Lighting Compositing via Multi-Image Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional lighting techniques for photography require extensive coordination and timing, and existing image processing systems struggle to effectively manage and combine images with varying light sources, leading to suboptimal results.
Innovation Solution
A method and system that combine multiple images of a scene, each exposed to ambient light and one or more discrete light sources, to form a composite image, allowing for the identification and modification of discrete light sources, including the assignment of control regions to each source for deletion or modification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If traditional lighting techniques are used to stage and capture lighting in one or two images, then the lighting effect can be achieved, but it requires large lights and an extreme amount of timing and coordination
Solution Approach 1:
The patent segments the lighting capture process by taking multiple separate images, each illuminated by a different light source, and then combining them computationally. This divides the complex single-shot lighting problem into simpler individual captures that can be processed separately and merged, eliminating the need for large lights and complex timing coordination.
Solution Approach 2:
The patent replaces the mechanical system of physical lighting coordination and timing with a computational approach. Instead of using large lights and coordinating their operation mechanically, the system uses image processing algorithms to combine multiple images captured under different lighting conditions, substituting computational complexity for mechanical complexity.
2Adaptability or versatility
If multiple images with varying light sources are captured and combined, then a composite image with multiple discrete light sources can be formed, but the process becomes more complex
Solution Approach 1:
The patent creates a computational model or representation of light sources by analyzing patterns in the captured images. It copies the essential characteristics of multiple light sources and their spatial relationships into a processed composite image, allowing the system to represent and manipulate complex lighting scenarios through image data rather than physical light sources.
Solution Approach 2:
The patent develops a universal image processing framework that can handle multiple types of light sources and lighting conditions through a single computational approach. The system is designed to be multi-functional, capable of processing images with various lighting configurations and producing consistent results across different scenarios, thereby managing complexity through generalization.
3Ease of manufacture
If traditional composite photography apparatus with polarized light is used, then background projection can be achieved, but it requires complex polarized light management
Solution Approach 1:
The patent replaces the optical-mechanical system of polarized light management with a computational image processing approach. Instead of using polarized filters and controlled light reflection, the system captures images under different lighting conditions and uses algorithms to separate and combine background and foreground elements, substituting computational processing for optical complexity.
Data Source
AI summary
A method and system. A base photographic image of a scene is combined with N additional photographic images of the scene to form a composite image including M discrete light sources (N≥2; M≥N). The scene in the base image is exposed to ambient light. The scene of the base image is exposed, in each of the N additional images, to the ambient light and to at least one discrete light source to which the base image is not exposed. The M discrete light sources in the composite image include the discrete light sources to which the scene is exposed in the N additional images. The composite image includes a region surrounding each discrete light source and has an area that correlates with an intensity of light from the discrete light source surrounded by the region. The intensity of light is the intensity at the discrete light source where the light is emitted.