Adaptive Reference Image Selection for HDR Artifact Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional HDR imaging techniques often result in artifacts due to the assumption that pixels in the EV0 image with incomplete information can be replaced by corresponding pixels from the EV- image, which is not always true, especially due to motion and occlusion, leading to unwanted artifacts in the resulting HDR image.
Innovation Solution
The adaptive reference image selection (ARIS) logic/module compares the number of potential artifact pixels in the EV0 and EV- images and selects one as the reference image based on a threshold value, determining the appropriate image registration technique to reduce or eliminate artifacts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If the EV0 image is used as the reference image for image registration and fusion, then the HDR image quality is improved, but artifacts are generated when pixels with incomplete information cannot be replaced by corresponding pixels from the EV- image
Solution Approach 1:
The patent dynamically selects the reference image between EV0 and EV- based on the presence of moving objects detected through ghost map analysis. This dynamic selection adapts to different scene conditions: using EV0 when no motion artifacts are present (maintaining high quality) and using EV- when motion artifacts are detected (eliminating artifacts), thus resolving the contradiction between image quality and artifact generation.
2Object-generated harmful factors
If the EV- image is used as the reference image to avoid artifacts, then artifact generation is reduced, but the HDR image quality deteriorates due to incomplete information in underexposed regions
Solution Approach 1:
The system dynamically switches between EV0 and EV- as reference images based on motion detection. When motion is detected, EV- is selected to avoid artifacts; when no motion is detected, EV0 is selected to maintain high image quality. This dynamic approach ensures optimal HDR image quality while minimizing artifacts in different scenarios.
Solution Approach 2:
The ghost map serves as an intermediary element that analyzes pixel consistency between EV0 and EV- images to detect moving objects. Based on the ghost map analysis results, the system determines which image to use as reference, thus mediating between the conflicting requirements of image quality and artifact reduction.
3Productivity
If image registration and fusion techniques are performed using the EV0 image as reference, then processing speed is improved, but computational resources are wasted on de-ghosting algorithms when artifacts are present
Solution Approach 1:
The patent performs preliminary ghost map analysis and reference image selection before the main HDR fusion process. By detecting potential motion artifacts in advance and selecting the appropriate reference image (EV0 or EV-), the system prevents artifact generation at the source, eliminating the need for subsequent de-ghosting algorithms and conserving computational resources.
Solution Approach 2:
The system uses feedback from ghost map analysis to determine the optimal reference image selection. The ghost map provides information about motion artifacts, which feeds back into the reference image selection decision, enabling the system to adaptively choose between EV0 and EV- to avoid artifacts and reduce the need for post-processing de-ghosting operations.
Data Source
AI summary
Techniques of reducing or eliminating artifact pixels in high dynamic range (HDR) imaging are described. One embodiment includes obtaining a first image of a scene at a first time with first exposure settings and obtaining a second image of the scene at a second time with second exposure settings that differ from the first exposure settings. The obtained images may be downsampled. The images may be compared to each other to assist with determining a number of potential artifact pixels in the scene. Depending on a relationship between the number of potential artifact pixels and a threshold value, the first image or second image may be selected as a reference image for registering the images with each other. A type of registration performed between the images may depend on which of the two images is the selected reference image. The registered images may be used to generate an HDR image.