Adobe filed 1 patent in Graphics.
The application covers an AI-powered progressive mesh system that uses machine learning to upscale coarse 3D polygon meshes in real-time, then refines them with residual data for high-fidelity rendering. This technology addresses the challenge of efficiently generating detailed 3D models by combining neural network processing with traditional mesh refinement techniques.
Adobe's single Graphics patent tackles the challenge of delivering detailed 3D models over networks without forcing users to wait for complete data transfer. The system trains a machine learning model on specific polygon mesh datasets, allowing it to intelligently predict how a low-resolution mesh should look when upscaled rather than relying on generic interpolation. When a coarse mesh arrives, the ML component immediately generates a higher-fidelity version for rendering, while residual correction data transmitted afterward restores the original accuracy. This progressive approach means viewers see plausible high-quality geometry almost instantly, with the final refinements arriving seamlessly in the background.
All data sourced from USPTO patent filings. Google Patents may take several weeks to index recent publications. If a link is unavailable, search for the patent number at USPTO Patent Public Search.