Depth-Based Image Fusion to Suppress NIR See-Through
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image fusion techniques fail to effectively address the see-through effect caused by combining near-infrared (NIR) and RGB images, particularly in close-up shots, leading to privacy concerns due to materials appearing transparent under NIR light.
Innovation Solution
The method employs depth information to assign fusion weights, reducing the see-through effect by selectively combining NIR and RGB image layers based on depth values, using a pyramid approach to separate and combine images at different resolution levels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If NIR and RGB images are fused to enhance details and provide different sensory information, then image information completeness is improved, but see-through effect occurs causing privacy concerns
Solution Approach 1:
The patent applies different fusion weights to different spatial regions of the image based on depth information. Close-up regions (foreground objects) use lower NIR fusion weights to suppress see-through effect, while background regions use higher NIR fusion weights to maintain detail enhancement. This local differentiation resolves the contradiction by allowing NIR information to be useful where appropriate (background) while minimizing harm where critical (foreground).
Solution Approach 2:
The patent dynamically adjusts the fusion weight parameter based on depth distance. By changing the fusion weight parameter according to the depth map, the system optimizes the balance between information completeness and see-through prevention for each spatial location, transforming a static fusion approach into an adaptive one that resolves the contradiction contextually.
2Object-affected harmful factors
If depth-based fusion weights are applied to suppress see-through effect in close-up shots, then privacy protection is improved, but image processing complexity increases
Solution Approach 1:
The patent segments the image into different depth layers using a depth map, separating foreground (close-up) regions from background regions. This segmentation allows selective application of fusion weights, enabling privacy protection in close-up shots while avoiding unnecessary processing complexity in background regions where see-through effect is less problematic.
Solution Approach 2:
The patent introduces a depth map as an intermediary element that guides the fusion process. Rather than directly analyzing material properties or implementing complex see-through detection, the depth map serves as a mediator that provides simple depth-based cues to control fusion weights, reducing processing complexity while maintaining privacy protection effectiveness.
3Object-affected harmful factors
If NIR fusion weight is reduced for close-up objects to suppress see-through effect, then see-through suppression is improved, but detail enhancement capability deteriorates
Solution Approach 1:
The patent applies local quality differentiation by using low NIR fusion weights specifically for close-up objects to suppress see-through effect, while simultaneously using high NIR fusion weights for background regions to maintain detail enhancement. This spatially-varying approach ensures that see-through suppression and detail enhancement are both optimized in their respective regions without mutual interference.
Data Source
AI summary
This disclosure is directed to image fusion. A computer system obtains a near infrared (NIR) image and an RGB image of a scene. A first NIR image layer is generated from the NIR image. A first RGB image layer and a second RGB image layer are generated from the RGB image. The first NIR image layer and first RGB image layer have a first resolution. A depth map is also generated and has the first resolution. Each pixel of the first NIR image layer and a corresponding pixel of the first RGB image layer are combined based on a respective weight to generate a first combined image layer used to reconstruct a fused image. For each pair of pixels of the first NIR and RGB layers, the respective weight is determined based on a depth value of a respective pixel of the depth map and a predefined cutoff depth.