Photon counting CT geometric correction method

Abstract

The invention relates to the technical field of photon counting CT scanning, and discloses a photon counting CT geometric correction method, which is applied to correction equipment, and specifically comprises the following steps: S101, obtaining original CT projection data through CT scanning, and obtaining a corresponding relationship between photon energy corresponding to each current pixel in the original CT projection data and a threshold value code; s102, inputting the original CT projection data into a preset content text recognition model to obtain a geometric offset vector of the projection data; and S103, obtaining a text recognition result output by the text recognition model based on the geometric offset vector. According to the method, the geometric offset is acquired and transmitted to the CT system to complete correction of the projection light beam, the micro-calibration effect is achieved, the corrected numerical value of the CT system is in the micron level, and therefore when the CT system uses projection in the later period, the projected light beam is accurate, and the accuracy of distinguishing a projection picture in the later period of the CT system is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of photon counting CT scanning, and in particular to a photon counting CT geometric correction method. Background Art

[0002] Mechanical vibrations during CT scanning or patient breathing, heartbeat and other movements will cause motion artifacts in the reconstructed image, seriously affecting diagnosis and treatment. For rigid motion, external motion detection equipment is usually used for geometric correction or geometric offset calculation based on projection image information; for non-rigid motion such as breathing and heartbeat, the mainstream solution is to use a respiratory detector and electrocardiogram to detect breathing and heartbeat, so that CT can scan at the same time when breathing is still and the heartbeat is. The disadvantage is that it requires the use of an external model, which is relatively expensive.

[0003] Currently, there are some technologies that do not require external detectors, but use algorithms to perform geometric correction, such as an automatic system calibration method for X-ray CT with application number 201680026497.1. However, it is time-consuming to calculate and requires the latest hardware to complete the huge amount of calculations, and the correction effect is average.

[0004] Therefore, it is urgent to design a photon counting CT geometric correction method, which relies on geometric correction to automatically obtain the projection offset at various angles in real time and eliminate the geometric motion artifacts of the reconstructed image. Summary of the invention

[0005] The purpose of the present invention is to provide a photon counting CT geometric correction method, which achieves the effect of micro-calibration, so that the value corrected by the CT system is at the micron level, so that the light beam projected by the CT system when used for projection in the later stage is accurate, and the accuracy of the CT system in identifying the projection image in the later stage is improved. It is intended to solve the problems in the prior art that the calculation is time-consuming, the latest hardware is required to complete the huge amount of calculation, and the correction effect is general.

[0006] The present invention is implemented as follows: a photon counting CT geometric correction method is applied to a correction device and specifically comprises the following steps: S101: obtaining original CT projection data through CT scanning, and obtaining a corresponding relationship between the photon energy corresponding to each pixel in the original CT projection data and the threshold code; S102: Inputting the original CT projection data into a preset content text recognition model to obtain a geometric offset vector of the projection data; S103: Obtaining a text recognition result output by the text recognition model based on the geometric offset vector, and obtaining an offset curve of the original CT projection data in the text recognition result; S104: Obtaining a shift curve of the original CT projection data, and then obtaining a preset correspondence relationship between the photon energy corresponding to each pixel and the threshold code, determining whether the correspondence relationship between the photon energy corresponding to each pixel and the threshold code is consistent, and drawing a shift curve of the ideal CT projection data; S105: Obtain the geometric offset of the current CT projection by comparing the offset curve of the original CT projection data with the offset curve of the ideal CT projection data, and correct and reconstruct the original CT projection data according to the geometric offset vector.

[0007] Furthermore, in S101, original CT projection data is obtained through CT scanning, including: A connection request to the CT system via a preset network, wherein the connection request is used to request to establish a connection with the correction device; Detecting whether the current account of the CT system is the target account; If the current account of the CT system is the target account, a connection is made with the calibration device according to the connection request, and after the connection is completed, a projection data acquisition request is sent to the CT system.

[0008] Furthermore, the corresponding relationship between the photon energy corresponding to each pixel in the original CT projection data and the threshold code is obtained, including: Identify each pixel in the original CT projection data; Obtaining a threshold code corresponding to a k-edge value feature for each pixel, and fitting the k-edge value feature and the corresponding threshold code; The k-edge value feature and the corresponding threshold coding fitting curve are obtained, and the corresponding relationship between the photon energy corresponding to each pixel and the threshold coding is obtained in the fitting curve.

[0009] Furthermore, in S102, the original CT projection data is input into a preset content text recognition model to obtain a geometric offset vector of the projection data, including: Acquire multiple text recognition models, and sort the multiple text recognition models in order of projection slope from small to large; The original CT projection data content is obtained, and the original CT projection data is input into the first-ranked text recognition model.

[0010] Furthermore, the plurality of text recognition models are sorted in order of projection slope from small to large, including: Acquire a text recognition model pre-stored in the CT system, and identify a preset ideal CT projection data curve in the text recognition model; Each ideal CT projection data curve is extracted, and the projection slope of the current projection data curve is calculated according to the projection data curve to complete the sorting of multiple text recognition models.

[0011] Further, inputting the original CT projection data into the first-ranked text recognition model includes: Acquire the original CT projection data, input the original CT projection data into the first-ranked text recognition model, and acquire a first offset feature; Acquire the original CT projection data, input the original CT projection data into the second ranked text recognition model, and acquire a second offset feature; Compare the deviations of the first offset feature and the second offset feature, retain the offset feature with the smaller deviation, until all text recognition models are recognized, obtain the offset feature with the smallest deviation, and then calculate and convert the offset feature into a geometric offset vector.

[0012] Furthermore, the preset network includes one or a combination of a 3G network, a 4G network, a 5G network, and a WIFI network.

[0013] Furthermore, in S103, obtaining a shift curve of the original CT projection data includes: Obtain the corresponding relationship between the photon energy and the threshold code corresponding to each pixel, identify and draw a discrete curve of the photon count value with respect to the threshold code; The geometric offset vector is inserted into the obtained discrete curve of the photon count value with respect to the threshold code to perform average calculation to obtain reference corresponding relationship points, and each corresponding relationship point is linked to obtain the offset curve of the original CT projection data.

[0014] Furthermore, in S104, it is determined whether the corresponding relationship between the photon energy corresponding to each pixel and the threshold code is consistent, including: Obtaining a deviation between the discrete curve of each pixel with respect to the threshold coding and a preset correspondence between the photon energy corresponding to each pixel and the threshold coding; The total deviation amount of the plurality of pixels is determined according to the deviation amount of each pixel.

[0015] Furthermore, in S105, the geometric offset of the current CT projection is obtained by comparing the offset curve of the original CT projection data with the offset curve of the ideal CT projection data, including: Obtaining an identification model of a deviation curve of ideal CT projection data; Obtaining a deviation curve of original CT projection data, and putting the deviation curve of the original CT projection data into a recognition model of a deviation curve of ideal CT projection data; According to the offset curve of the original CT projection data, the line segments are discretized in the recognition model of the offset curve of the ideal CT projection data to obtain the geometric offset of the current CT projection.

[0016] Compared with the prior art, the photon counting CT geometric correction method provided by the present invention has the following beneficial effects: 1. By identifying and reading the original CT projection data obtained in the early stage, the offset curve of the original CT projection data is obtained, and the corresponding relationship between the photon energy and the threshold code corresponding to each pixel of the original CT projection data is obtained. The corresponding relationship between the photon energy and the threshold code is formed by using the fitting curve formed by the threshold code corresponding to the k-edge value feature to obtain the corresponding relationship between the photon energy and the threshold code, so that the offset curve of the original CT projection data finally obtained is compared with the offset curve of the ideal CT projection data, and the geometric offset is obtained and transmitted to the CT system to complete the correction of the projection beam, achieving the effect of micro-calibration, so that the value corrected by the CT system is at the micron level, so that the CT system can accurately project the beam when using the projection in the later stage, and improve the accuracy of the CT system in identifying the projection screen in the later stage; 2. The offset curve of the original CT projection data is obtained by obtaining the corresponding relationship between the photon energy corresponding to each pixel and the threshold code, identifying and drawing a discrete curve of the photon count value with respect to the threshold code to ensure the accuracy of the output corresponding relationship points, and then using the corresponding relationship between the photon energy corresponding to each pixel and the threshold code to determine whether the corresponding relationship between the photon energy corresponding to each pixel and the threshold code is consistent, so as to obtain the total deviation of multiple pixels to draw the offset curve of the ideal CT projection data, so that the projection beam pixel point of the offset curve of the drawn ideal CT projection data is optimal, so as to serve as the correction base point of the offset curve of the original CT projection data. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A schematic flow chart of a photon counting CT geometric correction method proposed by the present invention; Figure 2 A flowchart of a photon counting CT geometric correction method proposed by the present invention for obtaining original CT projection data through CT scanning; Figure 3 A flowchart of obtaining the corresponding relationship between the photon energy corresponding to each pixel in the original CT projection data and the threshold code in a photon counting CT geometric correction method proposed by the present invention; Figure 4 A flowchart of a method for geometric correction of photon counting CT proposed by the present invention, in which the original CT projection data is input into a preset content text recognition model to obtain a geometric offset vector of the projection data; Figure 5 A flowchart of a photon counting CT geometric correction method proposed in the present invention for obtaining the geometric offset of the current CT projection by comparing the offset curve of the original CT projection data with the offset curve of the ideal CT projection data. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0019] The implementation of the present invention is described in detail below in conjunction with specific embodiments.

[0020] The same or similar numbers in the drawings of this embodiment correspond to the same or similar parts; in the description of the present invention, it should be understood that if the terms "upper", "lower", "left", "right" and the like indicate directions or positional relationships based on the directions or positional relationships shown in the drawings, it is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction. Therefore, the terms describing the positional relationship in the drawings are only used for illustrative purposes and cannot be understood as limitations on the present invention. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to specific circumstances.

[0021] Reference Figure 1-5 As shown, a photon counting CT geometric correction method is applied to a correction device and specifically includes the following steps: S101: obtaining original CT projection data through CT scanning, and obtaining a corresponding relationship between the photon energy corresponding to each pixel in the original CT projection data and the threshold code; The original CT projection data is obtained through CT scanning, including: A connection request to the CT system via a preset network, the connection request being used to request to establish a connection with the calibration device; Check whether the current account of the CT system is the target account; If the current account of the CT system is the target account, it connects with the calibration device according to the connection request, and after the connection is completed, it sends a projection data acquisition request to the CT system; The corresponding relationship between the photon energy corresponding to each pixel in the original CT projection data and the threshold code is obtained, including: Identify each pixel in the original CT projection data; Obtain the threshold code corresponding to the k-edge value feature for each pixel, and fit the k-edge value feature and the corresponding threshold code; Obtain the k-edge value feature and the corresponding threshold coding fitting curve, and obtain the corresponding relationship between the photon energy corresponding to each pixel and the threshold coding in the fitting curve; S102: inputting the original CT projection data into a preset content text recognition model to obtain a geometric offset vector of the projection data; S103: Obtaining a text recognition result output by a text recognition model based on a geometric offset vector, and obtaining an offset curve of original CT projection data in the text recognition result; The offset curve of the original CT projection data is obtained, including: Obtain the corresponding relationship between the photon energy and the threshold code corresponding to each pixel, identify and draw a discrete curve of the photon count value with respect to the threshold code; Inserting the geometric offset vector into the obtained discrete curve of the photon count value with respect to the threshold encoding to perform average calculation to obtain the reference corresponding relationship point, linking each corresponding relationship point, that is, obtaining the offset curve of the original CT projection data; S104: Obtaining a shift curve of the original CT projection data, and then obtaining a preset correspondence relationship between the photon energy corresponding to each pixel and the threshold code, determining whether the correspondence relationship between the photon energy corresponding to each pixel and the threshold code is consistent, and drawing a shift curve of the ideal CT projection data; The process of determining whether the photon energy corresponding to each pixel is consistent with the corresponding relationship of the threshold code includes: Obtaining the deviation between the discrete curve of each pixel with respect to the threshold coding and the preset corresponding relationship between the photon energy and the threshold coding corresponding to each pixel; Determine the total deviation of the plurality of pixels according to the deviation of each pixel; S105: By comparing the offset curve of the original CT projection data with the offset curve of the ideal CT projection data, the geometric offset of the current CT projection is obtained, and the CT original projection data is corrected and reconstructed according to the geometric offset vector. The technical solution obtains the offset curve of the original CT projection data by identifying and reading the original CT projection data obtained in the early stage, and the corresponding relationship between the photon energy and the threshold code corresponding to each pixel of the original CT projection data is obtained. The corresponding relationship between the photon energy and the threshold code is obtained by using a fitting curve formed by the threshold code corresponding to the k-edge value feature, so that the offset curve of the original CT projection data finally obtained is compared with the offset curve of the ideal CT projection data, and the geometric offset is obtained and transmitted to the CT system to complete the correction of the projection beam, thereby achieving the effect of micro-calibration, so that the value corrected by the CT system is at the micron level, so that the CT system can accurately project the beam when using the projection in the later stage, thereby improving the accuracy of the CT system in identifying the projection screen in the later stage.

[0022] In S102 of this embodiment, the original CT projection data is input into a preset content text recognition model to obtain a geometric offset vector of the projection data, including: Obtain multiple text recognition models, and sort the multiple text recognition models in order of projection slope from small to large; Among them, multiple text recognition models are sorted in order from small to large projection slope, including: Acquire a text recognition model pre-stored in the CT system, and identify a preset ideal CT projection data curve in the text recognition model; Extract each ideal CT projection data curve, and calculate the projection slope of the current projection data curve according to the projection data curve to complete the sorting of multiple text recognition models; Obtaining the original CT projection data content, and inputting the original CT projection data into the first-ranked text recognition model; The original CT projection data is input into the first-ranked text recognition model, including: Obtaining original CT projection data, inputting the original CT projection data into the first-ranked text recognition model, and obtaining a first offset feature; Obtaining original CT projection data, and inputting the original CT projection data into the second-ranked text recognition model to obtain a second offset feature; Compare the deviations of the first offset feature and the second offset feature, retain the offset feature with the smaller deviation, until all text recognition models are recognized, obtain the offset feature with the smallest deviation, and then calculate and convert the offset feature into a geometric offset vector.

[0023] In this embodiment, the preset network includes one or a combination of a 3G network, a 4G network, a 5G network, and a WIFI network.

[0024] In this embodiment, the offset curve of the original CT projection data is obtained by obtaining the corresponding relationship between the photon energy corresponding to each pixel and the threshold code, identifying and drawing a discrete curve of the photon count value with respect to the threshold code, so as to ensure the accuracy of the output corresponding relationship points; In S105 of this embodiment, the geometric offset of the current CT projection is obtained by comparing the offset curve of the original CT projection data with the offset curve of the ideal CT projection data, including: Obtaining an identification model of a deviation curve of ideal CT projection data; Obtaining a deviation curve of original CT projection data, and putting the deviation curve of the original CT projection data into a recognition model of a deviation curve of ideal CT projection data; According to the discretized line segments in the recognition model of the offset curve of the ideal CT projection data, the geometric offset of the current CT projection is obtained. The corresponding relationship between the photon energy corresponding to each pixel and the threshold code is used to determine whether the corresponding relationship between the photon energy corresponding to each pixel and the threshold code is consistent, so as to obtain the total deviation of multiple pixels to draw the offset curve of the ideal CT projection data, so that the projection beam pixel point of the offset curve of the drawn ideal CT projection data is optimal, so as to serve as the correction base point of the offset curve of the original CT projection data.

[0025] The technical solution obtains the offset curve of the original CT projection data by identifying and reading the original CT projection data obtained in the early stage, and the correspondence between the photon energy corresponding to each pixel of the original CT projection data and the threshold code, so as to compare the offset curve of the original CT projection data finally obtained with the offset curve of the ideal CT projection data, obtain the geometric offset, and transmit it to the CT system to complete the correction of the projection beam, thereby achieving the effect of micro-calibration, so that the value corrected by the CT system is at the micron level, so that the CT system can accurately project the beam when using the projection in the later stage, thereby improving the accuracy of the CT system in identifying the projection picture in the later stage.

[0026] In this embodiment, the entire operation process can be controlled by a computer to perform signal feedback and implement the steps in sequence. These are all common knowledge of current automated control and will not be described in detail in this embodiment.

[0027] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A photon counting CT geometric correction method, characterized in that: Applied to calibration equipment, specifically including the following steps: S101: obtaining original CT projection data through CT scanning, and obtaining a corresponding relationship between the photon energy corresponding to each pixel in the original CT projection data and the threshold code; S102: Inputting the original CT projection data into a preset content text recognition model to obtain a geometric offset vector of the projection data; S103: Obtaining a text recognition result output by the text recognition model based on the geometric offset vector, and obtaining an offset curve of the original CT projection data in the text recognition result; S104: Obtaining a shift curve of the original CT projection data, and then obtaining a preset correspondence relationship between the photon energy corresponding to each pixel and the threshold code, determining whether the correspondence relationship between the photon energy corresponding to each pixel and the threshold code is consistent, and drawing a shift curve of the ideal CT projection data; S105: Obtain the geometric offset of the current CT projection by comparing the offset curve of the original CT projection data with the offset curve of the ideal CT projection data, and correct and reconstruct the original CT projection data according to the geometric offset vector.

2. A photon counting CT geometric correction method as claimed in claim 1, characterized in that: In S101, original CT projection data is obtained through CT scanning, including: A connection request to the CT system via a preset network, wherein the connection request is used to request to establish a connection with the correction device; Detecting whether the current account of the CT system is the target account; If the current account of the CT system is the target account, a connection is made with the calibration device according to the connection request, and after the connection is completed, a projection data acquisition request is sent to the CT system.

3. A photon counting CT geometric correction method as claimed in claim 2, characterized in that: And obtain the corresponding relationship between the photon energy corresponding to each pixel in the original CT projection data and the threshold code, including: Identify each pixel in the original CT projection data; Obtaining a threshold code corresponding to a k-edge value feature for each pixel, and fitting the k-edge value feature and the corresponding threshold code; The k-edge value feature and the corresponding threshold coding fitting curve are obtained, and the corresponding relationship between the photon energy corresponding to each pixel and the threshold coding is obtained in the fitting curve.

4. A photon counting CT geometric correction method as claimed in claim 3, characterized in that: In S102, the original CT projection data is input into a preset content text recognition model to obtain a geometric offset vector of the projection data, including: Acquire multiple text recognition models, and sort the multiple text recognition models in order of projection slope from small to large; The original CT projection data content is obtained, and the original CT projection data is input into the first-ranked text recognition model.

5. A photon counting CT geometric correction method as claimed in claim 4, characterized in that: Sorting the multiple text recognition models in order of projection slope from small to large includes: Acquire a text recognition model pre-stored in the CT system, and identify a preset ideal CT projection data curve in the text recognition model; Each ideal CT projection data curve is extracted, and the projection slope of the current projection data curve is calculated according to the projection data curve to complete the sorting of multiple text recognition models.

6. A photon counting CT geometric correction method as claimed in claim 5, characterized in that: Inputting the original CT projection data into the first-ranked text recognition model comprises: Acquire the original CT projection data, input the original CT projection data into the first-ranked text recognition model, and acquire a first offset feature; Acquire the original CT projection data, input the original CT projection data into the second ranked text recognition model, and acquire a second offset feature; Compare the deviations of the first offset feature and the second offset feature, retain the offset feature with the smaller deviation, until all text recognition models are recognized, obtain the offset feature with the smallest deviation, and then calculate and convert the offset feature into a geometric offset vector.

7. A photon counting CT geometric correction method as claimed in claim 2, characterized in that: The preset network includes one or a combination of a 3G network, a 4G network, a 5G network, and a WIFI network.

8. A photon counting CT geometric correction method as claimed in claim 7, characterized in that: In S103, obtaining the offset curve of the original CT projection data includes: Obtain the corresponding relationship between the photon energy and the threshold code corresponding to each pixel, identify and draw a discrete curve of the photon count value with respect to the threshold code; The geometric offset vector is inserted into the obtained discrete curve of the photon count value with respect to the threshold code to perform average calculation to obtain reference corresponding relationship points, and each corresponding relationship point is linked to obtain the offset curve of the original CT projection data.

9. A photon counting CT geometric correction method as claimed in claim 8, characterized in that: In S104, judging whether the corresponding relationship between the photon energy corresponding to each pixel and the threshold code is consistent includes: Obtaining a deviation between the discrete curve of each pixel with respect to the threshold coding and a preset correspondence between the photon energy corresponding to each pixel and the threshold coding; The total deviation amount of the plurality of pixels is determined according to the deviation amount of each pixel.

10. A photon counting CT geometric correction method as claimed in claim 9, characterized in that: In S105, the geometric offset of the current CT projection is obtained by comparing the offset curve of the original CT projection data with the offset curve of the ideal CT projection data, including: Obtaining an identification model of a deviation curve of ideal CT projection data; Obtaining a deviation curve of original CT projection data, and putting the deviation curve of the original CT projection data into a recognition model of a deviation curve of ideal CT projection data; According to the offset curve of the original CT projection data, a discretized line segment is formed in the recognition model of the offset curve of the ideal CT projection data to obtain the geometric offset of the current CT projection.

Images