Image Generation TRIZ Cases
- 2D Floor Plan Generation From 3D Primitive Projection
Camera images and ML-extracted primitives generate accurate 2D room plans without depth sensors or time-consuming manual input.
- Point Cloud Splatting for Low-Bandwidth 3D Streaming
Adapted Gaussian splats replace redundant point sets to cut 3D streaming data while preserving visual fidelity for rendering.
- 3D Industrial Equipment Rendering Workflow for Format Conversion
A RaaS workflow converts industrial equipment 3D model data into metaverse-ready renderings with less manual format handling.
- Composite Image Rendering with Alpha Layers for Editable Objects
Generative image objects with alpha transparency can be edited, moved, and regenerated independently instead of reworking the full image.
- 2D Image to 3D Virtual Object for Interactive Shared Viewing
AI turns a 2D image into a manipulable 3D virtual object, giving users multiple views and real-time rotation in shared sessions.
- Floorplan Georeferencing From Blueprint Contours and 3D Maps
Machine learning separates floor and elevation sheets, extracts contours, and aligns them with 3D maps to locate buildings faster.
- Ray Tracing BVH Traversal Order for Grouped Rays
Grouped-ray BVH traversal reorders child node visits to cut memory bandwidth use and speed ray tracing in complex scenes.
- Precomputed Mesh Topology for Lower-Cost Real-Time Ray Tracing
Precomputed mesh groups and traversal likelihood metrics cut acceleration-structure cost and memory while keeping ray testing efficient.
- 3D Splat Generation Using Loss-Weighted Image Sampling
Loss-weighted image sampling focuses splat training on high-loss views to cut 3D model file size and wasted iterations.
- Low-Resolution BVH Traversal for Lower Ray Tracing Cost
A conservative low-resolution BVH cuts ray traversal cost by culling nodes early while preserving access to full-resolution bounding data.