KR102458523B1 - Method and server for processing energy spectrum data of photon counting x-ray detector based on silicon photomultiplier - Google Patents

Abstract

An energy spectrum data processing method performed by a server according to an aspect of the present application includes: receiving energy spectrum data measured by an X-ray detector; generating training data by applying a pre-processing process to the received energy spectrum data; and training the energy spectrum data processing model to discriminate energy bands by inputting the training data. The preprocessing process is to check the number of peaks in the energy spectrum data, to distinguish energy bands by applying a Gaussian curve function to the energy spectrum, and to identify a threshold value that is a junction of a Gaussian curve between each energy band.

Description

METHOD AND SERVER FOR PROCESSING ENERGY SPECTRUM DATA OF PHOTON COUNTING X-RAY DETECTOR BASED ON SILICON PHOTOMULTIPLIER

The present invention relates to a method and apparatus for improving the energy discrimination accuracy and uniformity of energy spectrum data, and more particularly, to improve the energy discrimination accuracy and uniformity of energy spectrum data measured by a silicon photomultiplier-based photon counting X-ray detector it's about how to do it.

Recently, interest in X-ray detectors is increasing due to the spread of COVID-19. For this reason, the phase of the X-ray detector in the medical device industry is increasing day by day. Accordingly, in order to accurately and quickly diagnose a patient's disease, the need for an X-ray detector with high reliability is increasing.

The X-ray detector is a photon counting detector (PCD) that obtains an energy value of X-rays by individually counting X-rays and an energy integration detector that integrates energy values of X-rays detected for a certain period of time , EID).

Compound semiconductor sensors such as CZT and CdTe are mainly used for photon counting detectors due to their high energy resolution, but have disadvantages such as high operating voltage, slow response signal, low gain, and high cost. Recently, a silicon photomultiplier sensor with advantages such as low operating voltage, fast response signal, high gain, and low cost has been in the spotlight as a new photon counting X-ray detector. This has a disadvantage of adversely affecting the contrast and resolution of the image. Therefore, there is a need for a method to solve this problem.

Republic of Korea Patent Publication No. 10-2004-0024912 (Title of the invention: Intelligent radiation dosimetry system and method employing energy spectrum analysis function)

The present invention is to solve the above problems, and it is a technical task to improve the energy discrimination accuracy and uniformity degradation of a photon counting X-ray detector based on a silicon photomultiplier sensor using a deep learning method.

In addition, it learns the energy spectrum of X-rays measured with a photon counting X-ray detector based on a silicon photomultiplier sensor to find a threshold value that can distinguish low and high energy bands, and matches the low and high energy bands according to the threshold to a Gaussian curve. It aims to increase the accuracy of energy discrimination by finding the window value of each energy band.

In addition, the uniformity of energy discrimination is improved by continuously analyzing the energy spectrum that fluctuates according to the temperature and changing the low and high energy window according to the fluctuation of the energy spectrum.

However, the technical task to be achieved by the present embodiment is not limited to the above-described technical task, and other technical tasks may exist.

As a technical means for solving the above technical problem, the energy spectrum data processing method performed by the server according to the first aspect of the present disclosure includes: receiving energy spectrum data measured by an X-ray detector; generating training data by applying a pre-processing process to the received energy spectrum data; and training the energy spectrum data processing model to discriminate energy bands by inputting the training data. The preprocessing process is to check the number of peaks in the energy spectrum data, to distinguish energy bands by applying a Gaussian curve function to the energy spectrum, and to check a threshold value that is a junction of the Gaussian curve between each energy band.

In addition, the energy spectrum data processing server according to the second aspect of the present disclosure includes a memory in which the energy spectrum data processing program is stored; A processor for executing a program stored in a memory is included. The processor receives the energy spectrum data measured by the X-ray detector by executing the program, generates training data by applying a preprocessing process to the received energy spectrum data, and uses the training data as an input to distinguish energy bands from energy spectrum data. Train the processing model. The preprocessing process is to check the number of peaks in the energy spectrum data, to distinguish energy bands by applying a Gaussian curve function to the energy spectrum, and to identify a threshold value that is a junction of a Gaussian curve between each energy band.

According to any one of the above-described problem solving means of the present application, the energy spectrum of the overlapping low-energy region and the high-energy region can be clearly distinguished due to the low energy resolution of the silicon photomultiplier, so that the accuracy can be increased, and through this, the contrast of the image can improve

Silicon photomultipliers react sensitively to temperature, and when the energy spectrum changes with temperature, the number of counts in each energy region moves, resulting in poor uniformity. According to the present invention, a uniform number of counts can be obtained by tracking an energy spectrum that changes according to temperature and setting an energy window of each energy spectrum, so that the contrast and resolution of an image can be improved.

1 is a block diagram illustrating the configuration of an energy spectrum data processing system according to an embodiment of the present invention.
2 is a block diagram illustrating the configuration of an energy spectrum data processing server according to an embodiment of the present invention.
3 is a flowchart illustrating a method for processing energy spectrum data according to an embodiment of the present invention.
4 is a flowchart illustrating a pre-processing process in an energy spectrum data processing method according to an embodiment of the present invention.
5 is a diagram illustrating an energy spectrum data processing model in a method for processing energy spectrum data according to an embodiment of the present invention.
6A and 6B are diagrams for explaining an energy spectrum in a method for processing energy spectrum data according to an embodiment of the present invention.
7A and 7B are flowcharts for explaining the energy domain versus count in the energy spectrum data processing method according to an embodiment of the present invention.

Hereinafter, embodiments of the present application will be described in detail with reference to the accompanying drawings so that those of ordinary skill in the art to which the present application pertains can easily implement them. However, the present application may be implemented in several different forms and is not limited to the embodiments described herein. And in order to clearly explain the present application in the drawings, parts irrelevant to the description are omitted, and similar reference numerals are attached to similar parts throughout the specification.

Throughout this specification, when a part is said to be "connected" with another part, it includes not only the case where it is "directly connected" but also the case where it is "electrically connected" with another element interposed therebetween. do.

Throughout this specification, when a member is said to be located “on” another member, this includes not only a case in which a member is in contact with another member but also a case in which another member is present between the two members.

Hereinafter, an embodiment of the present invention will be described in detail with reference to the accompanying drawings.

1 is a block diagram illustrating the configuration of an energy spectrum data processing system according to an embodiment of the present invention.

The energy spectrum data processing method may be implemented in the server 100 . In the energy spectrum data processing method, the X-ray energy spectrum data may be collected by the X-ray detector 200 . The X-ray detector may include a silicon photomultiplier-based photon counting X-ray detector.

The photon counting detector is a technology to classify and measure photons by energy of X-ray rays, and it makes it possible to distinguish lesions more clearly by using the fact that the energy response is different for each tissue component of the human body. Photon counting detectors are being applied to detectors of various medical devices such as X-ray, CT, DEXA, PET, Gamma Camera, and SPECT. The photon counting X-ray detector includes a photoconductor portion that converts X-rays into a charge signal, and a signal processing stage that converts and processes the generated charge signal into an electrical signal using the generated charge signal as an input.

The silicon photomultiplier sensor used in the photon counting X-ray detector consists of thousands of Geiger-mode avalanche photodiode structures connected in parallel each operating in Geiger mode. has the advantage of

The energy spectrum data processing model is a model for distinguishing energy bands by learning based on the collected data, and may be learned using a deep learning neural network such as a convolutional neural network (CNN).

2 is a block diagram illustrating the configuration of an energy spectrum data processing server according to an embodiment of the present invention.

As shown, the energy spectrum data processing server 100 may include a communication module 110 , a memory 120 , a processor 130 , a database 140 , and an input module 150 .

The communication module 110 may transmit/receive data to and from the connected terminal 200 . The communication module 110 may be a device including hardware and software necessary for transmitting and receiving signals such as control signals or data signals through wired/wireless connection with other network devices.

The memory 120 stores an energy spectrum data processing program. The energy spectrum data processing program receives the energy spectrum data measured by the X-ray detector, applies a preprocessing process to the received energy spectrum data to generate training data, and uses the training data as an input to distinguish energy bands from energy spectrum data. Train the processing model.

Various types of data generated during the execution of an operating system for driving the energy spectrum data processing server 100 or an energy spectrum data processing program are stored in the memory 120 .

In this case, the memory 120 collectively refers to a non-volatile storage device that continuously maintains stored information even when power is not supplied, and a volatile storage device that requires power to maintain the stored information.

In addition, the memory 120 may perform a function of temporarily or permanently storing data processed by the processor 130 . Here, the memory 120 may include magnetic storage media or flash storage media in addition to the volatile storage device requiring power to maintain stored information, but the scope of the present invention is limited thereto. it's not going to be

The processor 130 executes the program stored in the memory 120, but controls the entire process according to the execution of the energy spectrum data processing program. Each operation performed by the processor 130 will be described in more detail later.

The processor 130 may include any type of device capable of processing data. For example, it may refer to a data processing device embedded in hardware having a physically structured circuit to perform a function expressed as a code or an instruction included in a program. As an example of the data processing device embedded in the hardware as described above, a microprocessor, a central processing unit (CPU), a processor core, a multiprocessor, an application-specific (ASIC) An integrated circuit) and a processing device such as a field programmable gate array (FPGA) may be included, but the scope of the present invention is not limited thereto.

The database 140 stores or provides data necessary for the energy spectrum data processing system under the control of the processor 130 . The database 140 may be included as a component separate from the memory 120 , or may be built in a part of the memory 120 .

3 is a flowchart illustrating a method for processing energy spectrum data according to an embodiment of the present invention.

The processor 130 receives the energy spectrum data measured by the X-ray detector (S110). In this case, the energy spectrum data may be X-ray data detected by a silicon photomultiplier-based photon counting detector.

The processor 130 generates training data by applying a preprocessing process to the received energy spectrum data (S120).

The pre-processing process is to train an energy spectrum data processing model by selecting data in which energy bands are distinguished from the measured energy spectrum data. The preprocessing process may include identifying the number of peaks in the energy spectrum data, distinguishing energy bands by applying a Gaussian curve function to the energy spectrum, and identifying a threshold value that is a junction of a Gaussian curve between energy bands. Also, the process may include setting an energy window of a low/high band based on the Gaussian curve function and a threshold value, and selecting the energy spectrum data by checking whether energy bands can be distinguished according to the energy spectrum.

The processor 130 trains the energy spectrum data processing model to distinguish energy bands by inputting the training data (S130).

The processor 130 generates an energy spectrum data processing result value through the energy spectrum data processing model (S140).

4 is a flowchart illustrating a pre-processing process in an energy spectrum data processing method according to an embodiment of the present invention.

The processor 130 checks the number of energy spectrum peaks in the energy spectrum data (S121). Before applying the Gaussian curve function for energy band discrimination, the number of peaks in the energy spectrum measured for this purpose may be identified. According to the number of peaks, characteristics of the data such as the energy band of the measured data or the amount of X-rays can be confirmed.

The processor 130 applies a Gaussian curve function to the energy spectrum data (S122). Gaussian curve, also called normal distribution, is a continuous probability distribution and its shape is determined through two parameters, mean and standard deviation, and usually the frequency distribution curve shows a bell shape symmetrical about the mean point. In the present invention, a Gaussian curve is used to distinguish energy bands, and two energy regions can be distinguished from each other.

The processor 130 checks a threshold value in the energy spectrum data (S123). The threshold may be a junction part of a Gaussian curve between each energy band.

The processor 130 selects the data according to whether the energy band can be discriminated from the energy spectrum data ( S124 ). It is checked whether energy bands can be distinguished or discriminated according to the energy spectrum, and only possible data can be selected and used as training data for the energy spectrum data processing model.

5 is a diagram illustrating an energy spectrum data processing model in a method for processing energy spectrum data according to an embodiment of the present invention.

The energy spectrum data processing model uses a convolutional neural network and can be composed of an encoder network and a decoder network. The encoder network can consist of a 3 x 3 convolutional kernel and 6 convolutional layers, and includes batch normalization, ReLU activation function, and 2x2 max for normalizing inputs of layers after the input layer. It may include pooling and bilinear interpolation.

The decoder network may be configured in a way that mirrors the encoder network. The configuration of the network is not limited as described above, and may be configured differently in consideration of learning time, performance, and the like.

6A and 6B are diagrams for explaining an energy spectrum in a method for processing energy spectrum data according to an embodiment of the present invention.

6A shows an 8-channel energy spectrum of a photon counting X-ray detector based on a silicon photomultiplier. It is possible to check the channel-specific errors (structural and sensor performance) of the silicon photomultiplier, and there is a problem that the low and high energy bands cannot be completely distinguished.

Figure 6b displays the 8-channel energy spectrum in which the low- and high-energy band threshold (X) is found using the deep learning method and the low- and high-band energy window is set according to the Gaussian curve. An energy band can be determined, and an energy window can be applied to each region.

7A and 7B are flowcharts for explaining the energy domain versus count in the energy spectrum data processing method according to an embodiment of the present invention.

Photon counting In an X-ray detector, photon counting is a type of method for measuring weak light, and the X-ray intensity is measured by calculating X-rays as photons per unit time. Photons are converted into current and pulses by a photomultiplier tube and counted (counted).

7A is a count of a low energy region of an energy spectrum obtained by irradiating X-rays for 10 minutes. X is the count over time when the deep learning structure is not applied, and Y is the count over time when the deep learning structure is applied. The silicon photomultiplier sensor has a temperature-sensitive characteristic, so the spectrum shifts as the temperature rises over time, resulting in count fluctuation (X: circular line), resulting in reduced uniformity, but when deep learning structure is applied (Y: square line) The energy window can be corrected to improve uniformity.

7B shows counts versus high energy regions of an energy spectrum obtained by irradiating X-rays for 10 minutes. X is the count over time when the deep learning structure is not applied, and Y is the count over time when the deep learning structure is applied. The silicon photomultiplier sensor has a temperature-sensitive characteristic, so the spectrum shifts as the temperature rises over time, resulting in count fluctuation (X: circular line), resulting in reduced uniformity, but when deep learning structure is applied (Y: square line) The energy window can be corrected to improve uniformity.

As shown in Figures 7a and 7b, the measured energy spectrum data is uniform in both the high energy region and the low energy region due to the temperature sensitivity of the silicon photomultiplier sensor. can improve the uniformity of the data.

An embodiment of the present invention may also be implemented in the form of a recording medium including instructions executable by a computer, such as a program module executed by a computer. Computer-readable media can be any available media that can be accessed by a computer and includes both volatile and nonvolatile media, removable and non-removable media. Also, computer-readable media may include computer storage media. Computer storage media includes both volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data.

Although the methods and systems of the present invention have been described with reference to specific embodiments, some or all of their components or operations may be implemented using a computer system having a general purpose hardware architecture.

The foregoing description of the present application is for illustration, and those of ordinary skill in the art to which the present application pertains will understand that it can be easily modified into other specific forms without changing the technical spirit or essential features of the present application. Therefore, it should be understood that the embodiments described above are illustrative in all respects and not restrictive. For example, each component described as a single type may be implemented in a distributed manner, and likewise components described as distributed may also be implemented in a combined form.

The scope of the present application is indicated by the following claims rather than the above detailed description, and all changes or modifications derived from the meaning and scope of the claims and their equivalent concepts should be construed as being included in the scope of the present application.

110: communication module
120: memory
130: processor
140: database
150: input module
200: X-ray detector

Claims (13)

In the energy spectrum data processing method performed by the server,
(a) receiving energy spectrum data measured by a photon counting X-ray detector based on a silicon photomultiplier sensor;
(b) generating training data by applying a pre-processing process to the received energy spectrum data; and
(c) training an energy spectrum data processing model to be learned based on the learning data and to distinguish energy bands of input energy spectrum data; including,
The pre-processing process identifies the number of peaks in the energy spectrum data, applies a Gaussian curve function to the energy spectrum data to distinguish a low energy band from a high energy band, and identifies a threshold value that is a junction part of a Gaussian curve between each energy band will,
The energy spectrum data processing model will be learned based on the energy spectrum data that changes with temperature, the energy spectrum data processing method. The method of claim 1,
The pre-processing process is to set the energy window of the low/high band based on the Gaussian curve function and the threshold value,
Energy spectrum data processing method. delete The method of claim 1,
The energy spectrum data processing model is a deep learning network including an encoder and a decoder,
the encoder and decoder comprise one or more convolutional layers,
wherein the decoder is configured in a mirrored manner of the encoder,
Energy spectrum data processing method. delete The method of claim 1,
After step (c), receiving the energy spectrum data measured by the photon counting X-ray detector, and when the received energy spectrum data is input, generating an energy spectrum data processing result value through the energy spectrum data processing model; further comprising,
Energy spectrum data processing method. In the energy spectrum data processing server,
a memory storing an energy spectrum data processing program;
A processor for executing the program stored in the memory;
The processor receives energy spectrum data measured by a silicon photomultiplier sensor-based photon counting X-ray detector by executing the program, and applies a preprocessing process to the received energy spectrum data to generate training data, and the training data Learned based on and train the energy spectrum data processing model to distinguish the energy bands of the input energy spectrum data,
The pre-processing process identifies the number of peaks in the energy spectrum data, applies a Gaussian curve function to the energy spectrum data to distinguish a low energy band from a high energy band, and identifies a threshold value that is a junction part of a Gaussian curve between each energy band will,
The energy spectrum data processing model will be learned based on the energy spectrum data that changes with temperature, the energy spectrum data processing server. 8. The method of claim 7,
The pre-processing process is to set the energy window of the low/high band based on the Gaussian curve function and the threshold value,
Energy Spectrum Data Processing Server. delete 8. The method of claim 7,
The energy spectrum data processing model is a deep learning network including an encoder and a decoder,
the encoder or decoder comprises one or more convolutional layers,
wherein the decoder is configured in a mirrored manner of the encoder,
Energy Spectrum Data Processing Server. delete 8. The method of claim 7,
The processor receives the energy spectrum data measured by the photon counting X-ray detector, and when the received energy spectrum data is input, generates an energy spectrum data processing result value through the energy spectrum data processing model,
Energy Spectrum Data Processing Server. A non-transitory computer-readable recording medium in which a computer program for performing the energy spectrum data processing method according to any one of claims 1, 2, 4 and 6 is recorded.

Images