Digital X-ray Sensor CMOS Pixel Calibration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current digital X-ray sensors face challenges such as non-homogeneous response due to CMOS technology, high noise levels, limited pixel count, and inefficient photon conversion, which hinder their use in radiological applications requiring high resolution and low radiation doses.
Innovation Solution
A digital X-ray sensor with a CMOS ASIC collection layer and a calibration mechanism that uses digital-to-analog converters and logical means for iterative offset correction, enabling simultaneous calibration of all pixels and reducing noise, along with a honeycomb pixel arrangement for increased resolution and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Use of energy by stationary object
If CMOS technology is used for the collection layer, then power consumption is reduced and cost is lowered, but the device produces non-homogeneous response due to offset variations in each pixel
Solution Approach 1:
The patent applies preliminary action by performing self-calibration of offset values for each pixel before actual photon counting measurements. The calibration process determines and stores offset values that are subsequently subtracted from measurement signals, ensuring homogeneous response across all pixels while maintaining CMOS power efficiency
Solution Approach 2:
The patent implements self-service through automatic self-calibration functionality integrated into each pixel's circuitry. Each pixel autonomously determines its own offset value by analyzing dark current characteristics, eliminating the need for external calibration equipment and enabling mass production with consistent performance
2Quantity of substance
If integration panel with TFT pixel structure is used, then charge accumulation is achieved, but additional noise from dark current and electronics prevents detection of low charge levels
Solution Approach 1:
The patent extracts and removes the harmful offset component (dark current and electronic noise) from the measurement signal. By separately determining offset values through self-calibration and subtracting them from the total signal, only the genuine photon-induced charge remains, enabling detection of low charge levels
Solution Approach 2:
The patent uses feedback by continuously monitoring pixel response characteristics and using this information to adjust offset compensation. The self-calibration process measures actual pixel behavior and feeds this information back into the signal processing to correct for noise and drift
3Loss of information
If energy window discrimination is implemented, then spectroscopic features are obtained, but pixel response becomes non-homogeneous due to offset variations near thresholds
Solution Approach 1:
The patent applies preliminary action by pre-determining offset values for each pixel before energy window discrimination is applied. This calibration step ensures that subsequent threshold-based energy filtering operates on a homogeneous baseline, preventing offset variations from causing non-uniform counting responses across pixels
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 achieves homogeneous pixel responses, reduced noise, and enhanced sensitivity, allowing for high-resolution imaging with lower radiation doses and increased frame rates, suitable for radiological applications like bone densitometry and mammography.
Implementation Method 1
The conversion layer serves for transforming into an electric charge the photons of an X-ray beam that has travelled across an irradiated sample. This may occur directly or indirectly, as in the case of amorphous Selenium and of Cesium iodide
Data Source
Figure 1~2
Figure 3~3B
Figure 4
AI summary
An X-ray digital sensor (100), wherein a semiconductor conversion layer (10) configured for receiving and convert photons (2) into an electric charge (14) is integrated with a semiconductor collection layer (20), which is preferably a CMOS ASIC, and is formed by pixels (22) arranged preferably in a hexagonal honeycomb pattern, the pixels configured for receiving electrons (16) from the conversion layer (10), and wherein each collection pixel (22) comprises N discriminators (24i) each configured for carrying out a comparison between the inlet charge (14) and an own threshold (25i), and for carrying out an instantaneous transition between 0 and 1 if the inlet charge (14) exceeds the threshold, while a counter (26i) associated to each discriminator (24i), leaves respectively unchanged or increases by 1 its own count if the collected charge is higher than the threshold in the respective discriminator (24i) and lower than the threshold (25i+1) of the immediately upper discriminator (24i+1), while the other counters are configured for keeping unchanged their own counts, so that the counters measure the photons (2) in N corresponding bands of energy defined by the neighbouring thresholds. A calibration means is provided that comprises a means for being simultaneously actuated for all the pixel (22) and comprises, in each pixel (22), a digital-to-analog converter (DAC, 28i) for each discriminator (24i), configured for receiving combinations of a predetermined number of bits and for generating electric currents corresponding to the combinations of bits; a supply means (206) for supplying a current to the amplification means (203) of each pixel (22); a logical means (34) resident in each pixel and configured for carrying out an iterative procedure (80) in order to identify an offset correction current, wherein combinations of bits are generated and transferred to the DAC (28i) that generates a corresponding trial current supplied to the amplification means (203) through the supply means (206), the logical means reads a counting of the counter (26i) and causes the procedure to be repeated, if the counting increases by the trial current, or store the latter as the correction current in a memory unit (35) of the pixel (22), if the counting does not increase by the trial current.