Brain midline bias information analysis system and method
The brain centerline bias information analysis system uses learning algorithms to generate heatmaps and centerline information, automating the assessment of midline shifts in brain imaging, enhancing diagnostic accuracy and efficiency.
Patent Information
- Application Number
- KR1020240008346
- Authority / Receiving Office
- KR · KR
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2024-01-18
- Publication Date
- 2026-11-06
- Estimated Expiration
- 2044-01-18
AI Technical Summary
Existing CT scans require manual verification by medical professionals to determine midline shifts in brain imaging, which is inefficient and prone to errors.
A brain centerline bias information analysis system and method that uses learning algorithms to analyze brain imaging images, generating heatmaps and centerline information to automatically assess midline shifts.
Automatically determines midline shifts with high accuracy, reducing the need for manual verification and improving diagnostic efficiency.
Smart Images

Figure 112024007153107-PAT00001_ABST
Abstract
Description
Technology Field
[0001] The present invention relates to a system and method for analyzing brain centerline bias information, and more specifically, to a system and method for analyzing brain centerline bias information capable of determining how much the captured brain centerline is shifted by learning and analyzing brain imaging images. Background Technology
[0002] Midline displacement refers to the movement of the brain across the midline, a sign that can be clearly identified in neuroimaging techniques such as CT scans. This sign is used as a key factor in diagnosing brain diseases because it is generally associated with distortions in the brainstem that can cause severe dysfunction, such as abnormal posture and a failure of the pupil to constrict in response to light. Such midline displacement may also be associated with high intracranial pressure (ICP), and indeed, ICP can be monitored using midline displacement. A midline displacement of 5 mm or more may be considered a critical condition requiring immediate surgery, and this sign can result from conditions such as traumatic brain injury, stroke, hematoma, or congenital malformations leading to elevated intracranial pressure.
[0003] Such shifts in the midline can be detected using various methods. A representative method for detection is CT (computed tomography) scanning. Since CT scans allow for easy identification of midline shifts, equivalent results can be obtained without the use of MRI. However, when acquiring brain images via CT, a drawback exists in that verification by medical professionals is required to accurately determine whether and to what extent the midline has shifted. Prior art literature
[0004] Korean Patent Publication No. 10-2019-0087455 The problem to be solved
[0005] To solve the problems of the conventional technology described above, one embodiment of the present invention aims to provide a brain centerline bias information analysis system and method capable of obtaining centerline bias information by performing learning using brain imaging images as learning data. means of solving the problem
[0006] According to one aspect of the present invention for solving the above problem, a brain centerline bias information analysis system is provided. The brain centerline bias information analysis system comprises: a basic data acquisition unit that acquires basic data from a brain cross-section image; a data learning unit that performs data learning on the brain cross-section image through the basic data and a preset algorithm, and acquires a heatmap for the basic data as a learning result; and a data analysis unit that acquires centerline information for the brain cross-section using the heatmap and acquires brain cross-section analysis information including bias information of the centerline; wherein the preset algorithm is an algorithm that learns the brain cross-section image and text information including a description of the brain cross-section image together.
[0007] The above basic data may include at least one of AF (Anterior Falx) data, PF (Posterior Falx) data, and SP (Septum Pellucidum) data.
[0008] The above basic data may include coordinate information based on a preset point of the brain cross-section image.
[0009] The above-mentioned algorithm can be configured to learn a compressed image using an encoder and to upsample the reduced resolution using a decoder.
[0010] The heatmap above is obtained through a Gaussian heatmap, and the heatmaps of the AF data, PF data, and SP data can each be obtained through the Gaussian heatmap.
[0011] The algorithm configured above may be CPAM (Cross Position Attention Module).
[0012] The above centerline information is obtained through the AF data heatmap and the PF data heatmap, and the above orientation information can be obtained using the above centerline information and the SP data heatmap.
[0013] According to one aspect of the present invention, a method for analyzing brain centerline bias information is provided. The method for analyzing brain centerline bias information comprises: a basic data acquisition step of acquiring basic data from a brain cross-section image using a basic data acquisition unit; a data learning step of performing data learning on the brain cross-section image using the basic data and a preset algorithm using a data learning unit, and acquiring a heatmap for the basic data as a learning result; and a data analysis step of acquiring centerline information for the brain cross-section using the heatmap through a data analysis unit, and acquiring brain cross-section analysis information including bias information of the centerline; wherein the preset algorithm is an algorithm that learns together the brain cross-section image and text information including a description of the brain cross-section image. Effects of the invention
[0014] A brain centerline bias information analysis system and method according to one embodiment of the present invention has the effect of obtaining centerline bias information by performing learning using brain imaging images as training data. Brief explanation of the drawing
[0015] FIG. 1 is a block diagram of a brain centerline bias information analysis system according to an embodiment of the present invention. FIG. 2 is a flowchart of a method for analyzing brain centerline bias information according to an embodiment of the present invention. FIG. 3 is an exemplary diagram showing the operation of a preset algorithm according to one embodiment of the present invention. FIG. 4 is an example diagram of a brain imaging image showing an AF data heatmap, a PF data heatmap, and an SP data heatmap according to an embodiment of the present invention. FIG. 5 is an example of a centerline according to one embodiment of the present invention. Specific details for implementing the invention
[0016] Hereinafter, some embodiments of the present disclosure will be described in detail with reference to the exemplary drawings. In assigning reference numerals to the components of each drawing, the same components may have the same reference numeral as much as possible, even if they are shown in different drawings. Furthermore, in describing the embodiments, if it is determined that a detailed description of related known components or functions may obscure the essence of the technical concept, such detailed description may be omitted. Where terms such as "comprising," "having," or "consisting of" are used in this specification, other parts may be added unless "only" is used. Where a component is expressed in the singular, it may include a plural unless otherwise specified.
[0017] Additionally, terms such as first, second, A, B, (a), (b), etc., may be used to describe the components of the present disclosure. These terms are used merely to distinguish the components from other components, and the nature, order, sequence, or number of the components are not limited by such terms.
[0018] In describing the positional relationship of components, where it is stated that two or more components are "connected," "combined," or "joined," it should be understood that while the two or more components may be directly "connected," "combined," or "joined," they may also be "connected," "combined," or "joined" with other components "intervened." Here, the other components may be included in one or more of the two or more components that are "connected," "combined," or "joined" with one another.
[0019] In describing the temporal flow relationship regarding components, methods of operation, or methods of production, for example, when the temporal or sequential relationship is described using "after," "following," "next," or "before," it may include cases where the relationship is not continuous unless "immediately" or "directly" is used.
[0020] Meanwhile, where numerical values or corresponding information regarding a component (e.g., levels, etc.) are mentioned, even without separate explicit notation, the numerical values or corresponding information may be interpreted as including a range of error that may occur due to various factors (e.g., process factors, internal or external shocks, noise, etc.).
[0021] FIG. 1 illustrates an embodiment of the brain centerline bias information analysis system of the present invention. The brain centerline bias information analysis system (1) according to an embodiment of the present invention is configured to acquire a brain imaging image, learn from it to acquire a heatmap for basic data, and then acquire brain cross-sectional analysis information including centerline information and bias information using the heatmap.
[0022] To this end, a brain centerline bias information analysis system (1) according to one embodiment of the present invention may be formed to include a basic data acquisition unit (11), a data learning unit (13), and a data analysis unit (15) as shown in FIG. 1.
[0023] The basic data acquisition unit (11) is configured to acquire basic data from a brain cross-section image. Here, the basic data may be a brain cross-section image acquired from training data providers included in a pre-set training data pool. In addition, the present invention may be configured to acquire only images in which a description of the image is included as text as basic data. That is, even if the brain cross-section image is acquired from training data providers included in a pre-set training data pool, information that does not include a description of the image is not selected as basic data.
[0024] In addition, in one embodiment of the present invention, the basic data acquisition unit (11) may acquire a brain cross-section image including at least one of AF (Anterior Falx) data, PF (Posterior Falx) data, and SP (Septum Pellucidum) data as basic data. Furthermore, the basic data may be formed to include coordinate information based on a preset point of the brain cross-section image, and the coordinate information may be coordinate information for the point where the above-described AF data, PF data, and SP data are located.
[0025] The data learning unit (13) of the present invention is configured to perform data learning on brain cross-sectional images through basic data and a preset algorithm, and to obtain a heatmap of the basic data as a result of the learning. A preset algorithm according to one embodiment of the present invention is illustrated in FIG. 3. Referring to FIG. 3, the data learning unit (13) of the present invention is configured to perform encoding on the acquired basic data, learn the encoding result, and then use a decoder to upsample the reduced resolution again in order to output the learning result.
[0026] A pre-configured algorithm according to one embodiment of the present invention may be a Cross Position Attention Module (CPAM). More specifically, in one embodiment of the present invention, the basic data is formed to include text information as described above when acquired through the data acquisition unit (11). The data learning unit (13) of the present invention may be formed to learn the text information regarding the corresponding brain cross-section image together during the process of learning the encoding result after performing encoding to utilize it. Subsequently, when performing decoding, it may be formed to output an image containing the learning result by including the learning information.
[0027] In one embodiment of the present invention, the data learning unit (13) is formed to obtain a heatmap for basic data as a learning result, and the heatmap can be obtained as an AF data heatmap, a PF data heatmap, and an FP data heatmap for AF data, PF data, and SP data, respectively. In addition, in one embodiment of the present invention, a Gaussian heatmap may be used to obtain this.
[0028] A heatmap according to one embodiment of the present invention can be represented as shown in FIG. 4. The heatmap is used, for example, to predict the location of a corresponding landmark. That is, since the basic data of the present invention includes location information of AF, PF, and SP, it is possible to check how the locations of the corresponding points show a distribution within the brain imaging image using this information, and by using this heatmap, it can be formed so that the location can be predicted even if an image that does not include the location information is acquired later.
[0029] Meanwhile, when using a heatmap, the positional information of the SP may vary depending on the degree of bias of the brain centerline. Therefore, in one embodiment of the present invention, learning is performed using additional text information to correct the heatmap by correcting this information, and through the results of the learning, the SP data heatmap may further include information regarding centerline bias.
[0030] The data analysis unit (15) is configured to obtain centerline information for a brain cross-section using a heatmap and to obtain brain cross-section analysis information including centerline bias information. The data analysis unit (15) according to one embodiment of the present invention is configured to obtain brain cross-section imaging data for which prediction is desired, obtain centerline information by comparing it with a learning result, and then determine the degree of centerline bias.
[0031] FIG. 5 shows an example of selecting a centerline in the present invention. Referring to FIG. 5, a data analysis unit (15) according to one embodiment of the present invention is formed to generate a centerline and obtain centerline information. As shown in FIG. 5, the centerline may be a line connecting AF and PF to each other. In an embodiment of the present invention, the data analysis unit (15) may generate a centerline using a heatmap, and in this case, the AF data heatmap and the PF data heatmap may be used.
[0032] The data analysis unit (15) may be configured to determine the degree of bias by using the distance between the center line and the SP data heatmap when a center line is generated. In one embodiment of the present invention, the data analysis unit (15) may calculate the distance to the SP data heatmap when a center line is selected through the method described above, and thereby obtain center line bias information. At this time, if the distance between the heatmap and the center line exceeds a preset distance, the data analysis unit may be configured to select the predicted subject who provided the image as the primary observation target and transmit the brain imaging image to medical staff.
[0033] FIG. 2 illustrates an embodiment of the brain centerline bias information analysis method of the present invention. For convenience of explanation, FIG. 1 will be used for the following description, but the present invention is not limited thereto and may use devices, systems, and terminals capable of performing various similar functions or operations.
[0034] A method (10) for analyzing brain centerline bias information according to one embodiment of the present invention is formed to acquire a brain imaging image, learn from it to acquire a heatmap for basic data, and then use the heatmap to acquire brain cross-sectional analysis information including centerline information and bias information.
[0035] To this end, a brain centerline bias information analysis method (10) according to one embodiment of the present invention may be formed to include a basic data acquisition step (S11), a data learning step (S13), and a data analysis step (S15) as shown in FIG. 2.
[0036] The basic data acquisition step (S11) is configured to acquire basic data from a brain cross-section image. Here, the basic data may be a brain cross-section image acquired from training data providers included in a pre-configured training data pool. Additionally, in the present invention, only images in which a description of the image is included as text may be acquired as basic data. That is, even if the brain cross-section image is acquired from training data providers included in a pre-configured training data pool, information that does not include a description of the image is not selected as basic data.
[0037] In addition, in one embodiment of the present invention, the basic data acquisition step (S11) may acquire a brain cross-sectional image including at least one of AF (Anterior Falx) data, PF (Posterior Falx) data, and SP (Septum Pellucidum) data as basic data. Furthermore, the basic data may be formed to include coordinate information based on a preset point of the brain cross-sectional image, and the coordinate information may be coordinate information for the point where the above-described AF data, PF data, and SP data are located.
[0038] The data learning step (S13) of the present invention is configured to perform data learning on brain cross-sectional images using basic data and a preset algorithm, and to obtain a heatmap of the basic data as a learning result. A preset algorithm according to one embodiment of the present invention is illustrated in FIG. 3. Referring to FIG. 3, the data learning step (S13) of the present invention is configured to perform encoding on the acquired basic data, learn the encoding result, and then use a decoder to upsample the reduced resolution again in order to output the learning result.
[0039] A pre-configured algorithm according to one embodiment of the present invention may be a Cross Position Attention Module (CPAM). More specifically, in one embodiment of the present invention, the basic data is formed to include text information as described above when acquired through the data acquisition step (S11). The data learning step (S13) of the present invention may be formed to learn text information regarding the corresponding brain cross-section image together during the process of learning the encoding result after performing encoding to utilize it. Subsequently, when performing decoding, the data may be formed to output an image containing the learning result by including the learning information.
[0040] In one embodiment of the present invention, the data learning step (S13) is formed to obtain a heatmap for basic data as a learning result, and the heatmap can be obtained as an AF data heatmap, a PF data heatmap, and an FP data heatmap for AF data, PF data, and SP data, respectively. In addition, in one embodiment of the present invention, a Gaussian heatmap may be used to obtain this.
[0041] A heatmap according to one embodiment of the present invention can be represented as shown in FIG. 4. The heatmap is used, for example, to predict the location of a corresponding landmark. That is, since the basic data of the present invention includes location information of AF, PF, and SP, it is possible to check how the locations of the corresponding points show a distribution within the brain imaging image using this information, and by using this heatmap, it can be formed so that the location can be predicted even if an image that does not include the location information is acquired later.
[0042] Meanwhile, when using a heatmap, the positional information of the SP may vary depending on the degree of bias of the brain centerline. Therefore, in one embodiment of the present invention, learning is performed using additional text information to correct the heatmap by correcting this information, and through the results of the learning, the SP data heatmap may further include information regarding centerline bias.
[0043] The data analysis step (S15) is configured to obtain centerline information for a brain cross-section using a heatmap and to obtain brain cross-section analysis information including centerline bias information. The data analysis step (S15) according to one embodiment of the present invention is configured to obtain brain cross-section imaging data for which prediction is desired, obtain centerline information by comparing it with a learning result, and then determine the degree of centerline bias.
[0044] FIG. 5 shows an example of selecting a centerline in the present invention. Referring to FIG. 5, a data analysis step (S15) according to one embodiment of the present invention is formed to generate a centerline and obtain centerline information. As shown in FIG. 5, the centerline may be a line connecting AF and PF. In an embodiment of the present invention, the data analysis step (S15) may generate a centerline using a heatmap, and in this case, an AF data heatmap and a PF data heatmap may be used.
[0045] The data analysis step (S15) can be configured to determine the degree of bias using the distance between the centerline and the SP data heatmap when a centerline is generated. In one embodiment of the present invention, the data analysis step (S15) can calculate the distance to the SP data heatmap and obtain centerline bias information by selecting the centerline through the method described above. At this time, if the distance between the heatmap and the centerline exceeds a preset distance, the prediction subject who provided the image may be selected as the primary observation subject and the corresponding brain imaging image may be transmitted to medical staff.
[0046] The method for analyzing brain centerline bias information according to the embodiments of the present invention described above can be implemented as an application (computer program) stored on a computer storage medium.
[0047] Here, the computer may include a brain centerline bias information analysis system.
[0048] The computer operating system may be an operating system such as Windows or Macintosh installed on general PCs such as desktops and laptops, or a mobile-specific operating system such as iOS or Android installed on mobile devices such as smartphones and tablet PCs.
[0049] The brain centerline bias information analysis method according to the embodiments of the present invention described above is implemented as an application (i.e., a computer program) that is basically installed on a computer or installed by a user, and can be stored (recorded) on a computer-readable storage medium.
[0050] In this way, in order for a computer to read a program recorded on a storage medium and execute the brain centerline bias information analysis method according to the embodiments implemented as a program, the aforementioned application (application program) may include code coded in a computer language such as C, C++, JAVA, or machine language that can be read by a computer processor (CPU).
[0051] Such code may include functional code related to functions that define the aforementioned functions, and may also include control code related to execution procedures necessary for a computer processor to execute the aforementioned functions according to a predetermined procedure.
[0052] In addition, this code may further include memory reference-related code regarding where (address) in the computer's internal or external memory additional information or media required for the computer's processor to execute the aforementioned functions should be referenced.
[0053] In addition, if the computer processor needs to communicate with any other computer or server located remotely in order to execute the aforementioned functions, the code may further include communication-related code regarding how the computer processor must communicate with any other computer or server located remotely using the computer's communication module (e.g., wired and / or wireless communication module), and what information or media must be transmitted or received during communication.
[0054] Furthermore, the functional program for implementing the embodiments and the related code and code segments, etc., may be easily inferred or modified by programmers skilled in the art to which the present invention belongs, taking into account the system environment of a computer that reads a storage medium and executes the program.
[0055] In addition, a computer-readable storage medium storing a program as described above may be distributed across networked computer systems, allowing computer-readable code to be stored and executed in a distributed manner. In this case, one or more of the distributed computers may execute some of the functions presented above and transmit the results to one or more of the other distributed computers, and the computer receiving the results may also execute some of the functions presented above and provide the results to other distributed computers.
[0056] A computer-readable storage medium storing an application for executing a brain centerline bias information analysis method according to embodiments of the present invention as described above may include, for example, ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical media storage device, etc.
[0057] In addition, a computer-readable storage medium storing an application, which is a program for executing a brain centerline bias information analysis method according to embodiments of the present invention, may be a storage medium (e.g., a hard disk, etc.) included in an application provider server including an application store server, a web server related to the application or the service, etc., or the application provider server itself, or another computer or storage medium storing the program.
[0058] A computer capable of reading a storage medium storing an application, which is a program for executing a method for analyzing brain centerline bias information according to embodiments of the present invention, may include not only general PCs such as general desktops or laptops, but also mobile terminals such as smartphones, tablet PCs, PDAs (Personal Digital Assistants), and mobile communication terminals, and should be interpreted as any device capable of computing.
[0059] The foregoing description is merely an illustrative explanation of the technical concept of the present invention, and those skilled in the art to which the present invention pertains will be able to make various modifications and variations within the scope of the essential characteristics of the present invention. Accordingly, the embodiments disclosed in the present invention are intended to explain, not limit, the technical concept of the present invention, and the scope of the technical concept of the present invention is not limited by these embodiments. The scope of protection of the present invention shall be interpreted by the claims below, and all technical concepts within an equivalent scope shall be interpreted as being included within the scope of rights of the present invention. Explanation of the symbols
[0060] 1: Brain Centerline Bias Information Analysis System 11: Basic Data Acquisition Unit 13: Data Learning Section 15: Data Analysis Department
Claims
Claim 1 A basic data acquisition unit that acquires basic data from a brain cross-section image; a data learning unit that performs data learning on the brain cross-section image through the basic data and a preset algorithm, and acquires a heatmap for the basic data as a learning result; The system includes a data analysis unit that acquires centerline information for a brain cross-section using the heatmap and acquires brain cross-section analysis information including bias information of the centerline; wherein the pre-set algorithm is an algorithm that learns together the brain cross-section image and text information including a description of the brain cross-section image; wherein the basic data includes at least one of AF (Anterior Falx) data, PF (Posterior Falx) data, and SP (Septum Pellucidum) data; wherein the pre-set algorithm is formed to learn a compressed image using an encoder and to upsample the reduced resolution using a decoder; wherein the heatmap is acquired through a Gaussian heatmap, and the heatmaps of the AF data, PF data, and SP data are each acquired through the Gaussian heatmap; wherein the pre-set algorithm is a CPAM (Cross Position Attention Module); wherein the centerline information is acquired through the AF data heatmap and the PF data heatmap, and the bias information is acquired using the centerline information and the SP data heatmap. Centerline Bias Information Analysis System. Claim 2 delete Claim 3 A brain centerline bias information analysis system according to claim 1, wherein the basic data includes coordinate information based on a preset point of the brain cross-section image. Claim 4 delete Claim 5 delete Claim 6 delete Claim 7 delete Claim 8 A basic data acquisition step of acquiring basic data from a brain cross-section image using a basic data acquisition unit; a data learning step of performing data learning on the brain cross-section image using the basic data and a preset algorithm using a data learning unit, and acquiring a heatmap for the basic data as a learning result; The method includes a data analysis step of obtaining centerline information for a brain cross-section using the heatmap through a data analysis unit, and obtaining brain cross-section analysis information including bias information of the centerline; wherein the pre-set algorithm is an algorithm that learns together the brain cross-section imaging image and text information including a description of the brain cross-section imaging image; wherein the basic data includes at least one of AF (Anterior Falx) data, PF (Posterior Falx) data, and SP (Septum Pellucidum) data; wherein the pre-set algorithm is formed to learn a compressed image using an encoder and to upsample the reduced resolution using a decoder; wherein the heatmap is obtained through a Gaussian heatmap, and the heatmaps of the AF data, PF data, and SP data are each obtained through the Gaussian heatmap; wherein the pre-set algorithm is a CPAM (Cross Position Attention Module); wherein the centerline information is obtained through the AF data heatmap and the PF data heatmap, and the bias information is obtained using the centerline information and the SP data heatmap Method for analyzing acquired brain centerline bias information.
Citation Information
Patent Citations
Brain midline recognition method and device, equipment and storage medium
CN113706447A
Three-dimensional brain midline detection method and device, computer equipment and storage medium
CN116758139A
GAN-based artificial intelligence high-quality video compression system
KR1020230103257A
Automatic positioning method and device of brain midline
CN114419031A
