Artificial intelligence (AI) is revolutionizing design realism with its ability to simulate tree growth and shape, according to researchers from Purdue University’s Department of Computer Science and Institute for Digital Forestry, in collaboration with Sören Pirk at Kiel University in Germany. By developing new AI models inspired by DNA, the team led by Professor Bedrich Benes compressed the information required to encode tree form into a megabyte-sized neural model. These models, after training, generate complex tree models with detailed geometry outputting gigabytes of data.
Unlike the traditional approach to creating digital tree models, which involves complex simulation algorithms, the AI-based models can compress complex behavior into a small amount of data. This advancement in encoding tree development has significant applications in architecture, urban planning, gaming, and the entertainment industries, enhancing design realism.
Professor Benes, who has been working with AI models for nearly a decade, expressed optimism about the potential for AI to improve existing methods for digital tree models. The AI-generated growth models cover a range of tree species, including maple, oak, pine, and walnut, with and without leaves.
One of the challenges identified by the researchers is the difficulty in quickly creating 3D models of real trees using AI. While AI has successfully modeled 3D geometries unrelated to nature, it has struggled with natural phenomena. To overcome this deficiency, the team aims to separate a tree’s intrinsic properties from its environmental response by understanding how trees interact with their environment and the genetic characteristics encoded in their DNA.
The lack of training data describing the 3D geometry of real trees has been a hurdle in developing AI tree models. However, the researchers aim to address this limitation by using AI models to reconstruct 3D geometry data from actual trees, aligning with the goals of digital forestry.
Professor Benes clarified that their AI models simulate tree developmental algorithms rather than nature itself, stating that users can take a picture of a tree with their cellphone, and the AI will generate a 3D geometry that can be explored and manipulated digitally.
By revolutionizing design realism through AI, the researchers at Purdue University and Kiel University are paving the way for advancements in architecture, urban planning, gaming, and entertainment industries. With their innovative approach to simulating tree growth and shape, these AI models offer exciting possibilities for creating realistic digital representations of various tree species with high detail and complexity.
