#twin3dpro
**Unveiling the Future of Digital Humans: Insights from Twin3D**
Alexey Belokopytov, from Twin3D, with co-speaker Yefim, presents their work on creating photorealistic human avatars, emphasizing its utility across various sectors such as film, gaming, banking, and advertising. The company’s products include full-body scanning, facial scanning, and on-location services. They foresee digital humans being used across industries in the future, potentially creating unique virtual ecosystems that could merge into a unified matrix in the next 10 to 15 years.
Three main challenges are identified: the cost, complexity, and duration of creating avatars; the limited mass-market appeal of avatars; and diverse industry requirements. Streamlining the creation process and creating industry-specific applications are key to success.
Yefim outlines the complex 3D scanning technology used to capture facial details, emphasizing the importance of high-quality results. They discuss efforts to improve the image quality and the development of algorithms for realistic textures and clean geometry. The ultimate goal is to create an automatic, fast, and intelligently driven photorealistic 3D scanning process.
The cost and duration of the scanning process vary according to project requirements, ranging from a few hours for a basic scan to several months for a complete character scan, including animation. The prices start from 20,000 rubles.
The speakers acknowledge that the use of 3D avatars is becoming prominent across various industries. However, they note the limited game development in Russia, the expense and complexity of devices, and the need to automate avatar creation for broader adoption.
They note the growing field of 3D computer vision, which aids the advancement of the field, and ongoing efforts to improve scanning processes. Challenges include capturing live subjects, as even slight movements can degrade the model quality. In contrast, static objects, like sculptures, are simpler to scan. Mobile phone-based scanning technology still has limitations, and in-depth scanning requires manual scanners or photogrammetry with good lighting conditions.
While the primary focus is on complex challenges of human subject scanning, the technology can be applied to scanning static objects for specific film projects. Further advancements in the field are expected to be showcased in the next conference.