models sign appropriate consent forms for the use of their image for commercial and AI training purposes.
Lets Create something amazing
31-056 Kraków
Poland