System-on-a-chip for fluorescence-magnetic resonance dual-mode imaging

Abstract

The present application relates to the technical field of medical imaging equipment, in particular to a system on chip for fluorescence-magnetic resonance dual-mode imaging, comprising: a fluorescence image preprocessing module and a compensation register, the fluorescence image preprocessing module is connected with the compensation register and reads diffusion compensation parameters from the compensation register; the fluorescence image preprocessing module carries out diffusion compensation on the current input fluorescence image according to the diffusion compensation parameters to obtain a preprocessed image. The beneficial effect is that the corresponding diffusion compensation parameters are called according to the input fluorescence image to carry out diffusion compensation on the fluorescence image, the influence of tissue diffusion on the actual imaging result is eliminated and a preprocessed image is generated, and the preprocessed image and the magnetic resonance image are synthesized, avoiding the problem of lesion size and position deviation in the two images during synthesis.

Description

Technical Field

[0001] This invention relates to the field of medical imaging equipment technology, and more specifically to a system-on-a-chip for fluorescence-magnetic resonance dual-mode imaging. Background Technology

[0002] Molecular imaging technology can specifically identify target molecules within living organisms and highly sensitively visualize their concentration and distribution in biological samples using signal imaging methods (magnetic, optical, or radioactive), obtaining relevant information at the cellular and molecular levels. Therefore, it plays a central role in life science research and clinical diagnosis. To achieve this process, specific tracers are typically used. Labels such as radioactive isotopes, fluorescent dyes, or magnetic materials are injected into the organism, and then imaging equipment is used to detect and image these labels, thereby obtaining information about the distribution, metabolism, and interactions of molecules within the organism. Tracers are substances used to label and track molecular activity within living organisms. They can be radioactive isotopes, fluorescent dyes, magnetic materials, etc.

[0003] In existing technologies, due to the weak tissue penetration of fluorescent tracers and the insufficient sensitivity of magnetic resonance imaging (MRI), it is often impossible to use a single imaging technique to determine all the functional and structural information of the analyte. Therefore, existing technologies employ dual-mode or multi-mode imaging techniques to detect tissues. For example, Chinese patent CN114656447A discloses a type of near-infrared fluorescence and magnetic resonance Aβ dual-mode imaging probe based on high spatiotemporal resolution, its preparation method, and its application. This involves pre-synthesizing a specific dual-mode imaging probe, injecting it into a biological organism, and then scanning it using both a fluorescence imaging system and an MRI system to acquire fluorescence and MRI images, which are then used to determine the activity within biological tissues.

[0004] However, in practice, the inventors discovered that the location of the lesion can lead to differences in fluorescence characterization. For example, probes at lesion sites on the superficial surface and inside the body will exhibit different fluorescence intensities and diffusion patterns due to variations in tissue distribution. These differences can cause certain deviations during imaging; for instance, the location and size of the fluorescent area may not match the lesion area obtained from magnetic resonance imaging. Such errors are easily identified and corrected by the human eye—for example, by comparing only areas with higher fluorescence intensity as the actual probe location and ignoring artifacts. However, when these images are further used to train related artificial intelligence models, the errors are sufficient to prevent the model from converging correctly. Summary of the Invention

[0005] To address the aforementioned problems in the existing technology, an on-chip system for fluorescence-magnetic resonance dual-mode imaging is provided.

[0006] The specific technical solution is as follows:

[0007] A system-on-a-chip (SoC) for fluorescence-magnetic resonance dual-mode imaging is disclosed. The SoC is connected to an external fluorescence imaging system and a magnetic resonance system, respectively. The SoC is used to reconstruct an image from a gradient sequence output by the magnetic resonance system and then synthesize it with a fluorescence image output by the fluorescence imaging system. The SoC includes a fluorescence image preprocessing module and a compensation register. The fluorescence image preprocessing module is connected to the compensation register and reads diffusion compensation parameters from the compensation register.

[0008] The fluorescence image preprocessing module performs diffusion compensation on the currently input fluorescence image according to the diffusion compensation parameters to obtain a preprocessed image.

[0009] On the other hand, the register pre-stores multiple sets of depth compensation parameters corresponding to different water contents;

[0010] Each set of depth compensation parameters contains multiple diffusion compensation parameters corresponding to different tissue depths.

[0011] On the other hand, the on-chip system also includes:

[0012] The measurement module, upon receiving the input fluorescence image, outputs a measurement command to the magnetic resonance system.

[0013] The magnetic resonance system acquires measurement sequences from the tissue under test according to the measurement instructions;

[0014] The measurement module receives the measurement sequence and determines the moisture content and tissue depth according to the measurement sequence;

[0015] The fluorescence image preprocessing module reads the diffusion compensation parameters according to the water content and the tissue depth.

[0016] On the other hand, the measurement module includes:

[0017] A signal strength extraction module, which acquires signal anomaly intervals and multiple sets of signal strength sample values ​​from the measurement sequence;

[0018] A depth generation module, which is connected to the signal intensity extraction module, determines the tissue depth according to the echo time corresponding to the signal abnormality interval;

[0019] A weighted image generation module, which generates a T1 weighted image and a T2 weighted image according to the measurement sequence;

[0020] A moisture content measurement module is connected to the signal intensity extraction module and the weighted image generation module, respectively. The moisture content measurement module generates the moisture content according to the signal intensity sampling value, the T1 weighted image and the T2 weighted image.

[0021] On the other hand, the fluorescence image preprocessing module includes:

[0022] An image segmentation module, wherein the image segmentation module segments at least one fluorescent region from the fluorescent image;

[0023] A region erosion module is connected to the image segmentation module. The region erosion module performs erosion processing on each fluorescent region based on the diffusion compensation parameters to obtain a diffusion-compensated preprocessed image.

[0024] On the other hand, the image segmentation module includes:

[0025] A grayscale extraction module extracts global grayscale values ​​from the fluorescence image;

[0026] A threshold generation module, which is connected to the grayscale extraction module, generates an adaptive threshold based on the global grayscale value;

[0027] A threshold segmentation module, wherein the threshold segmentation module segments the fluorescence image according to the adaptive threshold to filter out multiple candidate regions;

[0028] A connectivity detection module is connected to the threshold segmentation module. The connectivity detection module performs connectivity detection on each of the candidate regions and splits the candidate regions that are artifacts to obtain the fluorescent regions.

[0029] On the other hand, the regional corrosion module includes:

[0030] An edge extraction module extracts the edge contours of the fluorescent regions respectively;

[0031] The erosion module shrinks the edge contour along the normal direction according to the diffusion compensation parameters to form the preprocessed image.

[0032] On the other hand, the on-chip system also includes:

[0033] A magnetic resonance image reconstruction module, wherein the magnetic resonance image reconstruction module receives the gradient sequence and performs image reconstruction according to the gradient sequence to obtain a magnetic resonance image;

[0034] An image overlay module is provided, which is connected to the fluorescence image preprocessing module and the magnetic resonance image reconstruction module respectively. The image overlay module aligns the preprocessed image and the magnetic resonance image and then outputs them.

[0035] The above technical solution has the following advantages or beneficial effects:

[0036] To address the issue that the fluorescence portion of existing fluorescence-magnetic resonance dual-mode imaging systems may shift when superimposed with magnetic resonance images due to differences in fluorescence diffusion across different tissues, this solution introduces the aforementioned on-chip system during image processing. This on-chip system pre-configures a fluorescence image preprocessing module and a compensation register, which are used to call the corresponding diffusion compensation parameters to perform diffusion compensation on the fluorescence image based on the input fluorescence image. This eliminates the influence of tissue diffusion on the actual imaging results and generates a preprocessed image. The preprocessed image and the magnetic resonance image are then combined, avoiding the problem of lesion size and location shifts in the two images during the synthesis process. Attached Figure Description

[0037] Embodiments of the invention will be described more fully with reference to the accompanying drawings. However, the drawings are for illustration and explanation only and do not constitute a limitation on the scope of the invention.

[0038] Figure 1 This is an overall schematic diagram of an embodiment of the present invention;

[0039] Figure 2 This is a schematic diagram of the measurement module in an embodiment of the present invention;

[0040] Figure 3 This is a schematic diagram of the fluorescence image preprocessing module in an embodiment of the present invention;

[0041] Figure 4 This is a schematic diagram of the mode adjustment module in an embodiment of the present invention;

[0042] Figure 5 This is a schematic diagram of the regional corrosion module in an embodiment of the present invention. Detailed Implementation

[0043] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0044] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0045] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, but this is not intended to limit the scope of the invention.

[0046] This invention includes:

[0047] A system-on-chip A1 for fluorescence-magnetic resonance dual-mode imaging, such as Figure 1 As shown, the on-chip system A1 is connected to the external fluorescence imaging system A2 and magnetic resonance system A3 respectively. The on-chip system A1 is used to perform image reconstruction on the gradient sequence output by the magnetic resonance system A3 and then synthesize it with the fluorescence image output by the fluorescence imaging system A2. The on-chip system A1 includes a fluorescence image preprocessing module 1 and a compensation register 2. The fluorescence image preprocessing module 1 is connected to the compensation register 2 and reads the diffusion compensation parameters from the compensation register 2.

[0048] The fluorescence image preprocessing module 1 performs diffusion compensation on the currently input fluorescence image according to the diffusion compensation parameters to obtain a preprocessed image;

[0049] The compensation register 2 is pre-stored with multiple sets of depth compensation parameters corresponding to different water contents;

[0050] Each set of depth compensation parameters contains multiple diffusion compensation parameters corresponding to different tissue depths.

[0051] Specifically, in response to the problem that image position shifts may occur due to fluorescence diffusion during image overlay in existing dual-mode imaging methods, this embodiment addresses this issue by setting the fluorescence image preprocessing module 1 to read diffusion compensation parameters from the compensation register 2 in the on-chip system A1 used for image synthesis. Based on the diffusion compensation parameters, diffusion correction is performed on the fluorescence region in the fluorescence image, thereby removing artifacts caused by tissue diffusion in the fluorescence image. This avoids the problem of position shifts during subsequent image synthesis, which could affect the construction of the artificial intelligence model training set.

[0052] Furthermore, to effectively remove the diffuse portion, in this embodiment, multiple sets of depth compensation parameters corresponding to different water contents are pre-stored in the compensation register 2, and each set of depth compensation parameters contains multiple diffusion compensation parameters corresponding to different tissue depths. The aforementioned diffusion compensation parameters are obtained in advance through experimental measurements. The experimental process includes: dividing different parts of the target organism according to water content to obtain multiple tissue structure water content conditions; using simulated materials with corresponding water contents, embedding fluorescent indicators at different tissue depths, and controlling the fluorescent indicators to have the same fluorescence intensity as the target biological probe; taking images under the same illumination conditions to obtain sample images; measuring the fluorescence diffusion at the corresponding tissue depth by measuring the fluorescence center region and surrounding diffusion region in the sample image, and taking negative values ​​as the corresponding diffusion compensation parameters. The experiment is repeated multiple times to obtain the depth compensation parameter sets, which are then stored in the compensation register 2. By calling the diffusion compensation parameters, the fluorescence image preprocessing module 1 can easily process the fluorescence images of biological tissues at any location, thereby achieving the effect of eliminating diffusion.

[0053] In one embodiment, the on-chip system A1 further includes:

[0054] Magnetic resonance image reconstruction module 3 receives gradient sequences and performs image reconstruction according to the gradient sequences to obtain magnetic resonance images;

[0055] Image overlay module 4 is connected to fluorescence image preprocessing module 1 and magnetic resonance image reconstruction module 3 respectively. Image overlay module 4 aligns the preprocessed image and magnetic resonance image and then outputs them.

[0056] Specifically, to implement the corresponding image processing procedures and facilitate subsequent training of artificial intelligence models, this embodiment includes a magnetic resonance image reconstruction module 3 on the system-on-a-chip. This module 3 can perform signal processing on gradient sequences according to existing technologies to reconstruct magnetic resonance images. These magnetic resonance images may be two-dimensional or three-dimensional magnetic resonance image sequences. The image overlay module 4, after generating the preprocessed image and the magnetic resonance image, can extract the fluorescence distribution in the fluorescence image and the signal intensity distribution in the magnetic resonance image, thereby registering and aligning the two images. When the magnetic resonance image is a three-dimensional magnetic resonance image sequence, the image of the central frame is extracted for alignment.

[0057] In one embodiment, the on-chip system A1 further includes:

[0058] Measurement module 5, upon receiving the input fluorescence image, outputs a measurement command to magnetic resonance system A3;

[0059] The magnetic resonance imaging system A3 acquires measurement sequences from the tissue under test according to the measurement instructions;

[0060] Measurement module 5 receives the measurement sequence and determines the moisture content and tissue depth according to the measurement sequence;

[0061] The fluorescence image preprocessing module 1 reads diffusion compensation parameters based on water content and tissue depth.

[0062] Specifically, in order to determine the tissue depth and water content and thus invoke the corresponding diffusion compensation parameters, a measurement module 5 is provided in the system-on-a-chip A1 in this embodiment. The measurement module 5 is enabled by the image input pin of the system-on-a-chip A1. When the system-on-a-chip A1 receives a fluorescence image, the measurement module 5 controls the magnetic resonance system A3 to acquire a specific measurement sequence of the current tissue to be tested, and calculates the tissue depth with probe and the water content of the tissue to be tested according to the measurement sequence, so that the fluorescence image preprocessing module 1 can read the correct diffusion compensation parameters according to the water content and tissue depth.

[0063] In one embodiment, such as Figure 2 As shown, the measurement module 5 includes:

[0064] Signal strength extraction module 51 acquires signal abnormality intervals and multiple sets of signal strength sample values ​​from the measurement sequence;

[0065] Depth generation module 52 is connected to signal strength extraction module 51. Depth generation module 52 determines tissue depth according to the echo time corresponding to the signal abnormality interval.

[0066] The weighted image generation module 53 generates a T1 weighted image and a T2 weighted image according to the measurement sequence.

[0067] The moisture content measurement module 54 is connected to the signal intensity extraction module 51 and the weighted image generation module 53 respectively. The moisture content measurement module 54 generates the moisture content according to the signal intensity sampling value, the T1 weighted image and the T2 weighted image.

[0068] Specifically, to achieve effective measurement of tissue water content and probe depth, this embodiment pre-configures a signal intensity extraction module 51, which can extract signal abnormality intervals in the measurement sequence. Generally, since a dual-mode probe has been pre-injected into the organism, it can generate a specific magnetic response at the target location, which is then reflected as a relatively strong signal peak in the gradient sequence. The depth generation module 52 easily calculates the corresponding tissue depth by receiving the extracted signal abnormality intervals and determining the echo time. Simultaneously, the measurement sequence generated based on specific parameters can be used by the weighted image generation module 53 to generate T1-weighted and T2-weighted images, respectively. Finally, the water content measurement module 54 estimates the water content within the tissue based on the signal intensity sampling value, the T1-weighted image, and the T2-weighted image.

[0069] In one embodiment, such as Figure 3 As shown, the fluorescence image preprocessing module 1 includes:

[0070] Image segmentation module 11, which segments at least one fluorescent region from the fluorescence image;

[0071] The region erosion module 12 is connected to the image segmentation module 11. The region erosion module 12 performs erosion processing on each fluorescent region based on the diffusion compensation parameters to obtain a preprocessed image after diffusion compensation.

[0072] Specifically, to achieve better fluorescence region correction, in this embodiment, an image segmentation module 11 and a region erosion module 12 are respectively set in the fluorescence image preprocessing module 1. The image segmentation module 11 can segment the fluorescence image based on a corresponding image segmentation algorithm to obtain the fluorescence region. This fluorescence region includes a central region corresponding to the fluorescence imaging portion and a surrounding tissue diffusion region, and irrelevant parts of the background are removed. Subsequently, the region erosion module 12 performs erosion processing on each fluorescence region based on diffusion compensation parameters, thereby removing and filling the surrounding tissue diffusion region to obtain a preprocessed image, avoiding subsequent alignment errors.

[0073] In one embodiment, such as Figure 4 As shown, the image segmentation module 11 includes:

[0074] Grayscale extraction module 111 extracts global grayscale values ​​from the fluorescence image;

[0075] Threshold generation module 112 is connected to grayscale extraction module 111. Threshold generation module 112 generates an adaptive threshold based on the global grayscale value.

[0076] Threshold segmentation module 113 segments the fluorescence image according to an adaptive threshold to filter out multiple candidate regions;

[0077] The connected domain detection module 114 is connected to the threshold segmentation module 113. The connected domain detection module 114 performs connected domain detection on each candidate region and splits the candidate regions that are artifacts to obtain the fluorescent regions.

[0078] Specifically, to achieve better extraction of fluorescent regions, in this embodiment, the grayscale extraction module 111 extracts the fluorescent image using the sliding window method and then calculates its global grayscale value. Subsequently, the threshold generation module 112 adjusts the segmentation threshold according to the global grayscale value to generate an adaptive threshold, avoiding segmentation errors caused by changes in tissue brightness under different shooting conditions. Then, the threshold segmentation module 113 segments the fluorescent image according to the adaptive threshold, extracting the brighter parts as candidate regions. At this point, the candidate regions may contain complete fluorescent regions, or they may be missegmented as the same fluorescent region due to adhesion or overlapping artifact regions. Therefore, the connectivity detection module 114 further performs connectivity detection based on pixel brightness and splits the candidate regions where the connectivity region is an artifact to obtain the fluorescent region, thereby achieving a better segmentation effect.

[0079] In one embodiment, such as Figure 5 As shown, the regional corrosion module 12 includes:

[0080] Edge extraction module 121 extracts the edge contours of the fluorescent regions respectively;

[0081] The erosion module 122 shrinks the edge contour along the normal direction according to the diffusion compensation parameters to form a preprocessed image.

[0082] Specifically, to achieve a better removal effect on artifact regions, in this embodiment, after determining the fluorescent regions, the edge extraction module 121 extracts the edge contours for each fluorescent region. Subsequently, the erosion module 122 shrinks the edge contours at each point along the normal direction according to the diffusion compensation parameters, thereby removing the artifact regions. After removing the artifact regions, depending on the embodiment, the removed parts can be supplemented by nearest pixel filling or other methods to form a new preprocessed image.

[0083] The above are merely preferred embodiments of the present invention and are not intended to limit the implementation methods and protection scope of the present invention. Those skilled in the art should recognize that any equivalent substitutions and obvious changes made based on the description and illustrations of the present invention should be included within the protection scope of the present invention.

Claims

1. A system-on-chip for fluorescence-magnetic resonance dual-mode imaging, wherein the system-on-chip is connected to an external fluorescence imaging system and a magnetic resonance system respectively, and the system-on-chip is used to perform image reconstruction on the gradient sequence output by the magnetic resonance system and then synthesize it with the fluorescence image output by the fluorescence imaging system, characterized in that, The system-on-chip includes a fluorescence image preprocessing module and a compensation register. The fluorescence image preprocessing module is connected to the compensation register and reads diffusion compensation parameters from the compensation register. The fluorescence image preprocessing module performs diffusion compensation on the currently input fluorescence image according to the diffusion compensation parameters to obtain a preprocessed image; The compensation register is pre-stored with multiple sets of depth compensation parameters corresponding to different water contents; Each set of depth compensation parameters contains multiple diffusion compensation parameters corresponding to different tissue depths; The on-chip system also includes: The measurement module, upon receiving the input fluorescence image, outputs a measurement command to the magnetic resonance system. The magnetic resonance system acquires measurement sequences from the tissue under test according to the measurement instructions; The measurement module receives the measurement sequence and determines the moisture content and tissue depth according to the measurement sequence; The fluorescence image preprocessing module reads the diffusion compensation parameters according to the water content and the tissue depth; The diffusion compensation parameters were obtained in advance through experimental measurement; The experimental procedure includes dividing different parts of the target organism according to their water content to obtain multiple tissue structures with varying water content; using simulated materials with corresponding water content, fluorescent indicators are embedded at different tissue depths, and the fluorescence intensity of the fluorescent indicators and the target biological probe is controlled to be the same. Sample images are obtained by taking pictures under the same lighting conditions; by measuring the fluorescence center region and the surrounding diffusion region in the sample images, the fluorescence diffusion at the corresponding tissue depth is obtained, and negative values ​​are taken as the corresponding diffusion compensation parameters.

2. The system-on-a-chip according to claim 1, characterized in that, The measurement module includes: A signal strength extraction module, which acquires signal anomaly intervals and multiple sets of signal strength sample values ​​from the measurement sequence; A depth generation module, which is connected to the signal intensity extraction module, determines the tissue depth according to the echo time corresponding to the signal abnormality interval; A weighted image generation module, which generates a T1 weighted image and a T2 weighted image according to the measurement sequence; A moisture content measurement module is connected to the signal intensity extraction module and the weighted image generation module, respectively. The moisture content measurement module generates the moisture content according to the signal intensity sampling value, the T1 weighted image and the T2 weighted image.

3. The system-on-a-chip according to claim 1, characterized in that, The fluorescence image preprocessing module includes: An image segmentation module, wherein the image segmentation module segments at least one fluorescent region from the fluorescent image; A region erosion module is connected to the image segmentation module. The region erosion module performs erosion processing on each fluorescent region based on the diffusion compensation parameters to obtain a diffusion-compensated preprocessed image.

4. The system-on-a-chip according to claim 3, characterized in that, The image segmentation module includes: A grayscale extraction module extracts global grayscale values ​​from the fluorescence image; A threshold generation module, which is connected to the grayscale extraction module, generates an adaptive threshold based on the global grayscale value; A threshold segmentation module, wherein the threshold segmentation module segments the fluorescence image according to the adaptive threshold to filter out multiple candidate regions; A connectivity detection module is connected to the threshold segmentation module. The connectivity detection module performs connectivity detection on each of the candidate regions and splits the candidate regions that are artifacts to obtain the fluorescent regions.

5. The system-on-a-chip according to claim 3, characterized in that, The regional corrosion module includes: An edge extraction module extracts the edge contours of the fluorescent regions respectively; The erosion module shrinks the edge contour along the normal direction according to the diffusion compensation parameters to form the preprocessed image.

6. The system-on-a-chip according to claim 1, characterized in that, The on-chip system also includes: A magnetic resonance image reconstruction module, wherein the magnetic resonance image reconstruction module receives the gradient sequence and performs image reconstruction according to the gradient sequence to obtain a magnetic resonance image; An image overlay module is provided, which is connected to the fluorescence image preprocessing module and the magnetic resonance image reconstruction module respectively. The image overlay module aligns the preprocessed image and the magnetic resonance image and then outputs them.

Images