#meshcapadecom
**Naureen Mahmood, the CEO of Meshcapade**, spoke about the company’s mission to revolutionize the creation, animation, and use of 3D avatars across various industries. She started by expressing her admiration for the stage she was presenting on, reminiscing about her childhood visits to planetariums and how they inspired her passion for 3D visualizations and astronomy.
She touched upon Owen’s talk and emphasized the importance of the regulations he discussed, highlighting Meshcapade’s origins as a spin-off from the Max Planck Institute in Germany. The company focuses on developing the future age of avatars, wherein digital humans will dominate e-commerce, telepresence, social networking, and gaming. Naureen spoke of the need for accurate digital doubles for online shopping, photorealistic avatars for virtual meetings, and fantasy characters for entertainment.
However, a significant gap remains in the market: a technology that can efficiently produce digital humans. Current methods to capture 3D motion and body shapes of real humans are challenging and time-consuming. Realistic characters seen in movies or games require months of work by teams of Hollywood creators and animators.
Meshcapade addresses these issues by leveraging computer vision, machine learning, and AI to develop a digital human layer vital for synthetic data e-commerce and the burgeoning metaverse. The company believes that digital humans will transform industries and even lead to the emergence of new sectors.
The technology Meshcapade is pioneering intends to make the creation of digital humans accessible to everyone, from artists and e-commerce businesses to those in the healthcare and fitness sectors. The primary challenge is not just acquiring vast amounts of 3D data but ensuring this data is in a consistent format, just like 2D images.
To solve this problem, Meshcapade developed “SMPL”, a patented generative body model. It captures the nuances of human movement, including facial details, expressions, and even soft tissue deformations in 3D. This model, trained on over a million 3D scans and motion capture frames, offers a consistent representation for training AI networks. Essentially, it converts human motion and form into mathematical equations.
Furthermore, the company has made the SMPL model free for academia and creative communities, believing in the power of collaboration to drive innovation. Using this model, one can train neural networks to generate 3D digital humans from various inputs like images, videos, or even text. This has made SMPL a standard in AI methods both in academia and the industry.
Meshcapade also introduced “Neural Avatars”, 3D models that leverage neural networks to depict the complexities of clothing and hair. Built upon the foundational SMPL model, these avatars look incredibly lifelike, replicating real human motions.
To conclude, Naureen emphasized the vast potential of their technology across different industries, noting the current fragmented approach to digital human creation. Meshcapade’s solution offers a consistent foundation for creating digital humans, bridging the gap between physical reality and digital realms, ensuring everyone has an equal footing in this digital revolution.