WO2024239285A1 - 数据采集方法、装置及系统 - Google Patents

数据采集方法、装置及系统 Download PDF

Info

Publication number
WO2024239285A1
WO2024239285A1 PCT/CN2023/096110 CN2023096110W WO2024239285A1 WO 2024239285 A1 WO2024239285 A1 WO 2024239285A1 CN 2023096110 W CN2023096110 W CN 2023096110W WO 2024239285 A1 WO2024239285 A1 WO 2024239285A1
Authority
WO
WIPO (PCT)
Prior art keywords
data
image
biological tissue
tissue sample
expression
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2023/096110
Other languages
English (en)
French (fr)
Inventor
刘超
夏天祎
史昶
郭俊甫
彭勇
张博涵
陆芹
李美
张勇
黎宇翔
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
BGI Shenzhen Co Ltd
Original Assignee
BGI Shenzhen Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by BGI Shenzhen Co Ltd filed Critical BGI Shenzhen Co Ltd
Priority to CN202380098640.8A priority Critical patent/CN121399661A/zh
Priority to PCT/CN2023/096110 priority patent/WO2024239285A1/zh
Publication of WO2024239285A1 publication Critical patent/WO2024239285A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
    • G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods

Definitions

  • the present disclosure relates to the field of data processing technology, and in particular to a data acquisition method, device and system.
  • spatiotemporal omics technology Since its development, spatiotemporal omics technology has been initially applied in neuroscience, pathology, developmental biology and other fields. It is difficult for spatiotemporal omics analysis on a two-dimensional plane to obtain the morphological structure and accurate position information of a three-dimensional space. How to break through the research on the two-dimensional plane and realize three-dimensional spatiotemporal omics research is of great technical challenge and important scientific significance. At present, the main research is on the three-dimensional reconstruction method of spatiotemporal omics to achieve a comprehensive analysis of the spatial distribution, interaction, gene protein expression and overall morphological structure of different cells at the three-dimensional level of tissue.
  • the current three-dimensional reconstruction of spatiotemporal omics is based only on the expression abundance information captured by the spatiotemporal group, and the expression data measured on the plane are rigidly registered. There are problems such as low accuracy of data reconstruction back to the original expression level, limited visualization effect, and insufficient data processing and post-display capabilities for large amounts of tissue samples.
  • the present disclosure proposes a data acquisition method, device and system, which is aimed at the spatiotemporal omics technology captured in situ, combines optical imaging information to obtain the internal and external tissue contours to assist in visual information positioning, and achieves cell-precision position alignment for each information acquisition site, and can effectively process and display large-scale high-resolution spatiotemporal group data.
  • a first aspect embodiment of the present disclosure provides a data acquisition method, which includes: slicing a biological tissue sample and performing spatiotemporal group sequencing on the slices to obtain expression data; performing image acquisition on the sliced biological tissue sample to obtain image data; aligning the expression data and/or image data to generate an outer contour and/or partition contour of the biological tissue sample; and generating three-dimensional display data of the biological tissue sample based on the outer contour and/or partition contour.
  • slicing a biological tissue sample and performing spatiotemporal group sequencing on the slices to obtain expression data includes: continuously slicing the biological tissue sample, performing spatiotemporal group sequencing once for each first predetermined number of slices to obtain expression data of the slices, wherein the first predetermined number is a positive integer.
  • the method further includes: preprocessing the expression data, wherein the preprocessing includes at least one of aggregation processing, quality control, logarithmic normalization, finding highly variable genes, and dimensionality reduction processing.
  • performing image acquisition on the sliced biological tissue sample to obtain image data includes: performing image acquisition on the sliced biological tissue sample once every second predetermined number of slices to obtain image data, wherein the second predetermined number is a positive integer.
  • registering expression data and/or image data to generate an outer contour and/or partition contour of a biological tissue sample includes: performing image segmentation processing on the image data to generate a segmentation result map; registering the segmentation result map to generate a registration result map; and generating the outer contour of the biological tissue sample based on the registration result map.
  • image segmentation processing of image data includes: converting the image data into HSV color space, extracting the image data of the H channel and converting it into a grayscale image; binarizing the image data through a preset image processing algorithm to determine a contour map of the binarization result; retaining the contour map that meets preset conditions and merging it with the image data.
  • the method further includes: acquiring a grayscale image and a histogram of the image data; performing maximum entropy threshold binarization processing on the image data; and merging the processed image with the image data.
  • registering the segmentation result image to generate a registration result image includes: performing spatial transformation processing and linear interpolation algorithm processing on the segmentation result image; performing mean square error similarity measurement on the processed segmentation result image and the registration reference image; and determining the optimal measurement solution as the registration result image.
  • generating the outer contour of the biological tissue sample based on the registration result image includes: stacking all the registration result images into an array file in a preset format; and processing the array file to obtain the outer contour of the biological tissue sample.
  • expression data and/or image data are registered to generate an outer contour and/or partition contour of a biological tissue sample, including: calculating the expression data to obtain an expression map; performing histogram equalization on the expression map and image data to obtain an image pair of the expression map-image data; and sequentially performing rigid registration and elastic registration on the image pair to generate a partition contour.
  • rigid registration of an image pair includes: downsampling the image pair respectively; extracting feature information of the image pair; and performing similarity estimation affine transformation on the image pair to obtain a rigid registration result.
  • elastic registration of image pairs includes: performing histogram equalization on the rigid registration results; inputting the rigid registration results into a deep learning model, outputting an intermediate result of the deep learning model, and obtaining a deformation field, wherein the deep learning model is obtained by training expression graph-image data pairs; and performing nearest neighbor interpolation transformation on the rigid registration results according to the deformation field to obtain an elastic registration result.
  • registering expression data and/or image data to generate an outer contour and/or partition contour of a biological tissue sample includes: registering expression data in two adjacent or similar slices to obtain registered expression data; and generating an outer contour and/or partition contour of the biological tissue sample based on the registered expression data.
  • generating three-dimensional display data of a biological tissue sample according to the outer layer contour and/or the partition contour includes: performing bioinformatics analysis on the registration results obtained by registering the expression data and/or the image data to obtain clustering and annotation results; constructing a three-dimensional array based on the clustering and annotation results, and adding spatial scale information; and using a three-dimensional rendering tool to render the three-dimensional display data of the biological tissue sample.
  • a second aspect of an embodiment of the present disclosure provides a data processing method, which includes: obtaining expression data and image data of a biological tissue sample, wherein the expression data is obtained by slicing the biological tissue sample and performing spatiotemporal group sequencing on the slices, and the image data is obtained by performing image acquisition on the biological tissue sample after slicing; calculating the expression data to obtain an expression graph; performing histogram equalization on the expression graph and the image data to obtain an image pair of the expression graph-image data; and sequentially performing rigid registration and elastic registration on the image pair to generate contour data of the biological tissue sample, and the contour data is used to generate three-dimensional display data of the biological tissue sample.
  • rigid registration of an image pair includes: downsampling the image pair respectively; extracting feature information of the image pair; and performing similarity estimation affine transformation on the image pair to obtain a rigid registration result.
  • elastic registration of image pairs includes: performing histogram equalization on the rigid registration results; inputting the rigid registration results into a deep learning model, outputting an intermediate result of the deep learning model, and obtaining a deformation field, wherein the deep learning model is obtained by training expression graph-image data pairs; and performing nearest neighbor interpolation transformation on the rigid registration results according to the deformation field to obtain an elastic registration result.
  • the third aspect of the present disclosure provides a data acquisition device, which includes: a sequencing module for slicing a biological tissue sample and performing spatiotemporal group sequencing on the slice to obtain expression data; a collection module for performing image collection on the sliced biological tissue sample to obtain image data; and a registration module for registering the expression data and/or the image data to generate a biological group data.
  • a fourth aspect of the present disclosure provides a data processing device, which includes: an acquisition module, used to acquire expression data and image data of a biological tissue sample, wherein the expression data is obtained by slicing the biological tissue sample and performing spatiotemporal group sequencing on the slices, and the image data is obtained by performing image acquisition on the biological tissue sample after slicing; a calculation module, used to calculate the expression data to obtain an expression map; an equalization module, used to perform histogram equalization on the expression map and the image data to obtain an image pair of the expression map-image data; and a registration module, which performs rigid registration and elastic registration on the image pair in sequence to generate contour data of the biological tissue sample, and the contour data is used to generate three-dimensional display data of the biological tissue sample.
  • an acquisition module used to acquire expression data and image data of a biological tissue sample, wherein the expression data is obtained by slicing the biological tissue sample and performing spatiotemporal group sequencing on the slices, and the image data is obtained by performing image acquisition on the biological tissue sample after slicing
  • An embodiment of the fifth aspect of the present disclosure provides an electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method described in the embodiment of the first aspect or the embodiment of the second aspect of the present disclosure.
  • the sixth aspect embodiment of the present disclosure provides a computer storage medium, wherein the computer storage medium stores computer executable instructions; after the computer executable instructions are executed by a processor, the method described in the first aspect or second aspect embodiment of the present disclosure can be implemented.
  • the biological tissue sample is sliced, and the slice is subjected to spatiotemporal group sequencing to obtain expression data; the sliced biological tissue sample is imaged to obtain image data; the expression data and/or image data are aligned to generate the outer contour and/or partition contour of the biological tissue sample; and three-dimensional display data of the biological tissue sample is generated according to the outer contour and/or partition contour.
  • the method disclosed in the present invention is aimed at the spatiotemporal group technology captured in situ, combined with optical imaging information, to assist in the rigid and elastic registration of continuous tissue slices, realize three-dimensional reconstruction, and then obtain the internal and external tissue contours to assist in the visualization information positioning, and effectively process and display large-scale high-resolution spatiotemporal group data.
  • FIG1 is a schematic diagram of a flow chart of a data collection method according to an embodiment of the present disclosure
  • FIG2 is a flow chart of a data collection method according to an embodiment of the present disclosure.
  • FIG3 is a schematic diagram of a data collection method according to an embodiment of the present disclosure.
  • FIG4 is a schematic diagram of a slicer according to an embodiment of the present disclosure.
  • FIG5 is a schematic diagram of a sliding position of a photographing device according to an embodiment of the present disclosure.
  • FIG6 is a schematic diagram of an expression data preprocessing method according to an embodiment of the present disclosure.
  • FIG7 is a schematic diagram of an image acquisition and spatiotemporal sequencing rule according to an embodiment of the present disclosure.
  • FIG8 is a schematic diagram of a photographing device taking pictures according to an embodiment of the present disclosure.
  • FIG9 is a schematic diagram of an image segmentation processing strategy according to an embodiment of the present disclosure.
  • FIG10 is a schematic diagram of an image segmentation processing strategy according to an embodiment of the present disclosure.
  • FIG11 is a schematic diagram of a rigid registration algorithm according to an embodiment of the present disclosure.
  • FIG12 is a schematic diagram of an image segmentation and registration result according to an embodiment of the present disclosure.
  • FIG13 is a schematic diagram of a tissue outer contour generation process according to an embodiment of the present disclosure.
  • FIG14 is a schematic diagram of an image rendering effect according to an embodiment of the present disclosure.
  • FIG15 is a schematic diagram of an image segmentation and registration effect according to an embodiment of the present disclosure.
  • FIG16 is a schematic diagram of an expression data registration process according to an embodiment of the present disclosure.
  • FIG17 is a schematic diagram of the effect of histogram equalization according to an embodiment of the present disclosure.
  • FIG18 is a schematic diagram of a rigid registration process according to an embodiment of the present disclosure.
  • FIG19 is a schematic diagram of a rigid registration result according to an embodiment of the present disclosure.
  • FIG20 is a schematic diagram of an elastic registration process according to an embodiment of the present disclosure.
  • FIG21 is a schematic diagram of an elastic registration result according to an embodiment of the present disclosure.
  • FIG22 is a schematic diagram of continuous registration of expression data according to an embodiment of the present disclosure.
  • FIG23 is a schematic diagram of a partition contour model generation process according to an embodiment of the present disclosure.
  • FIG24 is a schematic diagram of a basic data structure of a command file according to an embodiment of the present disclosure.
  • FIG25 is a schematic diagram of a single gene hierarchical block data structure in a command file according to an embodiment of the present disclosure
  • FIG26 is a schematic diagram of a cell type hierarchical block data in a command file according to an embodiment of the present disclosure
  • FIG27 is a schematic diagram of a mouse brain reconstruction rendering effect according to an embodiment of the present disclosure.
  • FIG28 is a schematic diagram of a list of single genes that can be displayed by mouse brain according to an embodiment of the present disclosure
  • FIG29 is a schematic diagram showing the effect of generating a tissue outer contour model according to an embodiment of the present disclosure
  • FIG30 is a schematic diagram showing the effect of generating a tissue partition outline according to an embodiment of the present disclosure.
  • FIG31 is a schematic diagram of an elastic registration result according to an embodiment of the present disclosure.
  • FIG32 is a schematic block diagram of a data acquisition device according to an embodiment of the present disclosure.
  • FIG33 is a schematic block diagram of a data processing device according to an embodiment of the present disclosure.
  • FIG34 is a schematic structural diagram of an electronic device according to an embodiment of the present disclosure.
  • the in situ capture-based method usually relies on physical contact between the chip and tissue slices and probe capture to obtain in situ spatial information and various omics information.
  • the destructive cutting of tissues and the planar capture of corresponding information limit it to two-dimensional spatiotemporal omics analysis of tissues, and it is difficult to obtain the morphological structure and accurate position information of three-dimensional space.
  • Different organs and tissues have unique three-dimensional cell distribution and gene expression maps.
  • the three-dimensional spatiotemporal omics research on tissues is extremely technically challenging and of great scientific significance.
  • the key point of three-dimensional spatiotemporal omics lies in how to break through the research of two-dimensional planes and achieve a comprehensive analysis of the spatial distribution, interactions, gene protein expression and overall morphological structure of different cells at the three-dimensional level of tissues.
  • the 3D reconstruction process based on spatiotemporal groups mainly involves rigid registration of adjacent or similar slices, and then generating partition outlines based on clustering or annotation information, and then generating the outline of the entire tissue/organism based on the slice stack.
  • the results of 3D reconstruction are usually presented in the form of charts in the article or demo images in the code library.
  • the restored three-dimensional spatial information needs to be projected onto a two-dimensional plane such as a computer screen or paper journal document for visualization.
  • a two-dimensional plane such as a computer screen or paper journal document for visualization.
  • the three-dimensional effect provided by the scattered points in the two-dimensional plane is limited. Especially when the number of scattered points is large and there is no obvious tissue contour information, it is not conducive to the human eye to recognize the three-dimensional spatial information.
  • the present disclosure proposes a data acquisition method, device and system, which is based on the spatiotemporal omics technology of in situ capture, combined with in situ optical imaging to obtain the original tissue contour, assist in the positioning of the biological information collected inside, so as to restore the three-dimensional position information before tissue cutting, reduce the impact of experimental operations on tissue morphology, achieve cell-precision position alignment for each information collection site on the chip, and effectively process and display tissue samples with larger sizes and data volumes.
  • Fig. 1 shows a schematic flow chart of a data collection method according to an embodiment of the present disclosure. As shown in Fig. 1, the method may include the following steps.
  • S101 slice the biological tissue sample, and perform spatiotemporal group sequencing on the slice to obtain expression data.
  • biological tissue samples are sliced continuously, and slices at designated locations are selected for spatiotemporal group sequencing according to actual conditions.
  • biological tissue samples are taken from biological samples, which generally refer to body fluids (such as urine, blood, saliva, bile, gastric juice, lymph and other secretions of organisms, etc.), hair, muscles and some tissue organs (such as thymus, pancreas, liver, lungs, brain, stomach, kidneys, etc.) of animals (including humans) and various microorganisms.
  • body fluids such as urine, blood, saliva, bile, gastric juice, lymph and other secretions of organisms, etc.
  • hair such as thymus, pancreas, liver, lungs, brain, stomach, kidneys, etc.
  • the device for slicing the biological tissue sample is not limited.
  • a biological tissue slicer is used to slice the biological tissue sample, which has higher accuracy and stability.
  • expression data is obtained based on the in situ capture method, and the expression data can be obtained by capturing the sequencing chip signal, wherein the performance of the chip determines the technical effects such as accuracy and resolution, preferably based on the DNA nanoball sequencing chip for in situ capture of tissues, the chip is large, the signal density is high to achieve better resolution and field of view.
  • other spatiotemporal group data based on in situ capture sequencing such as proteomics currently under development, can also be applied, and the present disclosure is not limited.
  • each time a slice is selected for empty group sequencing the corresponding position of the slice is recorded, and the corresponding position of the slice is used for subsequent three-dimensional reconstruction.
  • optical imaging may be used to collect images of biological tissue samples after slicing, or other in situ imaging techniques that do not affect spatiotemporal omics may be used, which are not limited by the present disclosure.
  • image acquisition is performed on the tissues after slicing corresponding to the spatiotemporal group sequencing of all selected slices, and images of the slice cross-sections are mainly acquired for subsequent registration.
  • other sliced biological tissue samples can also be selected for image acquisition.
  • the specific selection rules are not limited by the present disclosure. For example, image acquisition is performed on the sliced biological tissue samples every m cuts, and the slices are taken for spatiotemporal sequencing every n cuts, where m and n are positive integers and can be adjusted according to actual conditions, and n can be a multiple of m.
  • each time a slice is selected for image acquisition the corresponding position of the slice is recorded, and the corresponding position of the slice is used for subsequent three-dimensional reconstruction.
  • the expression data is preprocessed and the image data is segmented, wherein the preprocessing may include at least one of aggregation processing, quality control, logarithmic normalization, finding highly variable genes, and dimensionality reduction processing.
  • the image segmentation processing mainly refers to using an image segmentation algorithm to perform tissue segmentation, identify and delete unqualified segmentation samples, and retain qualified segmentation results as image data for subsequent registration.
  • the method may include registering the image data to generate an outer layer contour of the biological tissue sample.
  • the image data is segmented to generate a segmentation result map
  • the segmentation result map is registered to generate a registration result map
  • the outer contour of the biological tissue sample is generated based on the registration result map.
  • the first segmentation result map is selected and rotationally corrected, and the corrected image is used as the first reference map to register other segmentation maps.
  • a rigid registration method or other registration methods can be used, which are not limited by the present disclosure.
  • the method may include registering the expression data and the image data to generate a partition outline of the biological tissue sample.
  • the expression data is calculated to obtain an expression map
  • the expression map and image data are histogram-equalized to obtain an image pair of the expression map-image data
  • the image pair is rigidly aligned and elastically aligned in turn to generate a partition outline.
  • the method may also include aligning the expression data to generate an outer contour and/or partition contour of the biological tissue sample, but the spatiotemporal group slices involved in the alignment need to be adjacent or close.
  • expression data in two adjacent or close slices are registered to obtain registered expression data, and the outer contour and/or partition contour of the biological tissue sample is generated according to the registered expression data.
  • outer layer contour and/or partition contour of a biological tissue sample can be generated by any of the above methods, which can be selected according to actual research or application requirements and is not limited by the present disclosure.
  • the contour data of the biological tissue sample is generated based on the registered spatiotemporal group data. Specifically, the expression data and/or image data are registered and the registration results are subjected to bioinformatics analysis to obtain clustering and annotation results. Based on the clustering and annotation results, a three-dimensional array is constructed, and spatial scale information is added. Each cluster or partition result is segmented to generate a smooth contour surface of multiple tissues or regions of the internal structure of the sample. Among them, tools such as segment editor in 3D slicer can be used to segment each cluster or partition result, and the contour map finally generated is stored in common 3D vector model data such as vtk, obj, stl, etc.
  • a three-dimensional rendering tool is used to render the three-dimensional display data of the biological tissue sample.
  • the rendering tool used is not limited to the present disclosure.
  • the outer contour and/or partition contour data are integrated with other image rendering data, registered spatiotemporal group data, etc. to form a project command file, wherein the transmission efficiency and rendering effect are comprehensively considered, and the spatiotemporal group data are processed in layers and blocks according to the single gene expression amount, cell type, etc. when generating the project command file.
  • the generated command file is loaded into the renderer through a flask-based backend program, which can realize the intuitive display of information such as gene expression distribution and tissue contour in three-dimensional space.
  • the biological tissue sample is sliced, and the slice is subjected to spatiotemporal group sequencing to obtain expression data; the sliced biological tissue sample is imaged to obtain image data; the expression data and/or image data are aligned to generate the outer contour and/or partition contour of the biological tissue sample; and three-dimensional display data of the biological tissue sample is generated according to the outer contour and/or partition contour.
  • the present disclosure is directed to the spatiotemporal omics technology captured in situ, and is combined with optical imaging information to perform alignment to obtain the auxiliary visualization information positioning of the inner and outer tissue contours, which can effectively process and display large-scale high-resolution spatiotemporal group data, and can be applied to three-dimensional spatiotemporal omics technology to achieve simultaneous analysis of genes and images.
  • Fig. 2 shows a schematic diagram of a data collection method according to an embodiment of the present disclosure. Based on the embodiment shown in Fig. 1, as shown in Fig. 2, the method may include the following steps.
  • Fig. 3 shows a schematic diagram of a data collection method according to an embodiment of the present disclosure.
  • step S101 includes step S201 shown in FIG. 2 .
  • S201 continuously slice the biological tissue sample, and perform spatiotemporal group sequencing once for each first predetermined number of slices to obtain expression level data of the slices.
  • the first predetermined number is a positive integer, for example, one slice is taken for every 10 slices to perform spatiotemporal group sequencing.
  • a biological tissue slicer to slice the biological tissue sample, which has higher precision and stability.
  • the biological tissue sample is fixed at the freezing head position of the slicer, and the position and angle of the moving head are roughly adjusted according to the requirements of the subsequent photo position.
  • the tissue slice is trimmed to remove excess OCT embedding agent until the tissue area is exposed, and the tissue is aligned for continuous slicing.
  • a microtome is used for slicing, as shown in FIG4 , firstly, the experimenter prepares the monkey brain tissue sample, and fixes the monkey brain tissue sample at the center of the microtome frozen head. Secondly, the OCT (frozen section embedding agent) embedding block containing the monkey brain tissue to be cut is fixed on the sample tray, the section is adjusted to a suitable angle, and the trimming mode is set on the indicator screen shown in FIG4 , and the trimming thickness is adjusted.
  • OCT frozen section embedding agent
  • the right hand wheel is shaken to perform idling rough trimming of the section so that the OCT surface is cut flat; before focusing, it is necessary to ensure that the forward displacement display scale pointer of the sample head is about 15 mm, so when it is estimated that the embedding block may jump out of this range, the microtome frozen head needs to be retracted, and the retract button needs to be pressed at this time.
  • the amount of retraction needs to be based on the monkey brain tissue.
  • the remaining thickness of the tissue is determined by the situation.
  • the knife holder needs to be adjusted to make it flat with the tissue surface. Specifically, when the thickness of the embedding block is about 10mm, there is no need to adjust the knife holder.
  • the monkey brain tissue is larger than 10mm, for example, it is a 15mm thick embedding block, because the freezing head can only move forward 10mm, when the freezing head cannot move forward, it is necessary to retract the freezing head by at least 5mm.
  • the specific movement value can be read from the display screen on the right.
  • the knife holder should also move to roughly fit the tissue surface.
  • the first knife should be removed as a flatness adjustment, and a new round of experiments should be started from the second knife.
  • expression data is obtained based on the method of in situ capture, and the method of capturing sequencing chip signals can be used to obtain expression data, wherein the performance of the chip determines the technical effects such as accuracy and resolution, and preferably the in situ capture of tissues is performed based on a DNA nanoball sequencing chip, and the chip is large and the signal density is high to achieve better resolution and field of view.
  • other spatiotemporal group data based on in situ capture sequencing such as proteomics currently under development, can also be applied, and the present disclosure is not limited.
  • the expression data is also preprocessed, wherein the preprocessing includes at least one of aggregation processing, quality control, logarithmic normalization, finding highly variable genes, and dimensionality reduction processing.
  • preprocessing of data is beneficial to reduce the noise brought by expression data and the data scale effect caused by experimental conditions and operations, thereby obtaining more accurate annotation results, expression graphs that can better reflect the organizational structure, and more accurate display results in the renderer.
  • the data is also annotated, and the annotation results based on the preprocessing will be used for the generation of the partition outline and the display of the data in the renderer.
  • preprocessing and annotation methods are determined by the user of the process.
  • Existing processing processes such as binset aggregation, quality control, log normalization (also known as “sequencing depth normalization"), finding highly variable genes, dimensionality reduction, and clustering, can be used, or self-designed processes can be used.
  • the upper portion shows the original expression data of macaque brain slices, and the lower portion shows the corresponding expression data after preprocessing and annotation.
  • step S102 includes step S202 shown in FIG. 2 .
  • the second predetermined number is a positive integer.
  • sequencing and image acquisition are performed according to certain rules, wherein the first predetermined number is a multiple of the second predetermined number, and spatiotemporal group sequencing must occur together with photographing.
  • photographing is performed without sequencing, and the adjustment is made according to the actual situation.
  • the first predetermined number is 10
  • the second predetermined number is 5, and the biological tissue sample is cut into 13 slices.
  • the slice is subjected to spatiotemporal group sequencing, and the image of the biological sample after the slice is collected.
  • the image of the biological sample after the slice is collected.
  • the slice is subjected to spatiotemporal group sequencing, and the image of the biological sample after the slice is collected, until the slice is completed.
  • optical imaging may be used to collect images of biological tissue samples after slicing, or other in situ imaging techniques that do not affect spatiotemporal omics may be used, which are not limited by the present disclosure.
  • a camera is used to capture images of biological tissue samples after slicing. Specifically, as shown in FIG5 , the camera is aimed at the tissue slice, the system is calibrated, the sample section is in focus, and the current camera position is recorded through the slide software control panel. The tissue is sliced according to the plan, and the cutter flow is recorded. The first slice is taken for spatiotemporal group sequencing, and the remaining tissue sections are photographed by BlockFace. The specific photography and spatiotemporal group sequencing rules are performed as shown in FIG7 .
  • the image acquisition method adopts in-situ optical imaging, taking macaque brain data as an example, the camera device used is shown in Figure 4, first, connect the camera device and the computer. Specifically, select a suitable lens, adjust the scale value corresponding to the white dot on the lens to indicate the actual magnification of the lens at this time, connect the power cord and USB signal line of the camera and the slide, wherein the USB signal line is to be connected to the USB3.0 interface of the computer, open the camera photography software on the computer desktop, select the camera's "MV-CH250-90UC" to connect the camera normally. Secondly, by operating the slide software and camera software, control the camera device to perform image acquisition.
  • the control slide will extend into the slicer and the tissue section will appear in the display box. If the tissue surface is not in focus, set the sliding unit. You can click the single input box in the slide software box. Position "1", select "-” and “+” operations according to the actual situation (select “-” means that the slide extends outward, and select “+” means that the slide extends inward).
  • the black Mark point on the OCT (used for marking fine-tuning) can be clearly imaged in the camera display frame (the Mark point must ensure that the OCT has been cut flat by the slicer before marking).
  • the OCT is sliced continuously on the basis of the Mark point focus until the required tissue surface is slightly dropped.
  • the slicing can be formally performed according to the slicing process.
  • Steps S203-205 shown in FIG. 2 are further disclosure of step S103 in the embodiment of FIG. 1 : registering the expression data and/or the image data to generate the outer contour and/or the partition contour of the biological tissue sample.
  • BlockFace image taken by optical imaging is subjected to tissue segmentation using an image segmentation algorithm, unqualified segmentation samples are identified and deleted, and qualified BlockFace image segmentation results are retained for subsequent processing.
  • image segmentation processing of image data includes: converting the image data into HSV color space, extracting the image data of the H channel and converting it into a grayscale image; binarizing the image data through a preset image processing algorithm to determine a contour map of the binarization result; retaining the contour map that meets preset conditions and merging it with the image data.
  • the original BlockFace image is first segmented, the large area of background is removed, the image is converted to the HSV color space format, the H channel is extracted and converted into a grayscale image, the image is binarized by the Li algorithm, and the contour map of the binarization result is found.
  • the contour first-order moment is used to retain the contour greater than the threshold, and the target segmentation area is obtained by merging with the original image.
  • performing image segmentation processing on the image data may also include: acquiring a grayscale image and a histogram of the image data; performing maximum entropy threshold binarization processing on the image data; and merging the processed image with the image data.
  • the strategy shown in FIG. 9 is preferably used for image segmentation processing.
  • the strategy shown in FIG. 10 can be optionally used for image segmentation.
  • the threshold obtained by the Li algorithm may not be applicable to all cases.
  • the original BlockFace image is first segmented to remove the large area of background, and then the image after removing the large area of background is grayed, and the histogram of the grayed image is counted (the gray histogram is all the pixels in the digital image), and the frequency of occurrence is counted according to the size of the gray value.
  • the maximum entropy binarization threshold is obtained based on the maximization of the information metric between the object and the background, and it is used to segment the grayed image.
  • the target segmentation area is obtained based on the segmented mask image and the image after removing the large area of background.
  • the first BlockFace segmentation result image is taken and rotationally corrected, and the corrected image is used as the first reference image to register other segmentation images.
  • registering the segmentation result image to generate a registration result image includes: performing spatial transformation processing and linear interpolation algorithm processing on the segmentation result image; performing mean square error similarity measurement on the processed segmentation result image and the registration reference image; and determining the optimal measurement solution as the registration result image.
  • the BlockFace image segmentation result registration logic is shown in FIG11, and the maximum number of iterations of the fixed-step gradient descent optimization algorithm of the rigid registration algorithm is set to K and the learning rate is r, and the step size is t (the values of K, r, and t are adjusted according to actual conditions) to achieve registration.
  • the BlockFace image segmentation result image is subjected to spatial transformation and linear interpolation algorithms.
  • the first BlockFace segmentation result image is used as a reference image, and a mean square error similarity test is performed with the reference image. If optimization is required, a descent optimization algorithm with a fixed step size of r is performed, and spatial transformation and linear interpolation algorithms are performed again until the maximum number of iterations K is reached to obtain the optimal result output, that is, the registration result image is obtained.
  • step S203 and step S204 taking the BlockFace image of mouse heart tissue as an example, the effects of segmenting and registering the BlockFace image of mouse heart tissue are demonstrated, as shown in FIG. 12 .
  • generating the outer contour of the biological tissue sample based on the registration result image includes: stacking all the registration result images into an array file in a preset format; and processing the array file to obtain the outer contour of the biological tissue sample.
  • the Volume and Mesh are generated for the registration result map.
  • the registered BlockFace segmentation map is stacked into a .npy format file.
  • the obtained npy array file in which the spatial position occupied by the non-zero value elements is the sample tissue, and the zero value elements are the spatial background, can be stored in the medical image format NIfTI (Neuroimaging Informatics Technology Initiative).
  • the method further includes processing the optical imaging result after registration to obtain the format required for rendering.
  • the registered BlockFace segmentation images are stacked into a .npy or .tif file, and the Blockface segmentation images are processed into binary rendering files with a hierarchical block structure supported by the renderer through the Python toolkits CloudVolume and Igneous.
  • the final rendering effect is shown in Figure 14.
  • the BlockFace photograph is rotationally corrected using ImageJ and the ROI analysis area of interest is selected.
  • the BlockFace ROI image is segmented to organize the target object and the background, and the BlockFace segmentation map is registered to form a BlockFace image stack aligned in space.
  • the algorithm generates the outermost contour of the object based on the BlockFace segmentation map.
  • the method further includes: calculating the expression data to obtain an expression map; performing histogram equalization on the expression map and the image data to obtain an image pair of the expression map-image data; and performing rigid registration and elastic registration on the image pair in sequence to generate a partition contour.
  • the expression data of the spatiotemporal group is processed into an expression map reflecting the abundance of expression in different regions, and the expression map and the BlockFace segmentation map are subjected to histogram equalization, respectively.
  • the effect of the histogram equalization is shown in FIG17 .
  • the BlockFace segmentation map-expression map image pairs that are adjacent to each other in the slice position the following operations are performed: the BlockFace segmentation map is used as the reference image, and the expression map is used to perform rigid registration and elastic registration with it.
  • the transformation relationship determined in the registration (the transformation relationship of the rigid registration is in the form of a 3*3 transformation matrix, and the transformation relationship of the elastic registration is in the form of two deformation fields, which respectively represent the distribution of deformation in the x and y directions) is reused on the expression data to achieve the registration of the expression data and the BlockFace segmentation map.
  • performing rigid registration on the image pair includes: downsampling the image pair respectively; extracting feature information of the image pair; performing similarity estimation affine transformation on the image pair to obtain a rigid registration result.
  • the image pairs are downsampled and noise filtered to remove artifacts, and the edge features, shape features and regional segmentation information of the target tissue in the images of the two different modalities are extracted by the deep learning model, and then the similarity of the outermost contour of the target tissue is maximized at the output of the model to estimate the affine transformation, and finally the affine transformation is applied to the original image of the expression data to be registered, and the processed pair of expression data and BlockFace segmentation map results are sent to the elastic registration link for processing.
  • the result diagram of the rigid registration is shown in FIG19 .
  • performing elastic registration on the image pair includes: performing histogram equalization on the rigid registration result; inputting the rigid registration result into a deep learning model, outputting the intermediate result of the model, and obtaining a deformation field, wherein the deep learning model is obtained by training the expression graph-image data pairs; and performing nearest neighbor interpolation transformation on the rigid registration result according to the deformation field to obtain an elastic registration result.
  • the implementation of elastic registration is based on the derivation of a deep learning model, as shown in Figure 20.
  • the corresponding deep learning model is obtained through the following process: first, after the expression data is processed into an expression map, the expression map and the corresponding BlockFace segmentation map are respectively subjected to histogram equalization. Then, the expression map-BlockFace segmentation map is rigidly registered in the software as the input image for model training, and then elastic registration is performed as a label for model training. Based on the control point coordinates output by the software during manual elastic registration, random parameters are added to generate a randomly deformed expression map to achieve data synthesis.
  • a supervised model is trained using synthetic and real data, and the training index is stopped after the requirements are met, and the model is saved.
  • this model can be used to derive elastic deformation.
  • the expression data and BlockFace segmentation map in the actual problem are processed according to the above-mentioned processing method to obtain a histogram-equalized image.
  • the two images are input into the model, and the intermediate result of the model is output: deformation field. According to the deformation field, the corresponding displacement is obtained, and the nearest neighbor interpolation transformation is performed on the expression data to calculate the new expression data.
  • the red one is the BlockFace image
  • the green one is the expression image. After elastic registration, the deformation of the visible expression map is corrected.
  • the expression data can also be used to directly perform registration to generate the outer contour and/or partition contour of the biological tissue sample.
  • the method includes: registering the expression data in two adjacent or similar slices to obtain registered expression data; and generating the outer contour and/or partition contour of the biological tissue sample based on the registered expression data.
  • a bin set or a cell bin is calculated based on the expression data in two adjacent or similar slices, and after preprocessing such as logarithmic normalization, the optimal transmission problem is solved to obtain a transfer probability matrix (pi matrix for short) for optimal transmission; after post-processing the pi matrix, the rigidity or rigidity + elasticity transformation relationship can be calculated; finally, this transformation relationship is applied to the original expression data.
  • pi matrix for short a transfer probability matrix for short
  • an outer layer contour and/or partition contour model is generated based on a registration result obtained by registering expression data and/or image data.
  • the outer contour is generated according to the registration result obtained by performing registration based on the image data, specifically, as shown in FIG. 13 , refer to step S205 .
  • a partition outline is generated based on the registration result obtained by registering the expression data and/or image data, which is similar to step S205.
  • the registered expression data or each bin set or cell bin is subjected to conventional bioinformatics analysis to obtain clustering and annotation results.
  • a three-dimensional npy array is constructed with the value of the annotation or clustering result of each bin set or cell bin as a unit, and spatial scale information is added to obtain a NIfTI file based on the annotated clustering result.
  • the generated outer contour and/or partition contour model is processed into a binary 3D model file bound with a partition label by using a python toolkit multiresolution-mesh-creator.
  • the above-mentioned image rendering data, registered spatiotemporal group data, tissue contour data, partition contour data, etc. are integrated into a project management file, and the file content structure is determined according to research or application requirements.
  • the aligned spatiotemporal group expression data is used to generate a cgef format file through geftools, and then the binary image data, binary partition contour model data, tissue outer contour model data, etc. generated in the above steps are used to generate a command file through gef3d.
  • the file content structure is shown in Figure 24, which mainly shows the first three levels of the command file.
  • the spatiotemporal group data can be processed in layers and blocks according to single gene expression, cell type, etc. when generating a project management file.
  • layered block segmentation refers to sampling a large number of data point clouds according to spatial positions in a pyramid-like structure, and ensuring that the data points in all blocks of all layers are combined into all data. This process is automatically performed when gef3d generates the command file, and the threshold of the number of bins/cellbins per block needs to be specified.
  • the hierarchical block method may include: first traversing all cgef to obtain the overall space size, setting a maximum number of layers, calculating the division scale according to the length of each data point in the three-dimensional space, and pre-dividing the number of blocks in each layer from top to bottom according to the 8-point method, and each block is approximately a cube. By default, there is only one block in the first layer. After the traversal of the first step, the data of each gene in each slice can be obtained, that is, the attribute value of the data point in the entire space, including coordinates and expression abundance.
  • the data of a single gene is hierarchically divided into blocks, and according to the result calculated in the first step, from the top layer to the bottom layer, each small block of each layer is filled in turn from the overall data of the gene until the small block is filled (controlled by the threshold). The remaining part continues to fill the next layer. If a block of data in the next layer is too small (threshold control), the data of this block will be added to the parent block of the previous layer. Distribute downward in sequence until the data of the gene is distributed.
  • the hierarchical block of the annotation results of cell types such as celltype is similar to that of single genes.
  • blk represents block, which is the abbreviation of block. This directory is used to store the block data.
  • offset represents offset, which stores the starting index offset of the block data
  • count represents number, which stores the number of points of the block data.
  • Cell represents cell, which is used to store all bin/cellbin data related information at this level.
  • x, y, and z respectively store the x, y, and z coordinate values of the bin/cellbin spatial coordinates
  • umicnt stores the bin/cellbin expression
  • cellid represents the cell number, which stores the bin/cellbin number.
  • ctype_list represents the cell type list, which stores the list of cell type numbers, names, and colors
  • maxlevelnum represents the maximum number of layers, which is the maximum number of layers in all cell type data
  • blkcnt represents the number of blocks, which stores the number of blocks in each axis at different levels
  • maxumi represents the maximum expression value, which stores the maximum expression value.
  • the command file generated by the above steps is loaded into the renderer through the flask-based backend program, which can realize the intuitive display of information such as gene expression distribution and tissue outline in three-dimensional space.
  • the effect is shown in Figure 27, where the overall expression data and tissue outline model are loaded respectively.
  • the in situ captured spatiotemporal omics technology combined with optical imaging information, it assists in the rigid and elastic alignment of continuous tissue sections, realizes three-dimensional reconstruction, and then obtains the internal and external tissue contours to assist in the visualization information positioning, and effectively processes and displays large-size, high-resolution spatiotemporal group data.
  • step S204 and step S205 the collected sample image data is subjected to image segmentation processing method and image registration.
  • image segmentation processing method and image registration Taking the BlockFace image of mouse heart tissue as an example, the effect of segmenting and registering the BlockFace image of mouse heart tissue is demonstrated, as shown in FIG12 .
  • FIG. 29 The left side shows the mouse brain tissue partition outline model, and the right side shows the mouse brain tissue partition outline model and the corresponding cell type bin.
  • the processing and display of tissue samples with larger size and data volume are shown in Figure 30.
  • the selected display example is more than 100 spatiotemporal chip data of mouse brain.
  • the chip size is 1cm x 1cm, and the displayed data has about 900K coordinate points, including about 22K genes.
  • the method disclosed in the present invention can better restore the tissue contour and related expression data before tissue sample slicing, and be applied to three-dimensional reconstruction of spatiotemporal omics, which will be conducive to promoting the research and application of three-dimensional spatiotemporal omics technology.
  • Fig. 31 is a flow chart of a data processing method according to an embodiment of the present disclosure. As shown in Fig. 31, the method may include the following steps.
  • the expression data are obtained by slicing biological tissue samples and performing spatiotemporal group sequencing on the slices, and the image data are obtained by capturing images of the biological tissue samples after slicing.
  • biological tissue samples are sliced continuously, and slices at designated positions are selected for spatiotemporal group measurement according to actual conditions to obtain expression data.
  • optical imaging can be used to capture images of biological tissue samples after slicing to obtain image data
  • other in situ imaging techniques that do not affect spatiotemporal omics can also be used, which are not limited by the present disclosure.
  • biological tissue samples are taken from biological samples, which generally refer to body fluids (such as urine, blood, saliva, bile, gastric juice, lymph and other secretions of organisms, etc.), hair, muscles and some tissue organs (such as thymus, pancreas, liver, lungs, brain, stomach, kidneys, etc.) of animals (including humans) and various microorganisms.
  • body fluids such as urine, blood, saliva, bile, gastric juice, lymph and other secretions of organisms, etc.
  • hair such as thymus, pancreas, liver, lungs, brain, stomach, kidneys, etc.
  • the device for slicing the biological tissue sample is not limited.
  • a biological tissue slicer is used to slice the biological tissue sample, which has higher accuracy and stability.
  • expression data is obtained based on the in situ capture method, and the expression data can be obtained by capturing the sequencing chip signal, wherein the performance of the chip determines the technical effects such as accuracy and resolution, preferably based on the DNA nanoball sequencing chip for in situ capture of tissues, the chip is large, the signal density is high to achieve better resolution and field of view.
  • other spatiotemporal group data based on in situ capture sequencing such as proteomics currently under development, can also be applied, and the present disclosure is not limited.
  • image acquisition is performed on the tissues after slicing corresponding to all selected slices for spatiotemporal group sequencing, and images of the slice cross sections are mainly acquired for subsequent alignment.
  • other biological tissue samples after slicing can also be selected for image acquisition.
  • the specific selection rules are not limited in the present disclosure.
  • each time a slice is selected for spatiotemporal group sequencing or image acquisition the corresponding position of the slice is recorded, and the corresponding position of the slice is used for subsequent three-dimensional reconstruction.
  • the expression data may be preprocessed, wherein the preprocessing includes at least one of aggregation processing, quality control, logarithmic normalization, finding highly variable genes, and dimensionality reduction processing.
  • Preprocessing the data is conducive to reducing the noise brought by the expression data, reducing the data scale effect caused by experimental conditions and operations, and thereby obtaining more accurate annotation results, expression graphs that can better reflect the organizational structure, and more accurate display results in the renderer.
  • the image data after acquiring the image data, the image data may also be segmented. Specifically, the image data is converted into an HSV color space, the image data of the H channel is extracted and converted into a grayscale image, and the image data is binarized through a preset image processing algorithm to determine a contour map of the binarization result, and the contour map that meets the preset conditions is retained and merged with the image data.
  • step S201 and step 202 Please refer to the specific implementation of step S201 and step 202, which will not be described in detail in this disclosure.
  • the correlation coefficient of the expression data of the spatiotemporal group is calculated to obtain an expression map that can reflect the expression abundance of different regions.
  • the specific calculation, statistics and drawing methods are not limited in the present disclosure.
  • histogram equalization is a simple and effective image enhancement technology. It changes the grayscale of each pixel in the image by changing the histogram of the image. It is mainly used to enhance the contrast of images with a small dynamic range.
  • the grayscale distribution of the original image may be concentrated in a narrow range, resulting in unclear images. For example, the grayscale of an overexposed image is concentrated in a high brightness range, while underexposure will cause the grayscale of the image to be concentrated in a low brightness range.
  • histogram equalization the histogram of the original image can be transformed into a uniformly distributed (balanced) form, thereby increasing the dynamic range of the grayscale value difference between pixels, thereby achieving the effect of enhancing the overall contrast of the image.
  • a BlockFace image is collected as image data, and the image is segmented to obtain a BlockFace segmentation map, and the BlockFace segmentation map-expression image pairs with adjacent slice positions are taken as image pairs of expression map-image data.
  • S304 performing rigid registration and elastic registration on the image pair in sequence to generate contour data of the biological tissue sample, and the contour data is used to generate three-dimensional display data of the biological tissue sample.
  • the BlockFace segmentation map is used as a reference image, and the expression map is used to perform rigid registration and elastic registration with it.
  • the transformation relationship determined in the registration (the transformation relationship of the rigid registration is in the form of a 3*3 transformation matrix, and the transformation relationship of the elastic registration is in the form of two deformation fields, which respectively represent the distribution of deformation in the x and y directions) is reused on the expression data to achieve registration of the expression data with the BlockFace segmentation map.
  • rigid registration of the image pair includes: downsampling the image pair respectively; extracting feature information of the image pair; performing similarity estimation affine transformation on the image pair to obtain a rigid registration result.
  • the image pairs are downsampled and noise filtered to remove artifacts, and the edge features, shape features and regional segmentation information of the target tissue in the images of the two different modalities are extracted by the deep learning model, and then the similarity of the outermost contour of the target tissue is maximized at the output of the model to estimate the affine transformation, and finally the affine transformation is applied to the original image of the expression data to be registered, and the processed pair of expression data and BlockFace segmentation map results are sent to the elastic registration link for processing.
  • the result diagram of the rigid registration is shown in FIG19 .
  • elastic registration of image pairs includes: performing histogram equalization on the rigid registration result; inputting the rigid registration result into a deep learning model, outputting an intermediate result of the deep learning model, and obtaining a deformation field, wherein the deep learning model is obtained by training expression graph-image data pairs; and performing nearest neighbor interpolation transformation on the rigid registration result according to the deformation field to obtain an elastic registration result.
  • the implementation of elastic registration is based on the derivation of a deep learning model, as shown in Figure 20.
  • the corresponding deep learning model is obtained through the following process: first, after the expression data is processed into an expression map, the expression map and the corresponding BlockFace segmentation map are respectively subjected to histogram equalization. Then, the expression map-BlockFace segmentation map is rigidly registered in the software as the input image for model training, and then elastic registration is performed as a label for model training. Based on the control point coordinates output by the software during manual elastic registration, random parameters are added to generate a randomly deformed expression map to achieve data synthesis.
  • a supervised model is trained using synthetic and real data, and the training index is stopped after the requirements are met, and the model is saved.
  • this model can be used to derive elastic deformation.
  • the expression data and BlockFace segmentation map in the actual problem are processed according to the above-mentioned processing method to obtain a histogram-equalized image.
  • the two images are input into the model, and the intermediate result of the model is output: deformation field. According to the deformation field, the corresponding displacement is obtained, the nearest neighbor interpolation is performed on the expression data, and the new expression data is calculated.
  • the red one is the BlockFace image
  • the green one is the expression image. After elastic registration, the deformation of the visible expression map is corrected.
  • the contour data of biological tissue samples are generated based on the aligned spatiotemporal group data. Specifically, based on the aligned expression spatiotemporal data, each bin set or cell bin is clustered and annotated after conventional bioinformatics analysis. A three-dimensional npy array is constructed with the value of the annotation or clustering result of each bin set or cell bin as the unit, and spatial scale information is added to obtain a NIfTI file based on the annotated clustering result.
  • tools such as the segment editor in the 3D slicer are used to segment each cluster or partition result, and a smooth contour surface of multiple tissues or regions of the internal structure of the sample is generated, which is stored in common 3D vector model data such as vtk, obj, and stl.
  • the contour data is used to generate three-dimensional display data of biological tissue samples.
  • the contour data is integrated with other image rendering data, space-time group data after registration, etc. to form a project command file, in which the transmission efficiency and rendering effect are comprehensively considered.
  • the space-time group data is layered and blocked according to the single gene expression amount, cell type, etc.
  • the generated command file is loaded into the renderer through a flask-based backend program, which can realize the intuitive display of information such as gene expression distribution and tissue contour in three-dimensional space.
  • expression data and image data of a biological tissue sample are obtained, wherein the expression data is obtained by slicing the biological tissue sample and performing spatiotemporal group sequencing on the slices, and the image data is obtained by performing image acquisition on the biological tissue sample after slicing; the expression data is calculated to obtain an expression graph; the expression graph and the image data are histogram-equalized to obtain an image pair of the expression graph-image data; the image pair is sequentially rigidly registered and elastically registered to generate contour data of the biological tissue sample, and the contour data is used to generate three-dimensional display data of the biological tissue sample, providing an expression graph of spatiotemporal sequencing.
  • the registration method of volume data and tissue section image data can be applied to obtain auxiliary visualization information positioning of internal and external tissue contours, effectively process and display large-size and high-resolution spatiotemporal group data, and promote the research and application of three-dimensional spatiotemporal group technology.
  • the network device and the terminal device may include a hardware structure and a software module, and implement the above functions in the form of a hardware structure, a software module, or a hardware structure plus a software module.
  • One of the above functions may be executed in the form of a hardware structure, a software module, or a hardware structure plus a software module.
  • the present disclosure also provides a data acquisition device. Since the data acquisition device provided in the embodiment of the present disclosure corresponds to the data acquisition methods provided in the above-mentioned embodiments, the implementation method of the data acquisition method is also applicable to the data acquisition device provided in this embodiment and will not be described in detail in this embodiment.
  • Figure 32 is a structural schematic diagram of a data acquisition device 400 provided in an embodiment of the present disclosure.
  • the device 400 may include: a sequencing module 410, used to slice a biological tissue sample and perform spatiotemporal group sequencing on the slices to obtain expression data; an acquisition module 420, used to perform image acquisition on the sliced biological tissue sample to obtain image data; an alignment module 430, used to align the expression data and/or image data to generate an outer contour and/or partition contour of the biological tissue sample; a generation module 440, used to generate three-dimensional display data of the biological tissue sample based on the outer contour and/or partition contour.
  • a sequencing module 410 used to slice a biological tissue sample and perform spatiotemporal group sequencing on the slices to obtain expression data
  • an acquisition module 420 used to perform image acquisition on the sliced biological tissue sample to obtain image data
  • an alignment module 430 used to align the expression data and/or image data to generate an outer contour and/or partition contour of the biological tissue sample
  • a generation module 440 used to generate three-dimensional display data
  • the biological tissue sample is sliced, and the slices are subjected to spatiotemporal group sequencing to obtain expression data; the sliced biological tissue sample is imaged to obtain image data; the expression data and/or image data are aligned to generate the outer contour and/or partition contour of the biological tissue sample; based on the outer contour and/or partition contour, three-dimensional display data of the biological tissue sample is generated, and for the spatiotemporal group technology captured in situ, combined with optical imaging information, the inner and outer tissue contours are aligned to assist in the positioning of visualization information, so that large-scale and high-resolution spatiotemporal group data can be effectively processed and displayed.
  • the transmission configuration information further includes a recommended OAM mode quantity and/or a recommended OAM mode set, and the recommended OAM mode quantity and/or the recommended OAM mode set are obtained by the second device based on the recommended PMI.
  • the sequencing module 410 is specifically used to: continuously slice the biological tissue sample, and perform spatiotemporal group sequencing once for each first predetermined number of slices to obtain expression data of the slices, wherein the first predetermined number is a positive integer.
  • a preprocessing module is further included for preprocessing the expression data, wherein the preprocessing includes at least one of aggregation processing, quality control, logarithmic normalization, finding highly variable genes, and dimensionality reduction processing.
  • the acquisition module 420 is specifically configured to: perform an image acquisition on the sliced biological tissue sample once every second predetermined number of slices to obtain image data, wherein the second predetermined number is a positive integer.
  • the registration module 430 is specifically used to: perform image segmentation processing on the image data to generate a segmentation result map; perform registration on the segmentation result map to generate a registration result map; and generate an outer contour of the biological tissue sample based on the registration result map.
  • image segmentation processing of image data includes: converting the image data into HSV color space, extracting the image data of the H channel and converting it into a grayscale image; binarizing the image data through a preset image processing algorithm to determine a contour map of the binarization result; retaining the contour map that meets preset conditions and merging it with the image data.
  • the method further includes: acquiring a grayscale image and a histogram of the image data; performing maximum entropy threshold binarization processing on the image data; and merging the processed image with the image data.
  • registering the segmentation result image to generate a registration result image includes: performing spatial transformation processing and linear interpolation algorithm processing on the segmentation result image; performing mean square error similarity measurement on the processed segmentation result image and the registration reference image; and determining the optimal measurement solution as the registration result image.
  • generating the outer contour of the biological tissue sample based on the registration result images includes: stacking all the registration result images into an array file in a preset format; and processing the array file to obtain the outer contour of the biological tissue sample.
  • the registration module 430 is also used to: calculate the expression data to obtain an expression map; perform histogram equalization on the expression map and image data to obtain an image pair of expression map-image data; and perform rigid registration and elastic registration on the image pair in sequence to generate a partition outline.
  • performing rigid registration on the image pair includes: downsampling the image pair respectively; extracting feature information of the image pair; performing similarity estimation affine transformation on the image pair to obtain a rigid registration result.
  • elastic registration of image pairs includes: performing histogram equalization on the rigid registration results; inputting the rigid registration results into a deep learning model, outputting an intermediate result of the deep learning model, and obtaining a deformation field, wherein the deep learning model is obtained by training expression graph-image data pairs; and performing nearest neighbor interpolation transformation on the rigid registration results according to the deformation field to obtain an elastic registration result.
  • the registration module 430 is further used to: register the expression data in two adjacent or close slices to obtain registered expression data; and generate the outer layer contour and/or partition contour of the biological tissue sample according to the registered expression data.
  • the generation module 440 is specifically used to: perform bioinformatics analysis on the registration results obtained by registering the expression data and/or image data to obtain clustering and annotation results; construct a three-dimensional array based on the clustering and annotation results, and add spatial scale information; use a three-dimensional rendering tool to render and obtain three-dimensional display data of the biological tissue sample.
  • the biological tissue sample is sliced, and the slices are subjected to spatiotemporal group sequencing to obtain expression data; the sliced biological tissue sample is imaged to obtain image data; the expression data and/or image data are aligned to generate the outer contour and/or partition contour of the biological tissue sample; based on the outer contour and/or partition contour, three-dimensional display data of the biological tissue sample is generated, and for the spatiotemporal omics technology captured in situ, combined with optical imaging information, the rigid and elastic alignment of continuous tissue slices is assisted to achieve three-dimensional reconstruction, and then the internal and external tissue contours are obtained to assist in the visualization information positioning, and large-size high-resolution spatiotemporal group data are effectively processed and displayed, and the application of three-dimensional spatiotemporal omics technology can realize the simultaneous analysis of genes and images.
  • the present disclosure also provides a data processing device. Since the data processing device provided in the embodiment of the present disclosure corresponds to the data processing methods provided in the above-mentioned embodiments, the implementation method of the data processing method is also applicable to the data processing device provided in this embodiment and will not be described in detail in this embodiment.
  • FIG33 is a schematic diagram of the structure of a data processing device 500 provided in an embodiment of the present disclosure.
  • the device 500 may include: an acquisition module 510, used to acquire expression data and image data of a biological tissue sample, wherein the expression data is obtained by slicing the biological tissue sample and performing spatiotemporal group sequencing on the slice, and the image data is obtained by performing image acquisition on the sliced biological tissue sample; a calculation module 520, used to calculate the expression data to obtain an expression map; an equalization module 530, used to perform histogram equalization on the expression map and the image data to obtain an image pair of the expression map-image data; and a registration module 540, which performs rigid registration and elastic registration on the image pair in sequence to generate contour data of the biological tissue sample, and the contour data is used to generate three-dimensional display data of the biological tissue sample.
  • an acquisition module 510 used to acquire expression data and image data of a biological tissue sample, wherein the expression data is obtained by slicing the biological tissue sample and performing spatiotemporal group sequencing on
  • expression data and image data of a biological tissue sample are obtained, wherein the expression data is obtained by slicing the biological tissue sample and performing spatiotemporal group sequencing on the slice, and the image data is obtained by performing image acquisition on the sliced biological tissue sample; the expression data is calculated to obtain an expression graph; the expression graph and the image data are histogram-equalized to obtain an image pair of the expression graph-image data; the image pair is sequentially rigidly registered and elastically registered to generate contour data of the biological tissue sample, and the contour data is used to generate three-dimensional display data of the biological tissue sample, so as to realize the spatiotemporal sequencing of the expression data and
  • the registration of tissue section image data can obtain auxiliary visualization information positioning of internal and external tissue contours, effectively process and display large-scale high-resolution spatiotemporal group data, and promote the research and application of three-dimensional spatiotemporal omics technology.
  • the registration module 540 is used to perform rigid registration on the image pair, including: downsampling the image pair respectively; extracting feature information of the image pair; performing similarity estimation affine transformation on the image pair to obtain a rigid registration result.
  • the registration module 540 is used to perform elastic registration on image pairs, including: performing histogram equalization on the rigid registration results; inputting the rigid registration results into a deep learning model, outputting the intermediate results of the deep learning model, and obtaining a deformation field, wherein the deep learning model is obtained by training expression graph-image data pairs; according to the deformation field, performing nearest neighbor interpolation transformation on the rigid registration results to obtain an elastic registration result.
  • expression data and image data of a biological tissue sample are obtained, wherein the expression data is obtained by slicing the biological tissue sample and performing spatiotemporal group sequencing on the slices, and the image data is obtained by performing image acquisition on the biological tissue sample after slicing; the expression data is calculated to obtain an expression graph; the expression graph and the image data are histogram-equalized to obtain an image pair of the expression graph-image data; the image pair is rigidly aligned and elastically aligned in sequence to generate contour data of the biological tissue sample, and the contour data is used to generate three-dimensional display data of the biological tissue sample.
  • a registration method for spatiotemporal sequenced expression data and tissue section image data is provided. The present method can be used to obtain auxiliary visualization information positioning of internal and external tissue contours, effectively process and display large-size and high-resolution spatiotemporal group data, and promote the research and application of three-dimensional spatiotemporal group technology.
  • the electronic device may include a hardware structure and a software module, and implement the functions in the form of a hardware structure, a software module, or a hardware structure plus a software module.
  • a function of the functions may be executed in the form of a hardware structure, a software module, or a hardware structure plus a software module.
  • the electronic device 600 may be a mobile phone, a computer, a messaging device, a game console, a tablet device, a medical device, a fitness device, a personal digital assistant, etc.
  • the electronic device 600 may include one or more of the following components: a processing component 602 , a memory 604 , a power component 606 , a multimedia component 606 , an audio component 610 , an input/output (I/O) interface 612 , a sensor component 614 , and a communication component 616 .
  • the processing component 602 generally controls the overall operation of the electronic device 600, such as operations associated with display, phone calls, data communications, camera operations, and recording operations.
  • the processing component 602 may include one or more processors 620 to execute instructions to complete all or part of the steps of the above-mentioned method.
  • the processing component 602 may include one or more modules to facilitate the interaction between the processing component 602 and other components.
  • the processing component 602 may include a multimedia module to facilitate the interaction between the multimedia component 608 and the processing component 602.
  • the memory 604 is configured to store various types of data to support operations on the electronic device 600. Examples of such data include instructions for any application or method operating on the electronic device 600, contact data, phone book data, messages, pictures, videos, etc.
  • the memory 604 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), etc. (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), Magnetic Storage, Flash Memory, Magnetic or Optical Disk.
  • SRAM static random access memory
  • EEPROM electrically erasable programmable read-only memory
  • EPROM erasable programmable read-only memory
  • PROM Programmable Read-Only Memory
  • ROM Read-Only Memory
  • Magnetic Storage Flash Memory
  • Flash Memory Magnetic or Optical Disk.
  • the power supply component 606 provides power to the various components of the electronic device 600.
  • the power supply component 606 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to the electronic device 600.
  • the multimedia component 606 includes a screen that provides an output interface between the electronic device 600 and the user.
  • the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from the user.
  • the touch panel includes one or more touch sensors to sense touch, slide, and gestures on the touch panel. The touch sensor may not only sense the boundaries of the touch or slide action, but also detect the duration and pressure associated with the touch or slide operation.
  • the multimedia component 608 includes a front camera and/or a rear camera. When the electronic device 600 is in an operating mode, such as a shooting mode or a video mode, the front camera and/or the rear camera may receive external multimedia data. Each front camera and rear camera may be a fixed optical lens system or have a focal length and optical zoom capability.
  • the audio component 610 is configured to output and/or input audio signals.
  • the audio component 610 includes a microphone (MIC), and when the electronic device 600 is in an operating mode, such as a call mode, a recording mode, and a voice recognition mode, the microphone is configured to receive an external audio signal.
  • the received audio signal can be further stored in the memory 604 or sent via the communication component 616.
  • the audio component 610 also includes a speaker for outputting audio signals.
  • I/O interface 612 provides an interface between processing component 602 and peripheral interface modules, such as keyboards, click wheels, buttons, etc. These buttons may include but are not limited to: a home button, a volume button, a start button, and a lock button.
  • the sensor assembly 614 includes one or more sensors for providing various aspects of status assessment for the electronic device 600.
  • the sensor assembly 614 can detect the open/closed state of the electronic device 600, the relative positioning of components, such as the display and keypad of the electronic device 600, and the sensor assembly 614 can also detect the position change of the electronic device 600 or a component of the electronic device 600, the presence or absence of user contact with the electronic device 600, the orientation or acceleration/deceleration of the electronic device 600, and the temperature change of the electronic device 600.
  • the sensor assembly 614 may include a proximity sensor configured to detect the presence of nearby objects without any physical contact.
  • the sensor assembly 614 may also include an optical sensor, such as a CMOS or CCD image sensor, for use in imaging applications.
  • the sensor assembly 614 may also include an accelerometer, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.
  • the communication component 616 is configured to facilitate wired or wireless communication between the electronic device 600 and other devices.
  • the electronic device 600 can access a wireless network based on a communication standard, such as WiFi, 2G or 3G, 4G LTE, 5G NR (New Radio), or a combination thereof.
  • the communication component 616 receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel.
  • the communication component 616 also includes a near field communication (NFC) module to facilitate short-range communication.
  • the NFC module can be implemented based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology, and other technologies.
  • RFID radio frequency identification
  • IrDA infrared data association
  • UWB ultra-wideband
  • Bluetooth Bluetooth
  • the electronic device 600 may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above methods.
  • ASICs application-specific integrated circuits
  • DSPs digital signal processors
  • DSPDs digital signal processing devices
  • PLDs programmable logic devices
  • FPGAs field programmable gate arrays
  • controllers microcontrollers, microprocessors, or other electronic components to perform the above methods.
  • a non-transitory computer-readable storage medium including instructions is also provided, such as a memory 604 including instructions, and the above instructions can be executed by the processor 620 of the electronic device 600 to complete the above method.
  • the non-transitory computer-readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, etc.
  • the embodiments of the present disclosure further provide a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to enable a computer to execute the data collection or data processing method described in the above embodiments of the present disclosure.
  • the embodiments of the present disclosure further provide a computer program product, including a computer program, which executes the data collection or data processing method described in the above embodiments of the present disclosure when a processor is used.
  • An embodiment of the present disclosure also proposes a chip, which includes one or more interface circuits and one or more processors; the interface circuit is used to receive signals from a memory of an electronic device and send signals to the processor, the signals include computer instructions stored in the memory, and when the processor executes the computer instructions, the electronic device executes the data collection or data processing method described in the above embodiments of the present disclosure.
  • Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code that includes one or more executable instructions for implementing the steps of a specific logical function or process, and the scope of the preferred embodiments of the present invention includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present invention belong.
  • the logic and/or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processing module, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in combination with these instruction execution systems, devices or apparatuses.
  • "computer-readable medium” can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses.
  • computer-readable media include the following: an electrical connection with one or more wires (control method), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disk read-only memory (CDROM).
  • the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting or otherwise processing in a suitable manner if necessary, and then stored in a computer memory.
  • each functional unit in each embodiment of the present invention may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module.
  • the above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.
  • the above-mentioned storage medium may be a read-only memory, a disk or an optical disk, etc.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Image Analysis (AREA)
  • Image Processing (AREA)

Abstract

本公开实施例提供一种数据采集方法、装置及系统，涉及数据处理技术领域。对生物组织样本进行切片，并对切片进行时空组测序，以得到表达量数据；对切片后的生物组织样本进行图像采集，以得到图像数据；对表达量数据和/或图像数据进行配准，以生成生物组织样本的外层轮廓和/或分区轮廓；根据外层轮廓和/或分区轮廓，生成生物组织样本的三维展示数据。本公开提出一整套的数据采集装置与方法，以及时空测序的表达量数据与组织切面图像数据的配准方法，针对原位捕获的时空组学技术，结合光学成像信息，辅助连续组织切片的刚性、弹性配准，实现三维重构，进而得到内外组织轮廓辅助可视化信息定位，对大尺寸高分辨率时空组数据进行有效处理与展示。

Description

数据采集方法、装置及系统 技术领域
本公开涉及数据处理技术领域，特别涉及一种数据采集方法、装置及系统。
背景技术
时空组学技术自发展以来，已在神经科学、病理学、发育生物学等方向得到初步应用。二维平面的时空组学分析难以得到三维空间的形态结构及准确的位置信息，如何突破二维平面的研究实现三维时空组学研究具有极大的技术挑战性和重要的科学意义。目前主要研究时空组学三维重构方法，以实现对组织三维水平上不同细胞的空间分布、相互作用、基因蛋白质等表达情况以及整体形态结构的全面解析。当前的时空组学三维重构仅依据时空组捕获的表达量丰度信息，将测量于平面上的表达量数据进行刚性配准，存在数据重构回原位表达水平准确性低、可视化效果有限、对于大数据量的组织样本数据处理和后期展示能力不足等问题。
发明内容
本公开提出了一种数据采集方法、装置及系统，针对原位捕获的时空组学技术，结合光学成像信息得到内外组织轮廓辅助可视化信息定位，每个信息采集位点实现细胞精度的位置配准，而且可以对大尺寸高分辨率时空组数据进行有效处理与展示。
本公开的第一方面实施例提供了一种数据采集方法，该方法包括：对生物组织样本进行切片，并对切片进行时空组测序，以得到表达量数据；对切片后的生物组织样本进行图像采集，以得到图像数据；对表达量数据和/或图像数据进行配准，以生成生物组织样本的外层轮廓和/或分区轮廓；根据外层轮廓和/或分区轮廓，生成生物组织样本的三维展示数据。
在本公开的一些实施例中，对生物组织样本进行切片，并对切片进行时空组测序，以得到表达量数据包括：对生物组织样本进行连续切片，每切第一预定数量的切片进行一次时空组测序，以得到切片的表达量数据，其中，第一预定数量为正整数。
在本公开的一些实施例中，该方法还包括：对表达量数据进行预处理，其中，预处理包括聚合处理、质量控制、对数归一化、找高变基因、降维处理中的至少一项。
在本公开的一些实施例中，对切片后的生物组织样本进行图像采集，以得到图像数据包括：每切第二预定数量的切片，对切片后的生物组织样本进行一次图像采集，以得到图像数据，其中，第二预定数量为正整数。
在本公开的一些实施例中，对表达量数据和/或图像数据进行配准，以生成生物组织样本的外层轮廓和/或分区轮廓包括：对图像数据进行图像分割处理，以生成分割结果图；对分割结果图进行配准，以生成配准结果图；基于配准结果图，生成生物组织样本的外层轮廓。
在本公开的一些实施例中，对图像数据进行图像分割处理包括：将图像数据转换为HSV颜色空间，提取H通道的图像数据并转换为灰度图；通过预设图像处理算法，对图像数据进行二值化，以确定二值化结果轮廓图；保留符合预设条件的轮廓图，与图像数据进行合并。
在本公开的一些实施例中，该方法还包括：获取图像数据的灰度化图像和直方图；对图像数据进行最大熵阈值二值化处理；将处理后的图像与图像数据进行合并。
在本公开的一些实施例中，对分割结果图进行配准，以生成配准结果图包括：对分割结果图进行空间变换处理和线性插值算法处理；对经处理的分割结果图以及配准参考图进行均方差相似度测量；将测量最优解确定为配准结果图。
在本公开的一些实施例中，基于配准结果图，生成生物组织样本的外层轮廓包括：将所有配准结果图堆叠为预设格式的数组文件；对数组文件进行处理，以得到生物组织样本的外层轮廓。
在本公开的一些实施例中，对表达量数据和/或图像数据进行配准，以生成生物组织样本的外层轮廓和/或分区轮廓包括：对表达量数据进行计算，得到表达量图；将表达量图和图像数据进行直方图均衡化，得到表达量图-图像数据的图像对；对图像对依次进行刚性配准和弹性配准，以生成分区轮廓。
在本公开的一些实施例中，对图像对进行刚性配准包括：对图像对分别进行下采样；提取图像对的特征信息；对图像对进行相似性估计仿射变换，得到刚性配准结果。
在本公开的一些实施例中，对图像对进行弹性配准包括：对刚性配准结果进行直方图均衡化；将刚性配准结果输入深度学习模型，输出深度学习模型的中间结果，得到形变场，其中，深度学习模型是通过表达量图-图像数据对进行训练得到的；根据形变场，对刚性配准结果进行最近邻插值变换，得到弹性配准结果。
在本公开的一些实施例中，对表达量数据和/或图像数据进行配准，以生成生物组织样本的外层轮廓和/或分区轮廓包括：对两个相邻或相近切片中的表达量数据进行配准，以得到配准表达量数据；根据配准表达量数据，生成生物组织样本的外层轮廓和/或分区轮廓。
在本公开的一些实施例中，根据外层轮廓和/或分区轮廓，生成生物组织样本的三维展示数据包括：对表达量数据和/或图像数据进行配准得到的配准结果进行生物信息学分析，得到聚类和注释结果；基于聚类和注释结果，构造三维数组，并增加空间尺度信息；利用三维渲染工具，渲染得到生物组织样本的三维展示数据。
本公开实施例的第二方面提供一种数据处理方法，该方法包括：获取生物组织样本的表达量数据和图像数据，其中表达量数据是对生物组织样本进行切片并对切片进行时空组测序得到的，图像数据是对切片后的生物组织样本进行图像采集得到的；对表达量数据进行计算，得到表达量图；将表达量图和图像数据进行直方图均衡化，得到表达量图-图像数据的图像对；对图像对依次进行刚性配准和弹性配准，以生成生物组织样本的轮廓数据，轮廓数据用于生成生物组织样本的三维展示数据。
在本公开的一些实施例中，对图像对进行刚性配准包括：对图像对分别进行下采样；提取图像对的特征信息；对图像对进行相似性估计仿射变换，得到刚性配准结果。
在本公开的一些实施例中，对图像对进行弹性配准包括：对刚性配准结果进行直方图均衡化；将刚性配准结果输入深度学习模型，输出深度学习模型的中间结果，得到形变场，其中，深度学习模型是通过表达量图-图像数据对进行训练得到的；根据形变场，对刚性配准结果进行最近邻插值变换，得到弹性配准结果。
本公开第三方面实施例提供一种数据采集装置，该装置包括：测序模块，用于对生物组织样本进行切片，并对切片进行时空组测序，以得到表达量数据；采集模块，用于对切片后的生物组织样本进行图像采集，以得到图像数据；配准模块，用于对表达量数据和/或图像数据进行配准，以生成生物组 织样本的外层轮廓和/或分区轮廓；生成模块，用于根据外层轮廓和/或分区轮廓，生成生物组织样本的三维展示数据。
本公开第四方面实施例提供一种数据处理装置，该装置包括：获取模块，用于获取生物组织样本的表达量数据和图像数据，其中表达量数据是对生物组织样本进行切片并对切片进行时空组测序得到的，图像数据是对切片后的生物组织样本进行图像采集得到的；计算模块，用于对表达量数据进行计算，得到表达量图；均衡化模块，用于将表达量图和图像数据进行直方图均衡化，得到表达量图-图像数据的图像对；配准模块，对图像对依次进行刚性配准和弹性配准，以生成生物组织样本的轮廓数据，轮廓数据用于生成生物组织样本的三维展示数据。
本公开第五方面实施例提供一种电子设备，包括：至少一个处理器；以及与至少一个处理器通信连接的存储器；其中，存储器存储有可被至少一个处理器执行的指令，指令被至少一个处理器执行，以使至少一个处理器能够执行本公开第一方面实施例或第二方面实施例中描述的方法。
本公开第六方面实施例提供一种计算机存储介质，其中，计算机存储介质存储有计算机可执行指令；计算机可执行指令被处理器执行后，能够实现本公开第一方面或第二方面实施例中描述的方法。
根据本公开的数据采集方法，对生物组织样本进行切片，并对切片进行时空组测序，以得到表达量数据；对切片后的生物组织样本进行图像采集，以得到图像数据；对表达量数据和/或图像数据进行配准，以生成生物组织样本的外层轮廓和/或分区轮廓；根据外层轮廓和/或分区轮廓，生成生物组织样本的三维展示数据。本公开的方法针对原位捕获的时空组学技术，结合光学成像信息，辅助连续组织切片的刚性、弹性配准，实现三维重构，进而得到内外组织轮廓辅助可视化信息定位，对大尺寸高分辨率时空组数据进行有效处理与展示。
本公开附加的方面和优点将在下面的描述中部分给出，部分将从下面的描述中变得明显，或通过本公开的实践了解到。
附图说明
本公开上述的和/或附加的方面和优点从下面结合附图对实施例的描述中将变得明显和容易理解，其中：
图1为根据本公开实施例的一种数据采集方法的流程示意图；
图2为根据本公开实施例的一种数据采集方法的流程示意图；
图3为根据本公开实施例的一种数据采集方法的示意图；
图4为根据本公开实施例的一种切片机的示意图；
图5为根据本公开实施例的一种拍照装置滑动位置的示意图；
图6为根据本公开实施例的一种表达量数据预处理方法的示意图；
图7为根据本公开实施例的一种图像采集和时空测序规则的示意图；
图8为根据本公开实施例的一种拍照装置拍摄样例的示意图；
图9为根据本公开实施例的一种图像分割处理策略的示意图；
图10为根据本公开实施例的一种图像分割处理策略的示意图；
图11为根据本公开实施例的一种刚性配准算法的示意图；
图12为根据本公开实施例的一种图像分割和配准结果的示意图；
图13为根据本公开实施例的一种组织外轮廓生成流程的示意图；
图14为根据本公开实施例的一种图像渲染的效果示意图；
图15为根据本公开实施例的一种图像分割和配准效果的示意图；
图16为根据本公开实施例的一种表达量数据配准流程的示意图；
图17为根据本公开实施例的一种直方图均衡化的效果示意图；
图18为根据本公开实施例的一种刚性配准流程的示意图；
图19为根据本公开实施例的一种刚性配准结果的示意图；
图20为根据本公开实施例的一种弹性配准流程的示意图；
图21为根据本公开实施例的一种弹性配准结果的示意图；
图22为根据本公开实施例的一种表达量数据连续配准的示意图；
图23为根据本公开实施例的一种分区轮廓模型生成流程的示意图；
图24为根据本公开实施例的一种统领文件基本数据结构的示意图；
图25为根据本公开实施例的一种统领文件内单基因分层分块数据结构的示意图；
图26为根据本公开实施例的一种统领文件内细胞类型分层分块数据的示意图；
图27为根据本公开实施例的一种鼠脑重构渲染效果的示意图；
图28为根据本公开实施例的一种鼠脑可展示单基因列表的示意图；
图29为根据本公开实施例的一种生成组织外轮廓模型的效果示意图；
图30为根据本公开实施例的一种生成组织分区轮廓的效果示意图；
图31为根据本公开实施例的一种弹性配准结果的示意图；
图32为根据本公开实施例的一种数据采集装置的示意框图；
图33为根据本公开实施例的一种数据处理装置的示意框图；
图34为根据本公开实施例的一种电子设备的结构示意图。
具体实施方式
下面详细描述本公开的实施例，实施例的示例在附图中示出，其中自始至终相同或类似的标号表示相同或类似的元件或具有相同或类似功能的元件。下面通过参考附图描述的实施例是示例性的，旨在用于解释本公开，而不能理解为对本公开的限制。
时空组学技术自发展以来，已在神经科学、病理学、发育生物学等方向得到初步应用。基于原位捕获的方法，通常为依靠芯片和组织切片间的物理接触和探针捕获来获取原位空间信息及各种组学信息，对组织的破坏性切割以及平面捕获相应信息限制其只能用于组织的二维平面时空组学分析，难以得到三维空间的形态结构及准确的位置信息。而不同器官和组织具有独特的三维细胞分布和基因表达图谱，对组织进行三维时空组学研究具有极大的技术挑战性和重要的科学意义。
三维时空组学的关键点在于如何突破二维平面的研究，实现组织三维水平上不同细胞的空间分布、相互作用、基因蛋白质等表达情况以及整体形态结构的全面解析。
当前，基于时空组的三维重构流程主要是将相邻或相近切片做切片之间的刚性配准，然后基于聚类或注释信息，生成分区轮廓，再基于切片堆生成整个组织/生物体的轮廓。三维重构的结果通常是以文章中的图表或者代码库的demo图片形式进行展示。存在以下不足：
(1)仅依据时空组捕获的表达量丰度信息，将测量于平面上的表达量数据进行刚性配准，无法校正实验过程中因人员操作和生化过程等带来的表达量数据的形变，从而没有将数据重构回原位表达水平：
(2)采用基于表达量丰度信息的刚性配准的方法，需要捕获到的表达量丰度信息来自相邻或相近切片。如果切片距离较远，刚性配准的准确性会有所下降；
(3)还原后的三维空间信息需要投影到如电脑屏幕，纸质期刊文献等二维平面来可视化，在无其他辅助信息如边界线和面的情况下，二维平面内的散点可提供的三维效果有限，尤其当散点数量较大且无明显组织轮廓信息时，不利于人眼识别三维空间信息；
(4)受技术和样本等限制，单张组织切片上信息点数量较少，当样本物理尺寸和芯片分辨率提高后，下游分析和展示时受硬件条件限制，对于大数据量的组织样本数据处理和后期展示能力不足。
为此，本公开提出了一种数据采集方法、装置及系统，基于原位捕获的时空组学技术，结合原位光学影像得到原始组织轮廓，辅助内部采集到的生物信息的定位，以还原组织切割前的三维位置信息，减少实验操作对于组织形态的影响，对于芯片上每个信息采集位点实现细胞精度的位置配准，对于较大尺寸和数据量的组织样本进行有效的处理和展示。
下面结合附图对本申请所提供的传输配置方案进行详细介绍。
图1示出了根据本公开实施例的一种数据采集方法的流程示意图。如图1所示，该方法可以包括以下步骤。
S101，对生物组织样本进行切片，并对切片进行时空组测序，以得到表达量数据。
具体地，对生物组织样本进行连续切片，根据实际情况选取指定位置切片进行时空组测序。
在本公开的实施例中，生物组织样本取自生物样品，生物样品通常是指动物(包括人)的体液(如尿血、唾液、胆汁、胃液、淋巴液及生物体的其他分泌液等)、毛发、肌肉和一些组织器官(如胸腺、胰腺、肝、肺、脑、胃、肾等)以及各种微生物等。
在本公开的实施例中，对生物组织样本进行切片的设备不予限制，优选地采用生物组织切片机对生物组织样本进行切片，精度和稳定性更高。
在本公开的实施例中，基于原位捕获的方法获取表达量数据，可以采用捕获测序芯片信号的方法获取表达量数据，其中芯片的性能决定了精度和分辨率等技术效果，优选地基于DNA纳米球测序芯片进行组织的原位捕获，芯片大、信号密度高实现更优的分辨率和视野。另外，也可以应用目前处于开发中的蛋白组学等其他基于原位捕获测序的时空组数据，本公开不予限制。
本公开的实施例中，在每次选取切片进行空组测序时，记录切片对应位置，切片的对应位置用于后续三维重构。
S102，对切片后的生物组织样本进行图像采集，以得到图像数据。
本公开的实施例中，可以采用光学成像对切片后的生物组织样本进行图像采集，也可采取其他不影响时空组学的原位成像技术手段，本公开不予限制。
本公开的实施例中，对所有选取切片进行时空组测序对应的切片后的组织进行图像采集，主要采集切片截面的图像，用于后续配准，同时也可以选取其他切片后的生物组织样本进行图像采集，具体选取规则本公开不予限制。举例而言，每切m次对切片后的生物组织样本进行图像采集，每切n次取切下的切片做时空测序，其中m、n为正整数，可以根据实际情况调整，n可以为m的倍数。
本公开的实施例中，在每次选取切片进行图像采集时，记录切片对应位置，切片的对应位置用于后续三维重构。
S103，对表达量数据和/或图像数据进行配准，以生成生物组织样本的外层轮廓和/或分区轮廓。
可选地，在匹配之前，还包括对表达量数据进行预处理，对图像数据进行图像分割处理，其中预处理可以包括聚合处理、质量控制、对数归一化、找高变基因、降维处理中的至少一项。其中图像分割处理主要指的是对使用图像分割算法对其进行组织分割，判别删除不合格的分割样本，保留合格的分割结果作为后续配准的图像数据。
在本公开的实施例中，该方法可以包括对图像数据进行配准，以生成生物组织样本的外层轮廓。
在本公开的一种实施例中，通过对图像数据进行图像分割处理，生成分割结果图，对分割结果图进行配准，生成配准结果图，最后基于配准结果图，生成生物组织样本的外层轮廓实现。其中，选取第一张分割结果图对其进行旋转校正，校正后作为第一张参考的图对其他分割图进行配准，具体可以采用刚性配准的方式或其他配准方法，本公开不予限制。
在本公开的实施例中，该方法可以包括对表达量数据和图像数据进行配准，以生成生物组织样本的分区轮廓。
在本公开的一种实施例中，通过对表达量数据进行计算，得到表达量图，将表达量图和图像数据进行直方图均衡化，得到表达量图-图像数据的图像对，对图像对依次进行刚性配准和弹性配准，生成分区轮廓实现。
在本公开的实施例中，不使用图像数据时，例如当没有采集图像或其他原因，该方法还可以包括对表达量数据进行配准，以生成生物组织样本的外层轮廓和/或分区轮廓，但是需要参与配准的时空组切片相邻或相近。
在本公开的一种实施例中，通过对两个相邻或相近切片中的表达量数据进行配准，以得到配准表达量数据，根据配准表达量数据，生成生物组织样本的外层轮廓和/或分区轮廓实现。
应当的理解的是，通过上述任一方法可以生成生物组织样本的外层轮廓和/或分区轮廓，根据实际研究或应用的需求选择，本公开不予限制。
在本公开的实施例中，基于配准后的时空组数据生成生物组织样本的轮廓数据，具体地，对表达量数据和/或图像数据进行配准得到的配准结果进行生物信息学分析，得到聚类和注释结果，基于聚类和注释结果，构造三维数组，并增加空间尺度信息，对每个聚类或分区结果进行分割，生成样本内部结构多个组织或者区域的平滑轮廓面。其中，可以采用类似3D slicer中的segment editor等工具，对每个聚类或分区结果进行分割，最终生成的轮廓图以vtk、obj、stl等常见的3D矢量模型数据存储。
S104，根据外层轮廓和/或分区轮廓，生成生物组织样本的三维展示数据。
在本公开的实施例中，利用三维渲染工具，渲染得到生物组织样本的三维展示数据，所采用的渲染工具本公开不与限制
在本公开的一种实施例中，外层轮廓和/或分区轮廓数据整合其他图像渲染数据、配准后时空组数据等组成项目统领文件，其中综合考虑传输效率和渲染效果，生成项目统领文件时对时空组数据按单基因表达量、细胞类型等角度分别进行数据的分层分块处理。将生成好的统领文件，通过基于flask的后端程序加载到渲染器，可以实现对基因表达分布、组织轮廓等信息在三维空间内的情况的直观展示。
综上，根据本公开实施例提供的传输配置方法，对生物组织样本进行切片，并对切片进行时空组测序，以得到表达量数据；对切片后的生物组织样本进行图像采集，以得到图像数据；对表达量数据和/或图像数据进行配准，以生成生物组织样本的外层轮廓和/或分区轮廓；根据外层轮廓和/或分区轮廓，生成生物组织样本的三维展示数据。本公开针对原位捕获的时空组学技术，结合光学成像信息，进行配准，得到内外组织轮廓辅助可视化信息定位，可以对大尺寸高分辨率时空组数据进行有效的处理与展示，应用于三维时空组学技术可以实现基因与影像同时分析。
图2示出了根据本公开实施例的一种数据采集方法的流程示意图。基于图1所示实施例，如图2所示，该方法可以包括以下步骤。为方便理解，图3示出了根据本公开实施例的一种数据采集方法的示意图。
在本公开中，为了便于理解方案，还详细描述了本方法的一种实施方式、过程以及效果，以获取猕猴脑数据为例，将在每个步骤的实施方式中具体展开描述，应该理解的是，该示例不是对方法的限制。
基于图1所示实施例，步骤S101包括了图2所示的步骤S201。
S201，对生物组织样本进行连续切片，每切第一预定数量的切片进行一次时空组测序，以得到切片的表达量数据。
其中，第一预定数量为正整数，例如每切10张取一张切片做时空组测序。
本公开的实施例中，优选采用生物组织切片机对生物组织样本进行切片，精度和稳定性更高。具体地，将生物组织样本固定在切片机的冷冻头位置，根据后续拍照位置的要求，大致调整好动头的位置和角度，首先对组织切片进行修片操作以去掉多余的OCT包埋剂，直至露出组织区域，对准组织进行连续切片。
在本公开的一种实施方式中，以获取猕猴脑数据为例，采用切片机进行切片，如图4所示，首先实验人员准备猴脑组织样品，并将猴脑组织样品固定在切片机冻头的中心位置。其次将含待切猴脑组织的OCT(冰冻切片包埋剂)包埋块固定在样本托上，调节切面至合适角度，在图4所示指示屏上设置为修片模式，调整修片厚度，依据客户提供的样本包埋状态信息，摇动右侧手轮进行空转粗修切面，使OCT表面被切平整；调焦前，需确保样本头的前进位移显示刻度指针在15mm左右，因此在预估包埋块可能跳出此范围时，需要回退切片机冻头，此时要按回退按钮，回退到多少刻度需要根据猴脑组织的剩余厚度情况决定，接下来需要调整刀架使其与组织面贴平，具体为：当包埋块厚度在10mm左右厚度时，此时不需要调刀架；当猴脑组织大于10mm时，比如说为15mm厚的包埋块，由于冻头只能前进10mm，所以当冻头不能前进时需要将冻头回退至少5mm，具体移动的数值可以从右图显示屏读出，此时刀架也要随之前移到与组织面大致相贴，此时第一刀作为平整度调整应去掉，从第二刀开始新的一轮实验。
本公开的实施例中，基于原位捕获的方法获取表达量数据，可以采用捕获测序芯片信号的方法获取表达量数据，其中芯片的性能决定了精度和分辨率等技术效果，优选地基于DNA纳米球测序芯片进行组织的原位捕获，芯片大、信号密度高实现更优的分辨率和视野。另外，也可以应用目前处于开发中的蛋白组学等其他基于原位捕获测序的时空组数据，本公开不予限制。
在本公开的一些实施例中，得到切片的表达量数据后，还包括对表达量数据进行预处理，其中，预处理包括聚合处理、质量控制、对数归一化、找高变基因、降维处理中的至少一项。
在本公开中，对数据进行预处理有利于减少表达量数据所带的噪音，减少因实验条件和操作而带来的数据规模效应，并由此得到更准确的注释结果、更能体现组织结构的表达量图、以及渲染器中更准确的展示结果。
在本公开中，如图6所示，还包括对数据进行注释，基于预处理的注释结果将用于分区轮廓的生成，以及渲染器中数据的展示。
应当说明的是，具体预处理和注释方法由流程的使用者自行决定，可以采用既有的处理流程，例如binset聚合、质量控制、对数归一化(又称为‘测序深度标准化’)、找高变基因、降维、聚类的流程，也可以采用自行设计的流程。
在本公开的一种实施方式中，以获取猕猴脑数据为例，如图6所示，上部示出了猕猴脑切片的原始表达量数据，下部示出了预处理和注释后对应的表达量数据。
基于图1所示实施例，步骤S102包括了图2所示的步骤S202。
S202，每切第二预定数量的切片，对切片后的生物组织样本进行一次图像采集，以得到图像数据。
其中，第二预定数量为正整数。
本公开的实施例中，测序和图像采集按照一定规则进行，其中第一预定数量为第二预定数量的倍数，时空组测序必然与拍照一起发生，有时拍照但不测序，根据实际情况调整。举例而言，如图7所示，第一预定数量为10，第二预定数量为5，将生物组织样本切为13片，第一次切片后，对切片进行时空组测序、对切片后的生物样本进行图像采集，连续切5片后，对切片后的生物样本进行图像采集，再连续切5片后，对切片进行时空组测序、对切片后的生物样本进行图像采集，直到切完为止。
本公开的实施例中，可以采用光学成像对切片后的生物组织样本进行图像采集，也可采取其他不影响时空组学的原位成像技术手段，本公开不予限制。
本公开的一种实施例中，采用拍照设备对切片后的生物组织样本进行图像采集，具体地，如图5所示，将拍照设备对准组织切片，系统校准，使样本截面在焦，通过滑台软件控制面板记录当前相机位置。按照计划对组织进行切片，并记录切刀流水。取第一张切片进行时空组测序，对切完剩下的组织切面进行BlockFace拍照，具体拍照和时空组测序规则按图7所示进行。
在本公开的一种实施方式中，图像采集方式采用原位光学成像，以获取猕猴脑数据为例，采用的拍照设备如图4所示，首先，连接拍照设备和电脑。具体地，选取合适地镜头，调整镜头上白色圆点对应的刻度值大小表示此时镜头的实际放大倍率，连接好相机、滑台的电源线和USB信号线，其中USB信号线要接到电脑的USB3.0接口，在电脑桌面打开相机拍照软件，选择相机的“MV-CH250-90UC”，使相机正常连接。其次，通过操作滑台软件和相机软件，以控制拍照装置进行图像采集。具体地，启动滑台软件，选择Easy模式，接着确认伺服，出现滑台的操作界面，设置滑台前进和后退的数值大小，以100为例，实际过程中一般在80～100mm之间，具体大小以初始焦面为准(拍照装置还没移动到切片机拍照的位置时，只设置数值，不点击“+”)。如图5所示，将拍照装置从初始端，沿滑轨滑动至限位端，并打开舱门的安全位置。选择相机软件的视频键，此时在相机的视频显示窗口会出现视频实时显示画面，然后点击滑台“+”(输入值为上述的设置即100)，控制滑台会伸入到切片机内，则在显示框会出现组织截面，如果组织面不在焦面的，设置滑动单位，可在滑台软件框输入框单 位“1”，根据实际情况选择“-”和“+”操作(选择“-”则表示滑台向外伸出，选择“+”则表示滑台向内伸入)，通过这种微调，使OCT上的黑色Mark点(用于标记微调)能在相机显示框清晰的成像(Mark点在标记前需保证OCT已被切片机切平)，确定OCT处于焦面后，在Mark点聚焦的基础上连续切片OCT直至微落出需要的组织面，此时就可以按照切片流程正式进行切片了。最后，拍照保存图像，将拍照装置复位。具体地，对于需要拍照保存的Block Face图片，点击相机拍照界面的“抓拍图像按键”，则此时的图像被保存。执行完成一次拍照后，点击滑台软件的“原点复位”，如图5所示，使拍照装置从机箱内，安全退出机箱外，并关闭舱门，这时实验人员再将拍照装置移动至初始位置，则单次拍照完成，拍照效果如图8所示。
图2所示步骤S203-205，是对图1实施例中步骤S103：对表达量数据和/或图像数据进行配准，以生成生物组织样本的外层轮廓和/或分区轮廓的进一步公开。
S203，对图像数据进行图像分割处理，以生成分割结果图。
在本公开地一种实施例中，具体地，将光学成像拍的BlockFace图像，使用图像分割算法对其进行组织分割，判别删除不合格的分割样本，保留合格的BlockFace图像分割结果进行后续处理。
在本公开的一些实施例中，对图像数据进行图像分割处理包括：将图像数据转换为HSV颜色空间，提取H通道的图像数据并转换为灰度图；通过预设图像处理算法，对图像数据进行二值化，以确定二值化结果轮廓图；保留符合预设条件的轮廓图，与图像数据进行合并。
具体的，如图9所示，首先对原始BlockFace图进行分割处理，去掉大面积的背景，将图像转换为HSV颜色空间格式，提取H通道转换为灰度图，通过Li算法对图像进行二值化，寻找二值化结果轮廓图，通过轮廓一阶矩，保留大于阈值的轮廓，与原图取并获得目标分割区域。
可选地，对图像数据进行图像分割处理还可以包括：获取图像数据的灰度化图像和直方图；对图像数据进行最大熵阈值二值化处理；将处理后的图像与图像数据进行合并。
在本公开中，优先采用如图9所示策略进行图像分割处理，对于不适合采用该策略分割的图像，或者采用该策略分割错误的图像，可选的，可以采用如图10所示策略进行图像分割。例如，由于组织区域的BlockFace图本身的多样性，如果BlockFace图中背景像素和待分割的组织目标像素值很接近，那么使用Li算法得到的阈值就不可能适用于所有情况。
具体地，如图10所示，首先对原始BlockFace图进行分割处理，去掉大面积的背景，然后对去掉大面积背景后的图进行灰度化处理，统计灰度化后图的直方图(灰度直方图是将数字图像中的所有像素)，按照灰度值的大小，统计其出现的频率。基于对象和背景之间信息度量的最大化得到最大熵二值化阈值，使用它对灰度化的图像进行分割。根据分割后的mask图与去掉大面积背景后的图取并获取目标分割区域。
S204，对分割结果图进行配准，以生成配准结果图。
在本公开的实施例中，在进行配准之前，取第一张BlockFace分割结果图对其进行旋转校正，校正后作为第一张参考的图对其他分割图进行配准。
在本公开的一些实施例中，对分割结果图进行配准，以生成配准结果图包括：对分割结果图进行空间变换处理和线性插值算法处理；对经处理的分割结果图以及配准参考图进行均方差相似度测量；将测量最优解确定为配准结果图。
在本公开的一种实施例中，BlockFace图像分割结果配准逻辑如图11所示，设置刚性配准算法的固定步长的梯度下降优化算法的最大迭代次数为K和学习率为r，步长为t，(K、r、t的值根据实际情况调整)，实现配准。具体地，首先对BlockFace图像分割结果图进行空间变换、线性插值算法，第一张BlockFace分割结果图作为参考图，与参考图进行均方差相似度测试，需要优化则进行固定步长r的下降优化算法，再次进行空间变换、线性插值算法，直到达到最大迭代次数K，获得最优结果输出，即获得配准结果图。
对于步骤S203和步骤S204，以鼠心组织的BlockFace图像为例，展示对鼠心组织的BlockFace图进行分割和对其进行配准的效果，如图12所示。
S205，基于配准结果图，生成生物组织样本的外层轮廓。
在本公开的一些实施例中，基于配准结果图，生成生物组织样本的外层轮廓包括：将所有配准结果图堆叠为预设格式的数组文件；对数组文件进行处理，以得到生物组织样本的外层轮廓。
在本公开的一种实施例中，如图13所示，对配准结果图进行Volume和Mesh的生成，具体地，将配准后的BlockFace分割图堆叠成为一个.npy格式的文件。得到的npy数组文件，其中非0值的元素所占据的空间位置为样本组织，0值的元素为空间背景，因此可以以医学图像格式NIfTI(Neuroimaging Informatics Technology Initiative)来进行存储。通过SimpleITK、Nibabel等图像处理工具，辅以切片的三维空间尺度，存储为nii格式的文件，然后通过3Dslicer等相关工具，进行segment-make surface的操作，形成平滑的组织外部轮廓，以vtk、obj、stl等常见的3D矢量模型数据存储。
在本公开的实施例中，还包括处理配准后的光学成像结果得到渲染所需格式。
具体地，将配准后的BlockFace分割图堆叠成为一个.npy格式或.tif的文件,通过python的工具包CloudVolume和Igneous将Blockface分割图像处理成渲染器支持的带分层分块结构的二进制渲染文件，最终渲染效果如图14所示。
在本公开的一种实施方式中，如图15所示，以获取猕猴脑数据为例，首先利用ImageJ对BlockFace拍照图进行旋转校正并选择感兴趣的ROI分析区域，接着对BlockFace ROI图像进行组织目标对象和背景的分割，并将BlockFace分割图进行配准，形成空间中对齐的BlockFace图像堆，算法依据BlockFace分割图生成对象的最外层轮廓，依据BlockFace分割图生成对象的最外层轮廓。
在本公开的一些实施例中，在步骤S203-205生成生物组织样本的外层轮廓之后，还包括：对表达量数据进行计算，得到表达量图；将表达量图和图像数据进行直方图均衡化，得到表达量图-图像数据的图像对；对图像对依次进行刚性配准和弹性配准，以生成分区轮廓。
具体地，如图16所示，将时空组的表达量数据处理成为反映不同区域表达量丰度的表达量图，分别对表达量图和BlockFace分割图做直方图均衡化处理，直方图均衡化处理的效果如图17所示。然后在切片位置相互近邻的BlockFace分割图-表达量图图像对中，做如下操作：以BlockFace分割图为参考图像，使用表达量图与之进行刚性配准和弹性配准。完成两种配准后，将配准中决定的变换关系(刚性配准的变换关系形如一个3*3的变换矩阵，弹性配准的变换关系形如两个形变场，分别表示x和y方向的形变的分布)复用在表达量数据上，实现表达量数据与BlockFace分割图的配准。
可选地，对图像对进行刚性配准包括：对图像对分别进行下采样；提取图像对的特征信息；对图像对进行相似性估计仿射变换，得到刚性配准结果。
在本公开的一种实施例中，如图18所示，图像对分别进行下采样，并进行噪声过滤去除伪影，分别通过深度学习模型提取两种不同模态的图像中的目标组织的边缘特征和形状特征和区域分割信息，然后在模型输出处最大化目标组织最外层轮廓的相似性估计仿射变换，最后将仿射变换应用在待配准的表达量数据原始图像上，将处理完的一对表达量数据和BlockFace分割图结果送至弹性配准环节处理。其中刚性配准的结果图如图19所示。
可选地，对图像对进行弹性配准包括：对刚性配准结果进行直方图均衡化；将刚性配准结果输入深度学习模型，输出模型的中间结果，得到形变场，其中，深度学习模型是通过表达量图-图像数据对进行训练得到的；根据形变场，对刚性配准结果进行最近邻插值变换，得到弹性配准结果。
在本公开的一种实施例中，弹性配准的实现基于的是深度学习模型的推导，如20图。对应的深度学习模型要经过如下过程获得：首先将表达量数据处理成为表达量图后，分别对表达量图和对应的BlockFace分割图进行直方图均衡化。然后在软件中对表达量图-BlockFace分割图进行刚性配准，作为模型训练的输入图像，再进行弹性配准，作为模型训练的标签。再基于手动弹性配准时软件输出的控制点坐标，加入随机参数，生成随机变形的表达量图，实现数据合成。最后，使用合成的和真实的数据训练有监督模型，在训练指标达到要求后停止，保存模型。获得深度学习模型之后，就可以使用此模型进行弹性形变的推导。将实际问题中的表达量数据和BlockFace分割图按照上述的处理方法，获得直方图均衡化的图像。将两个图像输入模型，输出模型的中间结果：形变场。根据形变场，获取对应位移，对表达量数据进行最近邻插值变换，计算得到新的表达量数据。图21所示，以猴脑为例，展示弹性配准前后，表达量图和BlockFace图的对应情况变换：其中红色的为BlockFace图像，绿色的为表达量图像。弹性配准后，可视的表达量图的形变得到了校正。
在另一种可选的方式中，也可以使用表达量数据直接进行配准生成生物组织样本的外层轮廓和/或分区轮廓，方法包括：对两个相邻或相近切片中的表达量数据进行配准，以得到配准表达量数据；根据配准表达量数据，生成生物组织样本的外层轮廓和/或分区轮廓。
应当的理解的是，在没有进行图像采集的情况下，比如出于用户的习惯不进行拍照、没有拍照设备等情况下，执行该方法也可以生成生物组织样本的外层轮廓和/或分区轮廓，但是需要参与配准的时空组切片相邻或相近。
在本公开的一种实施例中，如图22所示，依据两张相邻或相近切片中的表达量数据计算bin set或cell bin，经过对数归一化等预处理后，解最优传输问题，得到最优传输的转移概率矩阵(简称pi矩阵)；对pi矩阵进行后处理后，可以计算刚性，或刚性+弹性的变换关系；最后将此变换关系应用到原始表达量数据上。
S206，对表达量数据和/或图像数据进行配准得到的配准结果进行生物信息学分析，得到聚类和注释结果。
S207，基于聚类和注释结果，构造三维数组，并增加空间尺度信息。
在本公开中，基于表达量数据和/或图像数据进行配准得到的配准结果生成外层轮廓和/或分区轮廓模型。
在本公开的一种实施例中，根据基于图像数据进行配准得到的配准结果生成外层轮廓，具体地，如图13所示，参照步骤S205。
在本公开的一种实施例中，根据表达量数据和/或图像数据进行配准得到的配准结果生成分区轮廓，具体与步骤S205相似，如图23所示，对配准好的表达量数据或，每一个bin set或者cell bin在通过常规的生物信息学分析，得到聚类和注释结果，以每个bin set或cell bin的注释或者聚类结果的值为单位，构造三维npy数组，同时增加空间尺度信息，得到基于注释聚类结果的NIfTI文件。再运用类似3D slicer中的segment editor等工具，对每个聚类或分区结果进行分割，生成样本内部结构多个组织或者区域的平滑轮廓面，以vtk、obj、stl等常见的3D矢量模型数据存储。对应于上述步骤S205。
在本公开中，生成后的外层轮廓和/或分区轮廓模型，通过python的工具包multiresolution-mesh-creator，将文件处理成与分区标签绑定后的二进制3D模型文件。
在本公开中，整合上述图像渲染数据、配准后时空组数据、组织轮廓数据、分区轮廓数据等成项目统领文件，文件内容结构根据研究或应用需求决定。
在本公开的一种实施例中，将配准后的时空组表达量数据通过geftools生成cgef格式文件，再与前述步骤中生成的二进制图像数据、二进制分区轮廓模型数据、组织外轮廓模型数据等通过gef3d生成统领文件，文件内容结构如图24所示，主要展示了统领文件的前三级目录。
在本公开的一些实施例中，综合考虑传输效率和渲染效果，可以在生成项目统领文件时对时空组数据按单基因表达量、细胞类型等角度分别进行数据的分层分块处理。
其中，分层分块是指对大量数据点云按照空间位置进行类似金字塔结构的抽样，并保证所有层的所有块内数据点并集为全部数据。该处理在gef3d生成统领文件时自动进行，需指定每块的bin/cellbin个数的阈值。
在本公开的一种实施例中，分层分块方法可以包括：首先遍历所有cgef获取到整体的空间大小，设定一个最大层数，根据数据点三维空间各自的长度，计算好划分尺度，从上到下按照8分法预先划分出每层的块数，每块近似立方体。默认第一层只有一块。经过第一步的遍历可以获取到每个gene在各个切片中的数据，也就是在整个空间的数据点的属性值，包括坐标及表达量丰度等。对单个gene的数据进行分层分块，根据第一步计算的结果，从顶层到下层，依次从gene的整体数据中去填充每一层的每一小块，直至小块填满(由阈值控制)。剩下的部分继续填充下一层，当下一层某一块数据太少(阈值控制)，会将本块数据添加到上一层的父块中去。依次向下分配，直至gene的数据分配完。对于celltype等细胞类型注释结果的分层分块与单基因类似。
分层分块后的表达量或细胞类型等数据在统领文件中存储结构如图25、图26所示，在L0目录下，其中，blk表示块，为block的简写，该目录用于储存分块数据情况，在blk目录下，offset表示偏移量，储存该分块数据的起始索引偏移量，count表示数目，储存该分块数据的点数量；cell表示细胞，该目录用于存储该层级下所有bin/cellbin数据相关信息，在cell目录下，x、y、z分别储存该bin/cellbin空间坐标的x、y、z坐标值，umicnt储存该bin/cellbin表达量，cellid表示细胞序号，储存该bin/cellbin序号。在attrs目录下，其中，ctype_list表示细胞类型列表，储存细胞类型序号、名称和颜色的列表；maxlevelnum表示最大分层数，所有细胞类型数据中分层数的最大值；blkcnt表示分块数，储存不同层级的各轴向分块数量；maxumi表示表达量最大值，储存表达量最大值。
S208，利用三维渲染工具，渲染得到生物组织样本的三维展示数据。
具体地，将经上述步骤生成好的统领文件，通过基于flask的后端程序加载到渲染器，可以实现对基因表达分布、组织轮廓等信息在三维空间内的情况的直观展示。在网页端进行展示，效果如图27所示，其中分别加载了整体表达量数据，组织外轮廓模型。
综上，根据本公开实施例提供的数据采集方法，针对原位捕获的时空组学技术，结合光学成像信息，辅助连续组织切片的刚性、弹性配准，实现三维重构，进而得到内外组织轮廓辅助可视化信息定位，对大尺寸高分辨率时空组数据进行有效处理与展示。
本公开的方法效果展示如下：
1.通过步骤S204和步骤S205对采集的样本图像数据进行图像分割处理方法和图像配准，以鼠心组织的BlockFace图像为例，展示对鼠心组织的BlockFace图进行分割和对其进行配准的效果，如图12所示。
2.参照上述图18对应方法，以猴脑为例，展示弹性配准前后，表达量图和采集图像的对应情况变换，效果如图19所示，弹性配准后，可视的表达量图的形变得到了校正。
3.生成组织外轮廓模型，与转录组等空间散点信息结合进行展示，如图28所示，从左到右依次为，鼠脑组织轮廓及细胞类型展示、鼠脑组织轮廓及单基因表达量展示以及鼠脑组织轮廓模型及基因总表达量展示。
4.生成组织分区轮廓模型，如图29所示，左边为鼠脑组织分区轮廓模型展示，右边为鼠脑组织分区轮廓模型及对应的细胞类型bin展示。
5.对于较大尺寸和数据量的组织样本的处理和展示如图30所示，选取的展示例为鼠脑的100张以上时空组芯片数据。芯片大小为1cm x 1cm，所展示的数据约900K个坐标点，包括约22K个基因。
因此，本公开的方法可以更好地还原组织样本切片前的组织轮廓以及相关表达量数据，应用于时空组学三维重构，将有利于推动三维时空组学技术的研究和应用。
图31为根据本公开实施例的一种数据处理方法的流程示意图。如图31所示，该方法可以包括以下步骤。
S301，获取生物组织样本的表达量数据和图像数据。
其中，表达量数据是对生物组织样本进行切片并对切片进行时空组测序得到的，图像数据是对切片后的生物组织样本进行图像采集得到的。
在本公开的实施例中，对生物组织样本进行连续切片，根据实际情况选取指定位置切片进行时空组测，获取表达量数据。
在本公开的实施例中，可以采用光学成像对切片后的生物组织样本进行图像采集，获取图像数据，也可采取其他不影响时空组学的原位成像技术手段，本公开不予限制。
在本公开的实施例中，生物组织样本取自生物样品，生物样品通常是指动物(包括人)的体液(如尿血、唾液、胆汁、胃液、淋巴液及生物体的其他分泌液等)、毛发、肌肉和一些组织器官(如胸腺、胰腺、肝、肺、脑、胃、肾等)以及各种微生物等。
在本公开的实施例中，对生物组织样本进行切片的设备不予限制，优选地采用生物组织切片机对生物组织样本进行切片，精度和稳定性更高。
在本公开的实施例中，基于原位捕获的方法获取表达量数据，可以采用捕获测序芯片信号的方法获取表达量数据，其中芯片的性能决定了精度和分辨率等技术效果，优选地基于DNA纳米球测序芯片进行组织的原位捕获，芯片大、信号密度高实现更优的分辨率和视野。另外，也可以应用目前处于开发中的蛋白组学等其他基于原位捕获测序的时空组数据，本公开不予限制。
本公开的实施例中，对所有选取切片进行时空组测序对应的切片后的组织进行图像采集，主要采集切片截面的图像，用于后续配准，同时也可以选取其他切片后的生物组织样本进行图像采集，具体选取规则本公开不予限制。
本公开的实施例中，在每次选取切片进行时空组测序或图像采集时，记录切片对应位置，切片的对应位置用于后续三维重构。
在本公开中，获取表达量数据后，还可以包括对表达量数据进行预处理，其中，预处理包括聚合处理、质量控制、对数归一化、找高变基因、降维处理中的至少一项。对数据进行预处理有利于减少表达量数据所带的噪音，减少因实验条件和操作而带来的数据规模效应，并由此得到更准确的注释结果、更能体现组织结构的表达量图、以及渲染器中更准确的展示结果。
在本公开中，获取图像数据后，还可以包括对图像数据进行分割处理，具体地，将图像数据转换为HSV颜色空间，提取H通道的图像数据并转换为灰度图，通过预设图像处理算法，对图像数据进行二值化，以确定二值化结果轮廓图，保留符合预设条件的轮廓图，与图像数据进行合并。
可以参照步骤S201和步骤202的具体实施方式，本公开中不再赘述。
S302，对表达量数据进行计算，得到表达量图。
本公开的实施例中，根据实际研究需求，对时空组的表达量数据的相关系数进行计算，得到可以反映不同区域表达量丰度的表达量图，具体的计算、统计及画图方式本公开不予限制。
S303，将表达量图和图像数据进行直方图均衡化，得到表达量图-图像数据的图像对。
其中，直方图均衡化是一种简单有效的图像增强技术，通过改变图像的直方图来改变图像中各像素的灰度，主要用于增强动态范围偏小的图像的对比度。原始图像由于其灰度分布可能集中在较窄的区间，造成图像不够清晰。例如，过曝光图像的灰度级集中在高亮度范围内，而曝光不足将使图像灰度级集中在低亮度范围内。采用直方图均衡化，可以把原始图像的直方图变换为均匀分布(均衡)的形式，这样就增加了像素之间灰度值差别的动态范围，从而达到增强图像整体对比度的效果。
在本公开的一种实施例中，经过步骤S301采集BlockFace图像作为图像数据，并对图像进行分割处理获取BlockFace分割图，取切片位置相互近邻的BlockFace分割图-表达量图像对中作为表达量图-图像数据的图像对。
S304，对图像对依次进行刚性配准和弹性配准，以生成生物组织样本的轮廓数据，轮廓数据用于生成生物组织样本的三维展示数据。
在本公开的一种实施例中，如图16所示，以BlockFace分割图为参考图像，使用表达量图与之进行刚性配准和弹性配准。完成两种配准后，将配准中决定的变换关系(刚性配准的变换关系形如一个3*3的变换矩阵，弹性配准的变换关系形如两个形变场，分别表示x和y方向的形变的分布)复用在表达量数据上，实现表达量数据与BlockFace分割图的配准。
在本公开的可选实施例中，对图像对进行刚性配准包括：对图像对分别进行下采样；提取图像对的特征信息；对图像对进行相似性估计仿射变换，得到刚性配准结果。
在本公开的一种实施例中，如图18所示，图像对分别进行下采样，并进行噪声过滤去除伪影，分别通过深度学习模型提取两种不同模态的图像中的目标组织的边缘特征和形状特征和区域分割信息，然后在模型输出处最大化目标组织最外层轮廓的相似性估计仿射变换，最后将仿射变换应用在待配准的表达量数据原始图像上，将处理完的一对表达量数据和BlockFace分割图结果送至弹性配准环节处理。其中刚性配准的结果图如图19所示。
在本公开的可选实施例中，对图像对进行弹性配准包括：对刚性配准结果进行直方图均衡化；将刚性配准结果输入深度学习模型，输出深度学习模型的中间结果，得到形变场，其中，深度学习模型是通过表达量图-图像数据对进行训练得到的；根据形变场，对刚性配准结果进行最近邻插值变换，得到弹性配准结果。
在本公开的一种实施例中，弹性配准的实现基于的是深度学习模型的推导，如20图。对应的深度学习模型要经过如下过程获得：首先将表达量数据处理成为表达量图后，分别对表达量图和对应的BlockFace分割图进行直方图均衡化。然后在软件中对表达量图-BlockFace分割图进行刚性配准，作为模型训练的输入图像，再进行弹性配准，作为模型训练的标签。再基于手动弹性配准时软件输出的控制点坐标，加入随机参数，生成随机变形的表达量图，实现数据合成。最后，使用合成的和真实的数据训练有监督模型，在训练指标达到要求后停止，保存模型。获得深度学习模型之后，就可以使用此模型进行弹性形变的推导。将实际问题中的表达量数据和BlockFace分割图按照上述的处理方法，获得直方图均衡化的图像。将两个图像输入模型，输出模型的中间结果：形变场。根据形变场，获取对应位移，对表达量数据进行最近邻插值，计算得到新的表达量数据。图21所示，以猴脑为例，展示弹性配准前后，表达量图和BlockFace图的对应情况变换：其中红色的为BlockFace图像，绿色的为表达量图像。弹性配准后，可视的表达量图的形变得到了校正。
在本公开中，基于配准后的时空组数据生成生物组织样本的轮廓数据，具体地，基于已经配准好的表达量时空数据，每一个bin set或者cell bin在通过常规的生物信息学分析后，得到聚类和注释的结果。以每个bin set或cell bin的注释或者聚类结果的值为单位，构造三维npy数组，同时增加空间尺度信息，得到基于注释聚类结果的NIfTI文件。再运用类似3D slicer中的segment editor等工具，对每个聚类或分区结果进行分割，生成样本内部结构多个组织或者区域的平滑轮廓面，以vtk、obj、stl等常见的3D矢量模型数据存储。
在本公开中，轮廓数据用于生成生物组织样本的三维展示数据，具体地，轮廓数据整合其他图像渲染数据、配准后时空组数据等组成项目统领文件，其中综合考虑传输效率和渲染效果，生成项目统领文件时对时空组数据按单基因表达量、细胞类型等角度分别进行数据的分层分块处理。将生成好的统领文件，通过基于flask的后端程序加载到渲染器，可以实现对基因表达分布、组织轮廓等信息在三维空间内的情况的直观展示。
可以参照基于图2中步骤S203-208对应内容的具体实施方式，本公开中不再赘述。
综上，根据本公开实施例提供的数据处理方法，获取生物组织样本的表达量数据和图像数据，其中表达量数据是对生物组织样本进行切片并对切片进行时空组测序得到的，图像数据是对切片后的生物组织样本进行图像采集得到的；对表达量数据进行计算，得到表达量图；将表达量图和图像数据进行直方图均衡化，得到表达量图-图像数据的图像对；对图像对依次进行刚性配准和弹性配准，以生成生物组织样本的轮廓数据，轮廓数据用于生成生物组织样本的三维展示数据，提供了时空测序的表达 量数据与组织切面图像数据的配准方法，应用本方法可以得到内外组织轮廓辅助可视化信息定位，对大尺寸高分辨率时空组数据进行有效处理与展示，推动三维时空组学技术的研究与应用。
为了实现上述本申请实施例提供的方法中的各功能，网络设备和终端设备可以包括硬件结构、软件模块，以硬件结构、软件模块、或硬件结构加软件模块的形式来实现上述各功能。上述各功能中的某个功能可以以硬件结构、软件模块、或者硬件结构加软件模块的方式来执行。
与上述几种实施例提供的数据采集方法相对应，本公开还提供一种数据采集装置，由于本公开实施例提供的数据采集装置与上述几种实施例提供的数据采集方法相对应，因此数据采集方法的实施方式也适用于本实施例提供的数据采集装置，在本实施例中不再详细描述。
图32为本公开实施例提供的一种数据采集装置400的结构示意图，如图32所示，该装置400可以包括：测序模块410，用于对生物组织样本进行切片，并对切片进行时空组测序，以得到表达量数据；采集模块420，用于对切片后的生物组织样本进行图像采集，以得到图像数据；配准模块430，用于对表达量数据和/或图像数据进行配准，以生成生物组织样本的外层轮廓和/或分区轮廓；生成模块440，用于根据外层轮廓和/或分区轮廓，生成生物组织样本的三维展示数据。
根据本公开实施例提供的数据采集装置，对生物组织样本进行切片，并对切片进行时空组测序，以得到表达量数据；对切片后的生物组织样本进行图像采集，以得到图像数据；对表达量数据和/或图像数据进行配准，以生成生物组织样本的外层轮廓和/或分区轮廓；根据外层轮廓和/或分区轮廓，生成生物组织样本的三维展示数据，针对原位捕获的时空组学技术，结合光学成像信息，配准生成内外组织轮廓辅助可视化信息定位，可以对大尺寸高分辨率时空组数据进行有效的处理与展示。
在一些实施例中，所述传输配置信息还包括推荐OAM模态数量和/或推荐OAM模态集合，所述推荐OAM模态数量和/或所述推荐OAM模态集合为所述第二设备基于所述推荐PMI得到。
在一些实施例中，测序模块410具体用于：对生物组织样本进行连续切片，每切第一预定数量的切片进行一次时空组测序，以得到切片的表达量数据，其中，第一预定数量为正整数。
在一些实施例中，还包括预处理模块，用于对表达量数据进行预处理，其中，预处理包括聚合处理、质量控制、对数归一化、找高变基因、降维处理中的至少一项。
在一些实施例中，采集模块420具体用于：每切第二预定数量的切片，对切片后的生物组织样本进行一次图像采集，以得到图像数据，其中，第二预定数量为正整数。
在一些实施例中，配准模块430具体用于：对图像数据进行图像分割处理，以生成分割结果图；对分割结果图进行配准，以生成配准结果图；基于配准结果图，生成生物组织样本的外层轮廓。
在一些实施例中，对图像数据进行图像分割处理包括：将图像数据转换为HSV颜色空间，提取H通道的图像数据并转换为灰度图；通过预设图像处理算法，对图像数据进行二值化，以确定二值化结果轮廓图；保留符合预设条件的轮廓图，与图像数据进行合并。
在一些实施例中，还包括：获取图像数据的灰度化图像和直方图；对图像数据进行最大熵阈值二值化处理；将处理后的图像与图像数据进行合并。
在一些实施例中，对分割结果图进行配准，以生成配准结果图包括：对分割结果图进行空间变换处理和线性插值算法处理；对经处理的分割结果图以及配准参考图进行均方差相似度测量；将测量最优解确定为配准结果图。
在一些实施例中，基于配准结果图，生成生物组织样本的外层轮廓包括：将所有配准结果图堆叠为预设格式的数组文件；对数组文件进行处理，以得到生物组织样本的外层轮廓。
在一些实施例中，配准模块430还用于：对表达量数据进行计算，得到表达量图；将表达量图和图像数据进行直方图均衡化，得到表达量图-图像数据的图像对；对图像对依次进行刚性配准和弹性配准，以生成分区轮廓。
在一些实施例中，对图像对进行刚性配准包括：对图像对分别进行下采样；提取图像对的特征信息；对图像对进行相似性估计仿射变换，得到刚性配准结果。
在一些实施例中，对图像对进行弹性配准包括：对刚性配准结果进行直方图均衡化；将刚性配准结果输入深度学习模型，输出深度学习模型的中间结果，得到形变场，其中，深度学习模型是通过表达量图-图像数据对进行训练得到的；根据形变场，对刚性配准结果进行最近邻插值变换，得到弹性配准结果。
在一些实施例中，配准模块430还用于：对两个相邻或相近切片中的表达量数据进行配准，以得到配准表达量数据；根据配准表达量数据，生成生物组织样本的外层轮廓和/或分区轮廓。
在一些实施例中，生成模块440具体用于：对表达量数据和/或图像数据进行配准得到的配准结果进行生物信息学分析，得到聚类和注释结果；基于聚类和注释结果，构造三维数组，并增加空间尺度信息；利用三维渲染工具，渲染得到生物组织样本的三维展示数据。
综上，根据本公开实施例提供的数据采集方法，对生物组织样本进行切片，并对切片进行时空组测序，以得到表达量数据；对切片后的生物组织样本进行图像采集，以得到图像数据；对表达量数据和/或图像数据进行配准，以生成生物组织样本的外层轮廓和/或分区轮廓；根据外层轮廓和/或分区轮廓，生成生物组织样本的三维展示数据，针对原位捕获的时空组学技术，结合光学成像信息，辅助连续组织切片的刚性、弹性配准，实现三维重构，进而得到内外组织轮廓辅助可视化信息定位，对大尺寸高分辨率时空组数据进行有效处理与展示，应用于三维时空组学技术可以实现基因与影像同时分析。
与上述几种实施例提供的数据处理方法相对应，本公开还提供一种数据处理装置，由于本公开实施例提供的数据处理装置与上述几种实施例提供的数据处理方法相对应，因此数据处理方法的实施方式也适用于本实施例提供的数据处理装置，在本实施例中不再详细描述。
图33为本公开实施例提供的一种数据处理装置500的结构示意图。如图33所示，该装置500可以包括：获取模块510，用于获取生物组织样本的表达量数据和图像数据，其中表达量数据是对生物组织样本进行切片并对切片进行时空组测序得到的，图像数据是对切片后的生物组织样本进行图像采集得到的；计算模块520，用于对表达量数据进行计算，得到表达量图；均衡化模块530，用于将表达量图和图像数据进行直方图均衡化，得到表达量图-图像数据的图像对；配准模块540，对图像对依次进行刚性配准和弹性配准，以生成生物组织样本的轮廓数据，轮廓数据用于生成生物组织样本的三维展示数据。
根据本公开实施例提供的数据处理装置，获取生物组织样本的表达量数据和图像数据，其中表达量数据是对生物组织样本进行切片并对切片进行时空组测序得到的，图像数据是对切片后的生物组织样本进行图像采集得到的；对表达量数据进行计算，得到表达量图；将表达量图和图像数据进行直方图均衡化，得到表达量图-图像数据的图像对；对图像对依次进行刚性配准和弹性配准，以生成生物组织样本的轮廓数据，轮廓数据用于生成生物组织样本的三维展示数据，实现时空测序的表达量数据与 组织切面图像数据的配准，可以得到内外组织轮廓辅助可视化信息定位，对大尺寸高分辨率时空组数据进行有效处理与展示，推动三维时空组学技术的研究与应用。
在一些实施例中，配准模块540用于对图像对进行刚性配准包括：对图像对分别进行下采样；提取图像对的特征信息；对图像对进行相似性估计仿射变换，得到刚性配准结果。
在一些实施例中，配准模块540用于对图像对进行弹性配准包括：对刚性配准结果进行直方图均衡化；将刚性配准结果输入深度学习模型，输出深度学习模型的中间结果，得到形变场，其中，深度学习模型是通过表达量图-图像数据对进行训练得到的；根据形变场，对刚性配准结果进行最近邻插值变换，得到弹性配准结果。
综上，根据本公开实施例提供的数据处理方法，获取生物组织样本的表达量数据和图像数据，其中表达量数据是对生物组织样本进行切片并对切片进行时空组测序得到的，图像数据是对切片后的生物组织样本进行图像采集得到的；对表达量数据进行计算，得到表达量图；将表达量图和图像数据进行直方图均衡化，得到表达量图-图像数据的图像对；对图像对依次进行刚性配准和弹性配准，以生成生物组织样本的轮廓数据，轮廓数据用于生成生物组织样本的三维展示数据，提供了时空测序的表达量数据与组织切面图像数据的配准方法，应用本方法可以得到内外组织轮廓辅助可视化信息定位，对大尺寸高分辨率时空组数据进行有效处理与展示，推动三维时空组学技术的研究与应用。
上述本申请提供的实施例中，对本申请实施例提供的方法及装置进行了介绍。为了实现上述本申请实施例提供的方法中的各功能，电子设备可以包括硬件结构、软件模块，以硬件结构、软件模块、或硬件结构加软件模块的形式来实现上述各功能。上述各功能中的某个功能可以以硬件结构、软件模块、或者硬件结构加软件模块的方式来执行。
图34是根据一示例性实施例示出的一种用于实现上述数据采集或数据处理方法的电子设备600的框图。例如，电子设备600可以是移动电话，计算机，消息收发设备，游戏控制台，平板设备，医疗设备，健身设备，个人数字助理等。
参照图34，电子设备600可以包括以下一个或多个组件：处理组件602，存储器604，电源组件606，多媒体组件606，音频组件610，输入/输出(I/O)的接口612，传感器组件614，以及通信组件616。
处理组件602通常控制电子设备600的整体操作，诸如与显示，电话呼叫，数据通信，相机操作和记录操作相关联的操作。处理组件602可以包括一个或多个处理器620来执行指令，以完成上述的方法的全部或部分步骤。此外，处理组件602可以包括一个或多个模块，便于处理组件602和其他组件之间的交互。例如，处理组件602可以包括多媒体模块，以方便多媒体组件608和处理组件602之间的交互。
存储器604被配置为存储各种类型的数据以支持在电子设备600的操作。这些数据的示例包括用于在电子设备600上操作的任何应用程序或方法的指令，联系人数据，电话簿数据，消息，图片，视频等。存储器604可以由任何类型的易失性或非易失性存储设备或者它们的组合实现，如静态随机存取存储器(SRAM)，电可擦除可编程只读存储器(EEPROM)，可擦除可编程只读存储器 (EPROM)，可编程只读存储器(PROM)，只读存储器(ROM)，磁存储器，快闪存储器，磁盘或光盘。
电源组件606为电子设备600的各种组件提供电力。电源组件606可以包括电源管理系统，一个或多个电源，及其他与为电子设备600生成、管理和分配电力相关联的组件。
多媒体组件606包括在电子设备600和用户之间的提供一个输出接口的屏幕。在一些实施例中，屏幕可以包括液晶显示器(LCD)和触摸面板(TP)。如果屏幕包括触摸面板，屏幕可以被实现为触摸屏，以接收来自用户的输入信号。触摸面板包括一个或多个触摸传感器以感测触摸、滑动和触摸面板上的手势。触摸传感器可以不仅感测触摸或滑动动作的边界，而且还检测与触摸或滑动操作相关的持续时间和压力。在一些实施例中，多媒体组件608包括一个前置摄像头和/或后置摄像头。当电子设备600处于操作模式，如拍摄模式或视频模式时，前置摄像头和/或后置摄像头可以接收外部的多媒体数据。每个前置摄像头和后置摄像头可以是一个固定的光学透镜系统或具有焦距和光学变焦能力。
音频组件610被配置为输出和/或输入音频信号。例如，音频组件610包括一个麦克风(MIC)，当电子设备600处于操作模式，如呼叫模式、记录模式和语音识别模式时，麦克风被配置为接收外部音频信号。所接收的音频信号可以被进一步存储在存储器604或经由通信组件616发送。在一些实施例中，音频组件610还包括一个扬声器，用于输出音频信号。
I/O接口612为处理组件602和外围接口模块之间提供接口，上述外围接口模块可以是键盘，点击轮，按钮等。这些按钮可包括但不限于：主页按钮、音量按钮、启动按钮和锁定按钮。
传感器组件614包括一个或多个传感器，用于为电子设备600提供各个方面的状态评估。例如，传感器组件614可以检测到电子设备600的打开/关闭状态，组件的相对定位，例如组件为电子设备600的显示器和小键盘，传感器组件614还可以检测电子设备600或电子设备600一个组件的位置改变，用户与电子设备600接触的存在或不存在，电子设备600方位或加速/减速和电子设备600的温度变化。传感器组件614可以包括接近传感器，被配置用来在没有任何的物理接触时检测附近物体的存在。传感器组件614还可以包括光传感器，如CMOS或CCD图像传感器，用于在成像应用中使用。在一些实施例中，该传感器组件614还可以包括加速度传感器，陀螺仪传感器，磁传感器，压力传感器或温度传感器。
通信组件616被配置为便于电子设备600和其他设备之间有线或无线方式的通信。电子设备600可以接入基于通信标准的无线网络，如WiFi，2G或3G，4G LTE、5G NR(New Radio)或它们的组合。在一个示例性实施例中，通信组件616经由广播信道接收来自外部广播管理系统的广播信号或广播相关信息。在一个示例性实施例中，通信组件616还包括近场通信(NFC)模块，以促进短程通信。例如，在NFC模块可基于射频识别(RFID)技术，红外数据协会(IrDA)技术，超宽带(UWB)技术，蓝牙(BT)技术和其他技术来实现。
在示例性实施例中，电子设备600可以被一个或多个应用专用集成电路(ASIC)、数字信号处理器(DSP)、数字信号处理设备(DSPD)、可编程逻辑器件(PLD)、现场可编程门阵列(FPGA)、控制器、微控制器、微处理器或其他电子元件实现，用于执行上述方法。
在示例性实施例中，还提供了一种包括指令的非临时性计算机可读存储介质，例如包括指令的存储器604，上述指令可由电子设备600的处理器620执行以完成上述方法。例如，非临时性计算机可读存储介质可以是ROM、随机存取存储器(RAM)、CD-ROM、磁带、软盘和光数据存储设备等。
本公开的实施例还提出了一种存储有计算机指令的非瞬时计算机可读存储介质，其中，计算机指令用于使计算机执行本公开上述实施例中描述的数据采集或数据处理方法。
本公开的实施例还提出一种计算机程序产品，包括计算机程序，计算机程序在被处理器执行本公开上述实施例中描述的数据采集或数据处理方法。
本公开的实施例还提出了一种芯片，该芯片包括一个或多个接口电路和一个或多个处理器；接口电路用于从电子设备的存储器接收信号，并向处理器发送信号，信号包括存储器中存储的计算机指令，当处理器执行计算机指令时，使得电子设备执行本公开上述实施例中描述的数据采集或数据处理方法。
需要说明的是，本公开的说明书和权利要求书及上述附图中的术语“第一”、“第二”等是用于区别类似的对象，而不必用于描述特定的顺序或先后次序。应该理解这样使用的数据在适当情况下可以互换，以便这里描述的本公开的实施例能够以除了在这里图示或描述的那些以外的顺序实施。以下示例性实施例中所描述的实施方式并不代表与本公开相一致的所有实施方式。相反，它们仅是与如所附权利要求书中所详述的、本公开的一些方面相一致的装置和方法的例子。
在本说明书的描述中，参考术语“一个实施方式”、“一些实施方式”、“示意性实施方式”、“示例”、“具体示例”或“一些示例”等的描述意指结合实施方式或示例描述的具体特征、结构、材料或者特点包含于本发明的至少一个实施方式或示例中。在本说明书中，对上述术语的示意性表述不一定指的是相同的实施方式或示例。而且，描述的具体特征、结构、材料或者特点可以在任何的一个或多个实施方式或示例中以合适的方式结合。
流程图中或在此以其他方式描述的任何过程或方法描述可以被理解为，表示包括一个或更多个用于实现特定逻辑功能或过程的步骤的可执行指令的代码的模块、片段或部分，并且本发明的优选实施方式的范围包括另外的实现，其中可以不按所示出或讨论的顺序，包括根据所涉及的功能按基本同时的方式或按相反的顺序，来执行功能，这应被本发明的实施例所属技术领域的技术人员所理解。
在流程图中表示或在此以其他方式描述的逻辑和/或步骤，例如，可以被认为是用于实现逻辑功能的可执行指令的定序列表，可以具体实现在任何计算机可读介质中，以供指令执行系统、装置或设备(如基于计算机的系统、包括处理模块的系统或其他可以从指令执行系统、装置或设备取指令并执行指令的系统)使用，或结合这些指令执行系统、装置或设备而使用。就本说明书而言，"计算机可读介质"可以是任何可以包含、存储、通信、传播或传输程序以供指令执行系统、装置或设备或结合这些指令执行系统、装置或设备而使用的装置。计算机可读介质的更具体的示例(非穷尽性列表)包括以下：具有一个或多个布线的电连接部(控制方法)，便携式计算机盘盒(磁装置)，随机存取存储器(RAM)，只读存储器(ROM)，可擦除可编辑只读存储器(EPROM或闪速存储器)，光纤装置，以及便携式光盘只读存储器(CDROM)。另外，计算机可读介质甚至可以是可在其上打印程序的纸或其他合适的介质，因为可以例如通过对纸或其他介质进行光学扫描，接着进行编辑、解译或必要时以其他合适方式进行处理来以电子方式获得程序，然后将其存储在计算机存储器中。
应当理解，本发明的实施方式的各部分可以用硬件、软件、固件或它们的组合来实现。在上述实施方式中，多个步骤或方法可以用存储在存储器中且由合适的指令执行系统执行的软件或固件来实现。例如，如果用硬件来实现，和在另一实施方式中一样，可用本领域公知的下列技术中的任一项或他们的组合来实现：具有用于对数据信号实现逻辑功能的逻辑门电路的离散逻辑电路，具有合适的组合逻辑门电路的专用集成电路，可编程门阵列(PGA)，现场可编程门阵列(FPGA)等。
本技术领域的普通技术人员可以理解实现上述实施例方法携带的全部或部分步骤是可以通过程序来指令相关的硬件完成，程序可以存储于一种计算机可读存储介质中，该程序在执行时，包括方法实施例的步骤之一或其组合。
此外，在本发明的各个实施例中的各功能单元可以集成在一个处理模块中，也可以是各个单元单独物理存在，也可以两个或两个以上单元集成在一个模块中。上述集成的模块既可以采用硬件的形式实现，也可以采用软件功能模块的形式实现。集成的模块如果以软件功能模块的形式实现并作为独立的产品销售或使用时，也可以存储在一个计算机可读取存储介质中。上述提到的存储介质可以是只读存储器，磁盘或光盘等。
尽管上面已经示出和描述了本发明的实施方式，可以理解的是，上述实施方式是示例性的，不能理解为对本发明的限制，本领域的普通技术人员在本发明的范围内可以对上述实施实施进行变化、修改、替换和变型。

Claims (23)

  1. 一种数据采集方法，其特征在于，所述方法包括：
    对生物组织样本进行切片，并对切片进行时空组测序，以得到表达量数据；
    对切片后的生物组织样本进行图像采集，以得到图像数据；
    对所述表达量数据和/或所述图像数据进行配准，以生成所述生物组织样本的外层轮廓和/或分区轮廓；
    根据所述外层轮廓和/或所述分区轮廓，生成所述生物组织样本的三维展示数据。
  2. 根据权利要求1所述的方法，其特征在于，所述对生物组织样本进行切片，并对切片进行时空组测序，以得到表达量数据包括：
    对生物组织样本进行连续切片，每切第一预定数量的切片进行一次时空组测序，以得到所述切片的表达量数据，
    其中，所述第一预定数量为正整数。
  3. 根据权利要求2所述的方法，其特征在于，所述方法还包括：
    对所述表达量数据进行预处理，
    其中，所述预处理包括聚合处理、质量控制、对数归一化、找高变基因、降维处理中的至少一项。
  4. 根据权利要求1至3中任一项所述的方法，其特征在于，所述对切片后的生物组织样本进行图像采集，以得到图像数据包括：
    每切第二预定数量的切片，对切片后的生物组织样本进行一次图像采集，以得到所述图像数据，
    其中，所述第二预定数量为正整数。
  5. 根据权利要求1至4中任一项所述的方法，其特征在于，所述对所述表达量数据和/或所述图像数据进行配准，以生成所述生物组织样本的外层轮廓和/或分区轮廓包括：
    对所述图像数据进行图像分割处理，以生成分割结果图；
    对所述分割结果图进行配准，以生成配准结果图；
    基于所述配准结果图，生成所述生物组织样本的外层轮廓。
  6. 根据权利要求5所述的方法，其特征在于，所述对所述图像数据进行图像分割处理包括：
    将所述图像数据转换为HSV颜色空间，提取H通道的图像数据并转换为灰度图；
    通过预设图像处理算法，对所述图像数据进行二值化，以确定二值化结果轮廓图；
    保留符合预设条件的轮廓图，与所述图像数据进行合并。
  7. 根据权利要求6所述的方法，其特征在于，所述方法还包括：
    获取所述图像数据的灰度化图像和直方图；
    对所述图像数据进行最大熵阈值二值化处理；
    将处理后的图像与所述图像数据进行合并。
  8. 根据权利要求5至7中任一项所述的方法，其特征在于，所述对所述分割结果图进行配准，以生成配准结果图包括：
    对所述分割结果图进行空间变换处理和线性插值算法处理；
    对经处理的分割结果图以及配准参考图进行均方差相似度测量；
    将测量最优解确定为配准结果图。
  9. 根据权利要求8所述的方法，其特征在于，所述基于所述配准结果图，生成所述生物组织样本的外层轮廓包括：
    将所有配准结果图堆叠为预设格式的数组文件；
    对所述数组文件进行处理，以得到所述生物组织样本的外层轮廓。
  10. 根据权利要求5至9中任一项所述的方法，其特征在于，所述对所述表达量数据和/或所述图像数据进行配准，以生成所述生物组织样本的外层轮廓和/或分区轮廓包括：
    对所述表达量数据进行计算，得到表达量图；
    将所述表达量图和所述图像数据进行直方图均衡化，得到表达量图-图像数据的图像对；
    对所述图像对依次进行刚性配准和弹性配准，以生成所述分区轮廓。
  11. 根据权利要求10所述的方法，其特征在于，所述对所述图像对进行刚性配准包括：
    对所述图像对分别进行下采样；
    提取所述图像对的特征信息；
    对所述图像对进行相似性估计仿射变换，得到刚性配准结果。
  12. 根据权利要求11所述的方法，其特征在于，所述对所述图像对进行弹性配准包括：
    对所述刚性配准结果进行直方图均衡化；
    将所述刚性配准结果输入深度学习模型，输出所述深度学习模型的中间结果，得到形变场，其中，所述深度学习模型是通过表达量图-图像数据对进行训练得到的；
    根据所述形变场，对所述刚性配准结果进行最近邻插值变换，得到弹性配准结果。
  13. 根据权利要求5至12中任一项所述的方法，其特征在于，所述对所述表达量数据和/或所述图像数据进行配准，以生成所述生物组织样本的外层轮廓和/或分区轮廓包括：
    对两个相邻或相近切片中的表达量数据进行配准，以得到配准表达量数据；
    根据所述配准表达量数据，生成所述生物组织样本的外层轮廓和/或分区轮廓。
  14. 根据权利要求1至13中任一项所述的方法，其特征在于，所述根据所述外层轮廓和/或所述分区轮廓，生成所述生物组织样本的三维展示数据包括：
    对所述表达量数据和/或所述图像数据进行配准得到的配准结果进行生物信息学分析，得到聚类和注释结果；
    基于所述聚类和注释结果，构造三维数组，并增加空间尺度信息；
    利用三维渲染工具，渲染得到所述生物组织样本的三维展示数据。
  15. 一种数据处理方法，其特征在于，所述方法包括：
    获取生物组织样本的表达量数据和图像数据，其中所述表达量数据是对生物组织样本进行切片并对切片进行时空组测序得到的，所述图像数据是对切片后的生物组织样本进行图像采集得到的；
    对所述表达量数据进行计算，得到表达量图；
    将所述表达量图和图像数据进行直方图均衡化，得到表达量图-图像数据的图像对；
    对所述图像对依次进行刚性配准和弹性配准，以生成所述生物组织样本的轮廓数据，所述轮廓数据用于生成所述生物组织样本的三维展示数据。
  16. 根据权利要求15所述的方法，其特征在于，所述对所述图像对进行刚性配准包括：
    对所述图像对分别进行下采样；
    提取所述图像对的特征信息；
    对所述图像对进行相似性估计仿射变换，得到刚性配准结果。
  17. 根据权利要求16所述的方法，其特征在于，所述对所述图像对进行弹性配准包括：
    对所述刚性配准结果进行直方图均衡化；
    将所述刚性配准结果输入深度学习模型，输出所述深度学习模型的中间结果，得到形变场，其中，所述深度学习模型是通过表达量图-图像数据对进行训练得到的；
    根据所述形变场，对所述刚性配准结果进行最近邻插值变换，得到弹性配准结果。
  18. 一种数据采集装置，其特征在于，所述装置包括：
    测序模块，用于对生物组织样本进行切片，并对切片进行时空组测序，以得到表达量数据；
    采集模块，用于对切片后的生物组织样本进行图像采集，以得到图像数据；
    配准模块，用于对所述表达量数据和/或所述图像数据进行配准，以生成所述生物组织样本的外层轮廓和/或分区轮廓；
    生成模块，用于根据所述外层轮廓和/或所述分区轮廓，生成所述生物组织样本的三维展示数据。
  19. 一种数据处理装置，其特征在于，所述装置包括：
    获取模块，用于获取生物组织样本的表达量数据和图像数据，其中所述表达量数据是对生物组织样本进行切片并对切片进行时空组测序得到的，所述图像数据是对切片后的生物组织样本进行图像采集得到的；
    计算模块，用于对所述表达量数据进行计算，得到表达量图；
    均衡化模块，用于将所述表达量图和图像数据进行直方图均衡化，得到表达量图-图像数据的图像对；
    配准模块，对所述图像对依次进行刚性配准和弹性配准，以生成所述生物组织样本的轮廓数据，所述轮廓数据用于生成所述生物组织样本的三维展示数据。
  20. 一种电子设备，其特征在于，包括：
    至少一个处理器；以及
    与所述至少一个处理器通信连接的存储器；其中，
    所述存储器存储有可被所述至少一个处理器执行的指令，所述指令被所述至少一个处理器执行，以使所述至少一个处理器能够执行权利要求1-17中任一项所述的方法。
  21. 一种存储有计算机指令的非瞬时计算机可读存储介质，其特征在于，所述计算机指令用于使所述计算机执行根据权利要求1-17中任一项所述的方法。
  22. 一种计算机程序产品，其特征在于，包括计算机程序，所述计算机程序在被处理器执行时实 现根据权利要求1-17中任一项所述的方法。
  23. 一种芯片，其特征在于，包括一个或多个接口电路和一个或多个处理器；所述接口电路用于从电子设备的存储器接收信号，并向所述处理器发送所述信号，所述信号包括存储器中存储的计算机指令，当所述处理器执行所述计算机指令时，使得所述电子设备执行权利要求1-17中任一项所述的方法。
PCT/CN2023/096110 2023-05-24 2023-05-24 数据采集方法、装置及系统 Ceased WO2024239285A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN202380098640.8A CN121399661A (zh) 2023-05-24 2023-05-24 数据采集方法、装置及系统
PCT/CN2023/096110 WO2024239285A1 (zh) 2023-05-24 2023-05-24 数据采集方法、装置及系统

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2023/096110 WO2024239285A1 (zh) 2023-05-24 2023-05-24 数据采集方法、装置及系统

Publications (1)

Publication Number Publication Date
WO2024239285A1 true WO2024239285A1 (zh) 2024-11-28

Family

ID=93588792

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/096110 Ceased WO2024239285A1 (zh) 2023-05-24 2023-05-24 数据采集方法、装置及系统

Country Status (2)

Country Link
CN (1) CN121399661A (zh)
WO (1) WO2024239285A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN120198515A (zh) * 2025-02-26 2025-06-24 东莞市丰真检测仪器有限公司 参数生成方法、纺织物色牢度计算方法及系统

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20100150414A1 (en) * 2005-08-29 2010-06-17 Riken Gene expression image constructing method and gene expression image constructing system
US20180137689A1 (en) * 2016-11-11 2018-05-17 Microbrightfield, Inc. Methods and Software For Creating a 3D Image From Images of Multiple Histological Sections and for Mapping Anatomical Information From a Reference Atlas to a Histological Image
CN112581457A (zh) * 2020-12-23 2021-03-30 武汉理工大学 一种基于三维点云的管道内表面检测方法及装置
CN113592925A (zh) * 2021-07-16 2021-11-02 华中科技大学 一种术中超声图像及其轮廓实时配准方法及系统

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20100150414A1 (en) * 2005-08-29 2010-06-17 Riken Gene expression image constructing method and gene expression image constructing system
US20180137689A1 (en) * 2016-11-11 2018-05-17 Microbrightfield, Inc. Methods and Software For Creating a 3D Image From Images of Multiple Histological Sections and for Mapping Anatomical Information From a Reference Atlas to a Histological Image
CN112581457A (zh) * 2020-12-23 2021-03-30 武汉理工大学 一种基于三维点云的管道内表面检测方法及装置
CN113592925A (zh) * 2021-07-16 2021-11-02 华中科技大学 一种术中超声图像及其轮廓实时配准方法及系统

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN120198515A (zh) * 2025-02-26 2025-06-24 东莞市丰真检测仪器有限公司 参数生成方法、纺织物色牢度计算方法及系统

Also Published As

Publication number Publication date
CN121399661A (zh) 2026-01-23

Similar Documents

Publication Publication Date Title
US11395715B2 (en) Methods and systems for generating and using 3D images in surgical settings
CN110569854B (zh) 图像处理方法及装置、电子设备和存储介质
CN108765278B (zh) 一种图像处理方法、移动终端及计算机可读存储介质
US20190102878A1 (en) Method and apparatus for analyzing medical image
US8953024B2 (en) 3D scene model from collection of images
KR102309995B1 (ko) 이미지 아틀라스 시스템 및 방법
CN103201769A (zh) 图像处理装置、图像处理方法、程序、集成电路
CN111209916B (zh) 病灶识别方法及系统、识别设备
CN112767329A (zh) 图像处理方法及装置、电子设备
CN110211086B (zh) 图像分割方法、装置及存储介质
CN105787872B (zh) 图像处理装置、图像处理装置的控制方法
WO2021259391A2 (zh) 图像处理方法及装置、电子设备和存储介质
US20210125334A1 (en) System and Method for Analyzing Three-Dimensional Image Data
JP6445000B2 (ja) Ｍｄｘのための画像ベースのｒｏｉ追跡
US20160148343A1 (en) Information Processing Method and Electronic Device
CN112927239A (zh) 图像处理方法、装置、电子设备及存储介质
WO2021097595A1 (zh) 图像的病变区域分割方法、装置及服务器
CN109598250B (zh) 特征提取方法、装置、电子设备和计算机可读介质
CN111402177A (zh) 一种清晰度检测方法、系统、设备和介质
CN114926479A (zh) 图像处理方法及装置
WO2024239285A1 (zh) 数据采集方法、装置及系统
CN117373603A (zh) 影像报告生成方法、装置、设备、存储介质和程序产品
CN108805838A (zh) 一种图像处理方法、移动终端及计算机可读存储介质
CN110910348A (zh) 一种肺结节的位置分类方法、装置、设备及存储介质
CN115410028B (zh) 特征组合策略选择、状态检测方法及装置、电子设备

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23937959

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE