Robust viewpoint compositing for unconstrained image data
A machine-learned viewpoint synthesis model with generative embeddings and uncertainty modeling addresses variable lighting and transient occlusions, enhancing realism and efficiency in generating composite images from uncontrolled datasets.
JP2026116289APending Publication Date: 2026-07-09GOOGLE LLC
Patent Information
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- GOOGLE LLC
- Filing Date
- 2026-03-30
- Publication Date
- 2026-07-09
Smart Images
Figure 2026116289000001_ABST
Abstract
This invention provides a system and method for synthesizing novel viewpoints in complex scenes (e.g., outdoor scenes). [Solution] In some implementations, the system and method may include or use a machine learning model that can learn from unstructured and / or unconstrained collections of images, such as “wild” photographs. In particular, exemplary implementations of this disclosure can learn volume scene density and luminance represented by a machine learning model such as one or more multilayer perceptrons (MLPs).
Need to check novelty before this filing date? Find Prior Art