X-ray Image Sensor Noise Reduction via Hot Pixel Gradient Interpolation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In x-ray imaging, the absorption of x-ray photons in the image sensor leads to high noise, which is difficult to filter without degrading image resolution, and existing solutions like fiber optic plates are costly and increase system thickness and weight.
Innovation Solution
A method and device that identify 'hot' pixels affected by x-ray absorption, calculate directional gradients, select directions with lowest gradients, and determine replacement values based on neighborhood pixel values to reduce noise while maintaining image resolution.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If a fiber optic plate is placed between the scintillator and image sensor to absorb x-ray photons, then noise is reduced, but system thickness and weight increase
Solution Approach 1:
The patent replaces the mechanical fiber optic plate (FOP) with a software-based noise reduction algorithm that processes pixel values in the image sensor. Instead of using physical material to absorb x-ray photons, the system identifies hot pixels caused by x-ray absorption and replaces their values with interpolated values from neighboring pixels, thereby eliminating the need for additional physical components.
Solution Approach 2:
The patent creates a virtual copy of the noise reduction function that was previously performed by the physical FOP. By calculating replacement values based on neighborhood pixel data and directional gradients, the system replicates the noise filtering effect of the FOP through computational methods rather than physical absorption.
2Object-affected harmful factors
If a fiber optic plate is placed between the scintillator and image sensor to absorb x-ray photons, then noise is reduced, but system thickness increases
Solution Approach 1:
The patent replaces the mechanical fiber optic plate (FOP) with a software-based noise reduction algorithm that processes pixel values in the image sensor. Instead of using physical material to absorb x-ray photons, the system identifies hot pixels caused by x-ray absorption and replaces their values with interpolated values from neighboring pixels, thereby eliminating the need for additional physical components.
3Object-affected harmful factors
If noise filtering is applied to reduce noise from x-ray photon absorption, then noise is reduced, but image resolution deteriorates
Solution Approach 1:
The patent applies local quality by treating different regions of the image differently based on their characteristics. The algorithm identifies hot pixels through statistical analysis and applies replacement only to those specific locations, while leaving other pixels unchanged. This localized approach preserves image resolution in non-noisy regions while reducing noise only where necessary.
Solution Approach 2:
The patent uses dynamic gradient calculation to determine the appropriate replacement values for hot pixels. By calculating directional gradients in multiple directions and selecting the direction with the lowest gradient, the algorithm adapts to local image structures and preserves edges and features, thereby maintaining resolution while reducing noise.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The solution effectively reduces noise in x-ray images while preserving high resolution, similar to using a fiber optic plate, but with a lightweight and cost-effective software-based approach.
Implementation Method 1
the scintillator can only absorb a certain fraction of all x-ray photons that enter the scintillator's surface
Implementation Method 2
Some of these x-ray photons are absorbed in the image sensor, which typically results in high noise affecting one or several pixels
Data Source
AI summary
There is provided a method and corresponding device for noise generated by absorption of x-ray photons in an image sensor having a number of pixels. The method is based on identifying (S1) so-called hot pixels affected by absorption of x-rays, and calculating (S2), for each hot pixel, directional gradients in a number of different directions in a pixel neighborhood of the hot pixel. The method further involves selecting (S3), for each hot pixel, at least one direction among those directions having lowest gradient, and determining (S4), for each hot pixel, a replacement value based on neighborhood pixel values in the selected direction(s). For each hot pixel, the value of the hot pixel is then replaced (S5) with the determined replacement value. In this way, noise generated by the absorption of x-ray photons in the image sensor may be reduced, while substantially maintaining the resolution (sharpness) in the image.