KR-102964018-B1 - Method for Image Correction Based on Heart Position Recognition Using Temporary Images and Computer Device for Performing the Same

KR102964018B1KR 102964018 B1KR102964018 B1KR 102964018B1KR-102964018-B1

Abstract

The present invention relates to a method for correcting heart images. In a method for correcting images based on the position recognition of the heart performed on a computing device according to the present invention, a first image can be acquired during the process of acquiring heart images according to a preset sequence using a medical imaging device. The position of the heart can be inferred from the acquired first image using a trained neural network, and the position of the heart within the acquired heart image according to the sequence can be corrected based on the inferred position of the heart. The present invention can reduce artifacts by estimating and correcting movement in real time by utilizing the segmentation results of intermediate images.

Inventors

심재윤
양영중
김판기

Assignees

주식회사 팬토믹스

Dates

Publication Date: 20260512
Application Date: 20250206

Claims (11)

In an image correction method based on heart location recognition performed on a computing device, A step of acquiring a first image as a temporary image acquired during the initial section of the magnetization recovery process following the application of a plurality of RF pulses, during the process of acquiring a cardiac image according to a preset sequence using a medical imaging device; A step of inferring the location of the heart from the acquired first image by segmenting the anatomical boundaries of the heart using a learned neural network that includes skip connections between an encoder and a decoder, with the acquired first image as input using the learned neural network; and Based on the position of the heart within the first image inferred above, A method comprising the step of correcting local distortions of the heart in a second image by repeatedly aligning them to match a corresponding region in a first image, using a model modeled by learning non-linear shape changes according to the contraction and relaxation cycles of the heart within the sequence, based on common points between extracted ventricles or regions according to pixel intensity changes. Image correction method based on heart location recognition.
In Article 1, The above sequence is, Determined by the reference recovery rate of protons within cardiac tissue inverted by RF (Radio Frequency) pulses, Image correction method based on heart location recognition.
In Article 1, The above first image is in the process of applying a plurality of RF pulses, Acquired from the recovery signal collected after the application of the first RF pulse, Image correction method based on heart location recognition.
In Article 1, The step of inferring the location of the heart is, Inferring the location of the heart by providing the output of the encoder within the neural network of the first image as the input of the decoder, Image correction method based on heart location recognition.
In Article 4, The step of correcting the position of the heart is, A step of extracting a region within the first image; A step comprising aligning the position of the heart within the second image using the region of the second image corresponding to the extracted region, Image correction method based on heart location recognition.
processor, and It includes a memory that communicates with the above processor, and The above memory stores instructions that cause the processor to perform operations, and The above operations are, A step of acquiring a first image as a temporary image acquired during the initial section of the magnetization recovery process following the application of a plurality of RF pulses, during the process of acquiring a cardiac image according to a preset sequence using a medical imaging device; A step of inferring the location of the heart from the acquired first image by segmenting the anatomical boundaries of the heart using a learned neural network that includes skip connections between an encoder and a decoder, with the acquired first image as input using the learned neural network; and Based on the position of the heart within the first image inferred above, A method comprising the step of correcting local distortions of the heart in a second image by repeatedly aligning them to match a corresponding region in a first image, using a model modeled by learning non-linear shape changes according to the contraction and relaxation cycles of the heart within the sequence, based on common points between extracted ventricles or regions according to pixel intensity changes. Computer device.
In Article 6, The above sequence is, Determined by the reference recovery rate of protons within cardiac tissue inverted by RF (Radio Frequency) pulses, Computer device.
In Article 6, The above first image is in the process of applying a plurality of RF pulses, Acquired from the recovery signal collected after the application of the first RF pulse, Computer device.
In Article 6, The step of inferring the location of the heart is, Inferring the location of the heart by providing the output of the encoder within the neural network of the first image as the input of the decoder, Computer device.
In Article 9, The step of correcting the position of the heart is, A step of extracting a region within the first image; A step comprising aligning the position of the heart within the second image using the region of the second image corresponding to the extracted region, Computer device.
In Article 10, A step comprising calculating a change in the contraction or relaxation cycle of the heart, Computer device.

Description

Method for Image Correction Based on Heart Position Recognition Using Temporary Images and Computer Device for Performing the Same The present invention relates to a method for correcting cardiac images. Motion issues during MRI (Magnetic Resonance Imaging) scans are a critical problem in medical imaging. MRI is an imaging technique highly sensitive to the movement of objects or patients; therefore, even slight patient movements during a scan, or the imaging of moving organs such as the heart or lungs, can significantly degrade image quality due to such motion. MRI collects data in a frequency domain known as k-space, and when motion occurs, this k-space data becomes distorted, resulting in the appearance of blur or ghost artifacts in the image. In particular, these problems frequently occur during high-resolution scans that require long durations, due to voluntary or unavoidable patient movements. MRI images the internal structures of the human body using precise magnetic fields and radio frequency signals; however, when movement occurs, not only rotational and translational displacements but also fluctuations in the magnetic field surrounding the object can take place. These fluctuations cause image distortion, and the problem is particularly severe in areas with frequent movement, such as the heart or abdomen. Various technologies are being researched to address this, one of which is motion estimation and correction technology utilizing navigator signals. The prior art patent (Japanese Patent Publication JP 2024-523778 A, (July 2, 2024)) relates to a method for estimating and correcting the movement of an object and changes in the surrounding magnetic field during MRI imaging, and an apparatus for performing the same. In the present invention, movement and changes in the magnetic field are estimated in real time using navigator gradient segments included in the MRI sequence, and image quality is corrected based on this. In particular, the rotation angle and translational displacement amount are estimated, and a field offset according to changes in the magnetic field is calculated to minimize image distortion that may occur during MRI imaging. However, while existing methods tracked object movement using navigator signals and corrected distortions caused by changes in the magnetic field, this approach has limitations in immediate motion detection or fine-tuning. FIG. 1 is a diagram schematically illustrating a cardiac image correction system according to one embodiment of the present invention. FIG. 2 is a flowchart illustrating a cardiac image correction method according to one embodiment of the present invention. FIG. 3 is a figure showing a sequence according to one embodiment of the present invention. FIG. 4 is a figure showing the generation of a T1 image through the acquisition of an intermediate image according to an embodiment of the present invention. FIG. 5 is a diagram showing the operation of a neural network for extracting the location of the heart according to one embodiment of the present invention. FIG. 6 is a detailed flowchart illustrating a method for correcting the position of a heart according to one embodiment of the present invention. FIG. 7 is an exemplary diagram illustrating the alignment process of a computer device according to one embodiment of the present invention. FIG. 8 is an exemplary diagram showing the implementation of a server as a computing device for performing an image correction method according to the present embodiment. The following description merely illustrates the principles of the invention. Therefore, those skilled in the art may invent various devices that embody the principles of the invention and are included within the concept and scope of the invention, even if they are not explicitly described or illustrated in this specification. Furthermore, all conditional terms and embodiments listed in this specification are, in principle, explicitly intended only for the purpose of enabling an understanding of the concept of the invention and should be understood as not being limited to the embodiments and conditions specifically listed elsewhere. The aforementioned objectives, features, and advantages will become clearer through the following detailed description in conjunction with the attached drawings, and accordingly, a person skilled in the art to which the invention pertains will be able to easily implement the technical concept of the invention. In addition, in describing the invention, if it is determined that a detailed description of known technology related to the invention may unnecessarily obscure the essence of the invention, such detailed description will be omitted. Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the attached drawings. Magnetic resonance imaging (MRI) can calculate recovery times based on components by distinguishing the components of the recovery process, which involves