How it all began
As imaging experts, we witnessed the rising popularity of digital 3D interactive experiences across various industries and verticals, fueled by the ubiquity of powerful computing devices and ever-increasing communication throughput. While exploring this emerging ecosystem, it became clear to us that the growth of this new medium was being throttled by the availability of 3D content.
Photo-realistic 3D models are the new media type, analogous to image and video files commonly used in traditional 2D modalities. Yet, when we explored the existing methods for 3D-model creation, we realized that the current, labor-intensive approaches are not sustainable and that a technological paradigm shift is necessary in order to truly solve the 3D content creation problem, at scale.
The intersection in time of several factors—significant growth in computing power, data transmission and storage capacity, and speed and maturation of deep-learning technology, all accompanied by the huge amount of existing 2D content—has created a unique opportunity in which the team could utilize the best bits and pieces of science and develop the leading technology that can automatically create high-quality 3D models from sparse data like text or a few images.