[go: up one dir, main page]
More Web Proxy on the site http://driver.im/

WO2022237746A1 - 低密度脂蛋白试剂浓度的确定方法、装置及存储介质 - Google Patents

低密度脂蛋白试剂浓度的确定方法、装置及存储介质 Download PDF

Info

Publication number
WO2022237746A1
WO2022237746A1 PCT/CN2022/091834 CN2022091834W WO2022237746A1 WO 2022237746 A1 WO2022237746 A1 WO 2022237746A1 CN 2022091834 W CN2022091834 W CN 2022091834W WO 2022237746 A1 WO2022237746 A1 WO 2022237746A1
Authority
WO
WIPO (PCT)
Prior art keywords
classification
node
concentration
image
value
Prior art date
Application number
PCT/CN2022/091834
Other languages
English (en)
French (fr)
Inventor
张心觉
高军晖
刘亚
张明
楼敬伟
Original Assignee
上海宝藤生物医药科技股份有限公司
上海宝藤医学检验所有限公司
上海张江医学创新研究院
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 上海宝藤生物医药科技股份有限公司, 上海宝藤医学检验所有限公司, 上海张江医学创新研究院 filed Critical 上海宝藤生物医药科技股份有限公司
Publication of WO2022237746A1 publication Critical patent/WO2022237746A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0012—Biomedical image inspection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/004—Artificial life, i.e. computing arrangements simulating life
    • G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/11—Region-based segmentation

Definitions

  • the present application relates to the technical field of medical reagent detection, for example, to a method, device and storage medium for determining the concentration of a low-density lipoprotein reagent.
  • the concentration of low-density lipoprotein reagents can represent a person's physical health level. With the improvement of medical level, there are more and more demands for the concentration detection of low-density lipoprotein reagents. However, when it is necessary to simultaneously detect the concentration of a large number of low-density lipoprotein reagents, the detection efficiency of the detection method is low.
  • the present application provides a method, device and storage medium for determining the concentration of a low-density lipoprotein reagent, which can improve the detection efficiency of the low-density lipoprotein reagent concentration.
  • the present application provides a method for determining the concentration of a low-density lipoprotein reagent, including: acquiring a target image of the low-density lipoprotein reagent; Set the gray threshold value of each classification node in the classification regression tree in the random forest model, determine the target classification node associated with the target image among the multiple classification nodes, and determine the low density Concentration of lipoprotein reagent.
  • the present application provides a device for determining the concentration of a low-density lipoprotein reagent, including an acquisition module and a determination module;
  • Obtain module be set to obtain the target image of low-density lipoprotein reagent
  • the determination module is set to use the preset random forest model, based on the gray value of the pixel in the target image acquired by the acquisition module and the gray threshold of each classification node in the multiple classification nodes of the classification regression tree in the preset random forest model , determining a target classification node associated with the target image among the plurality of classification nodes, and determining the concentration of the low-density lipoprotein reagent according to the concentration value of the target classification node.
  • the application provides a device for determining the concentration of a low-density lipoprotein reagent, including a memory, a processor, a bus, and a communication interface; the memory is configured to store computer-executed instructions, and the processor and the memory are connected through the bus; when the low-density lipoprotein When the device for determining the concentration of the protein reagent is running, the processor is configured to execute the computer-executed instructions stored in the memory, so that the device for determining the concentration of the low-density lipoprotein reagent executes the method for determining the concentration of the low-density lipoprotein reagent provided in the first aspect.
  • the device for determining the concentration of the low-density lipoprotein reagent may also include a transceiver, which is configured to perform sending and receiving of data, signaling or Informational steps, for example, acquire target images of LDL reagents.
  • the device for determining the concentration of the low-density lipoprotein reagent can be a physical machine configured to realize the determination of the low-density lipoprotein reagent concentration, or it can be a part of the physical machine, for example, it can be a chip in the physical machine system.
  • the system on a chip is set to support the device for determining the concentration of the low-density lipoprotein reagent to realize the functions involved in the first aspect, for example, receiving, sending or processing the data and/or involved in the method for determining the concentration of the low-density lipoprotein reagent above. information.
  • the chip system includes a chip, and may also include other discrete devices or circuit structures.
  • the present application provides a computer-readable storage medium, in which instructions are stored, and when the computer executes the instructions, the computer executes the method for determining the concentration of the low-density lipoprotein reagent provided in the first aspect.
  • the present application provides a computer program product, the computer program product includes computer instructions, and when the computer instructions are run on the computer, the computer executes the method for determining the concentration of the low-density lipoprotein reagent provided in the first aspect.
  • All or part of the above computer instructions may be stored on a computer-readable storage medium.
  • the computer-readable storage medium can be packaged together with the processor of the device for determining the concentration of the low-density lipoprotein reagent, or can be packaged separately with the processor of the device for determining the concentration of the low-density lipoprotein reagent. limited.
  • the names of the above-mentioned devices for determining the concentration of low-density lipoprotein reagents do not limit the devices or functional modules themselves. In actual implementation, these devices or functional modules may appear with other names. As long as the function of each device or functional module is similar to that of the device or functional module in this application.
  • Fig. 1 is a schematic flow chart of a method for determining the concentration of a low-density lipoprotein reagent provided in an embodiment of the present application;
  • FIG. 2 is a schematic diagram of an original scanned image provided in an embodiment of the present application.
  • Fig. 3 is a schematic diagram of an area image provided by an embodiment of the present application.
  • FIG. 4 is a schematic diagram of an image obtained by cropping a region image according to a first boundary line and a second boundary line provided in an embodiment of the present application;
  • FIG. 5 is a schematic diagram of a plurality of target images provided by an embodiment of the present application.
  • Figure 6 is a schematic flow chart of another method for determining the concentration of a low-density lipoprotein reagent provided in an embodiment of the present application
  • Figure 7 is a schematic flow chart of another method for determining the concentration of a low-density lipoprotein reagent provided in the embodiment of the present application.
  • FIG. 8 is a schematic structural diagram of a device for determining the concentration of a low-density lipoprotein reagent provided in an embodiment of the present application
  • Fig. 9 is a schematic structural diagram of another device for determining the concentration of a low-density lipoprotein reagent provided in an embodiment of the present application.
  • first and second in this application are used to distinguish different objects, or to distinguish different processes for the same object, rather than to describe a specific order of objects.
  • words such as “exemplary” or “for example” are used as examples, illustrations or descriptions. Any embodiment or design scheme described as “exemplary” or “for example” in the embodiments of the present application shall not be interpreted as being more preferred or more advantageous than other embodiments or design schemes. Rather, the use of words such as “exemplary” or “such as” is intended to present relevant concepts in a manner.
  • the concentration of low-density lipoprotein reagents can represent a person's physical health level. With the improvement of medical level, there are more and more demands for the concentration detection of low-density lipoprotein reagents. However, when the concentration detection of a large number of low-density lipoprotein reagents needs to be carried out at the same time, the detection efficiency of the detection method is low. Therefore, it is urgent to propose a method for determining the concentration of the low-density lipoprotein reagent, so as to improve the detection efficiency of the low-density lipoprotein reagent concentration.
  • the embodiments of the present application provide a method, device and storage medium for determining the concentration of low-density lipoprotein reagents.
  • This scheme can input the target image to be detected into the pre-trained random forest model, based on the gray value of the pixel in the target image and each of the multiple classification nodes of the classification regression tree in the preset random forest model
  • the grayscale threshold determines the concentration of the LDL reagent.
  • the manual detection of the concentration of the low-density lipoprotein reagent is replaced by a preset random forest model, so the detection efficiency of the concentration of the low-density lipoprotein reagent can be improved.
  • the method for determining the concentration of low-density lipoprotein reagents provided in this application can be applied to a device for determining the concentration of low-density lipoprotein reagents.
  • the device for determining the concentration of the low-density lipoprotein reagent may be a server configured to detect the concentration of the low-density lipoprotein reagent.
  • the server may be one server, or may be a server cluster composed of multiple servers, which is not limited in this embodiment of the present application.
  • the method for determining the concentration of the low-density lipoprotein reagent provided in the embodiment of the present application includes S101-S102:
  • the device for determining the concentration of the low-density lipoprotein reagent acquires a target image of the low-density lipoprotein reagent.
  • the target image may be a preprocessed image.
  • the device for determining the concentration of the low-density lipoprotein reagent may first obtain the original scanned image, and then cut the original scanned image to obtain at least one regional image, and then process each regional image, Get the target image for each region image.
  • the original scanned image includes an area where at least one reagent container is located, and the reagent container is set to hold the low-density lipoprotein reagent.
  • Each of the at least one area image corresponds to a reagent container.
  • Fig. 2 provides a kind of raw scanning image, as shown in Fig. 2, comprise a plurality of reagent containers in the raw scanning image, the determining device of low-density lipoprotein reagent concentration can be compared to Fig. 2 raw
  • the scanned image is cut to obtain a plurality of regional images as shown in FIG. 3 , and then the regional images are processed to obtain a target image of the regional images.
  • the abscissa of the area image in FIG. 3 is the width of the reagent container, and the ordinate represents the length of the reagent container, and the unit is mm.
  • the original scanned image is generally a tag image file format (Tag Image File Format, TIFF) file
  • the image processing software generally does not support processing the image of the TIFF file, so the device for determining the concentration of the low-density lipoprotein reagent is in
  • TIFF Tag Image File Format
  • the device for determining the concentration of the low-density lipoprotein reagent is in
  • the original scanned image is obtained, it can be converted into a format such as Portable Network Graphics (PNG) or Joint Photographic Experts Group (JPEG) supported by image processing software.
  • PNG Portable Network Graphics
  • JPEG Joint Photographic Experts Group
  • some reagent containers in the original scanned image may not hold reagents. It can be seen from FIG. The gray values of multiple pixels of the reagent container holding the reagent are distributed regularly. Therefore, optionally, when the device for determining the concentration of the low-density lipoprotein reagent cuts the original scanned image, it may cut out the area where the reagent container that does not contain the reagent in the original scanned image is located based on the gray value of the pixel.
  • the original scanned image may also include the test tube rack containing the reagent container (for example, the black area at the bottom of the reagent container in FIG. 2 ), and this part may be cropped based on the gray values of multiple pixels.
  • the test tube rack containing the reagent container for example, the black area at the bottom of the reagent container in FIG. 2 .
  • the device for determining the concentration of the low-density lipoprotein reagent can determine the first demarcation line and the second demarcation line of each region image; then based on the first demarcation line and the second demarcation line, clipping the region image, and then filling the cropped region image based on the second boundary line, so as to fill the cropped region image into a region image of a preset size, and determine the filled region image of a preset size as target image.
  • the first dividing line corresponds to the liquid level of the low-density lipoprotein reagent
  • the second dividing line corresponds to the layered level of the low-density lipoprotein reagent.
  • the preset size may be an artificially determined size in advance, for example, the preset size may be 20mm*300mm. In practical applications, the preset size may also be other sizes, which are not limited in this embodiment of the present application.
  • FIG. 4 is an image obtained by cropping an area image according to the first boundary line and the second boundary line.
  • the cropped area image is filled at the end of the second boundary line, and the filled area image with a preset size is determined as the target image.
  • FIG. 5 is a target image obtained by filling the cropped region image at the tail of the second boundary line.
  • the target image may also be obtained in other ways, which is not limited in this embodiment of the present application.
  • the device for determining the concentration of the low-density lipoprotein reagent uses the preset random forest model, based on the gray value of the pixel in the target image and the gray value of each classification node in the multiple classification nodes of the classification regression tree in the preset random forest model A degree threshold is used to determine the target classification node associated with the target image among the plurality of classification nodes, and to determine the concentration of the low-density lipoprotein reagent according to the concentration value of the target classification node.
  • the preset random forest model can be determined in the following manner: based on the training set, the test set and the fourth preset value N, train N classification and regression trees, and obtain Preset random forest model. After determining the preset random forest model, in the N classification and regression trees of the preset random forest model, respectively determine the N target classification nodes associated with the target image, and then the average value of the concentration values of the N target classification nodes can be calculated. Determine the concentration of the LDL reagent.
  • the fourth preset value N may be the number of classification and regression trees in the preset random forest model determined artificially in advance.
  • N may be 5.
  • N may also be other values, which are not discussed in this embodiment of the present application. Do limited.
  • the training set and the test set include sample images and the concentrations corresponding to the sample images.
  • the sample images in the training set are used to train the preset random forest model, and the sample images in the test set are used to adjust the gray threshold in the preset random forest model. Excellent, optimize the preset random forest model to improve the accuracy of detection.
  • the sample image is also a pre-processed image with a preset size, and the processing process is the same as that of the target image, which will not be repeated in this embodiment of the present application.
  • the classification regression tree may be determined in the following manner: first, the device for determining the concentration of the low-density lipoprotein reagent randomly selects the sample image of the first preset value from the training set, and the The sample images of the first preset value are all determined as the sample images of the root node in the classification nodes of the classification regression tree; then, step A is performed: based on the gray threshold of the root node, all the sample images of the root node are divided into the root node two sub-nodes of ; execute step B: determine whether each sub-node of the root node satisfies the termination condition; execute step C: obtain the concentration value of the sub-node of the root node when the sub-node of the root node satisfies the termination condition; Execute step D: in the case that the child node of the root node does not meet the termination condition, randomly select the sample image of the third preset value from all sample images of the child node of the root node without
  • the training set may include sample images of a second preset value and density values corresponding to the sample images of the second preset value.
  • the first preset value, the second preset value and the third preset value may be values determined artificially in advance, the first preset value is less than or equal to the second preset value, and the third preset value is less than or equal to the first Default value.
  • the termination condition can be determined according to the first preset value and the depth of the classification and regression tree. Exemplarily, if the depth of the classification and regression tree is 5, the termination condition may be to determine whether the depth of the child node of the root node is 5, and if the depth of the child node of the root node does not reach 5, from the root node Select the sample image of the third preset value from all the sample images of the child node, use the child node as the new root node, and use the sample image of the third preset value as the sample image of the new root node, and then for the new The root node re-executes step A and step B until the depth of the child nodes of the root node reaches 5.
  • the device for determining the concentration of the low-density lipoprotein reagent can determine the number of sample images divided into child nodes according to the first preset value and the gray threshold value. The number of sample images of the node is 1, that is, when the division cannot be continued, it is determined that the termination condition is met.
  • the density value of the classification node is determined according to the density value of the sample image of the node.
  • the concentration values of the child nodes of the root node are determined according to the concentration values of the sample images of the child nodes of the root node.
  • the concentration value of the child nodes of the root node may be the average value of the concentration values of all sample images of the child nodes of the root node.
  • the grayscale threshold may be determined in the following manner:
  • the device for determining the concentration of the low-density lipoprotein reagent traverses the gray values of the pixels of all sample images of the classification node, and uses M preset gray values as initial thresholds to classify all sample images of the classification node to obtain M classification results; one classification result corresponds to two classification sub-nodes, and M is a positive integer; determine the concentration value of each classification sub-node in the two classification sub-nodes corresponding to each classification result according to the concentration values of all sample images of the classification node , according to the concentration values corresponding to the two classification sub-nodes, determine the classification error of the classification result; then determine the preset gray value corresponding to the classification result whose classification error meets the preset condition among the M classification results as the classification The grayscale threshold of the node.
  • the device for determining the concentration of the low-density lipoprotein reagent may calculate the first average concentration of the sample images of the first node and the second average concentration of the sample images of the second node according to the concentration values of all sample images of the classification node.
  • the first node and the second node are two classification sub-nodes corresponding to the classification results obtained by classifying all sample images of the classification node with the target preset gray value as the initial threshold value; the target preset gray value is Any one of M preset grayscale values.
  • the device for determining the concentration of the low-density lipoprotein reagent determines the first mean error of the concentration value of the sample image of the first node according to the first concentration mean value, and determines the second error of the concentration value of the sample image of the second node according to the second concentration mean value. mean error; then, according to the first mean error and the second mean error, determine the classification error of the classification result with the target preset gray value as the initial threshold.
  • the M preset grayscale values may be grayscale values of the grayscale image, which are M values from 0-255, respectively.
  • the preset condition may be: among the M classification results, the preset gray value corresponding to the classification result with the smallest classification error is determined as the gray threshold.
  • the first average concentration of the sample image of the first node may be the average concentration value of all sample images of the first node
  • the second average concentration value of the sample image of the second node may be the concentration value of all sample images of the second node average of.
  • the classification error of the classification result with the target preset gray value as the initial threshold may be the sum of the first mean error and the second mean error.
  • the gray value of the pixel can be less than or equal to s
  • the expression (1) can be obtained:
  • R 1 (j,s) ⁇ x
  • R 2 (j,s) ⁇ x
  • y i represents the actual density value of the sample image i.
  • the classification result corresponding to the smallest classification error can be determined by expression (3).
  • the preset grayscale value in the division point (j, s) corresponding to the classification result with the smallest classification error may be determined as the grayscale threshold.
  • the device for determining the concentration of the low-density lipoprotein reagent can divide the sample image of the root node of the classification regression tree into nodes, continuously divide the node into two sub-nodes, and divide all the sample images of the root node into T region (R 1 , R 2 ... R T ), generate a classification and regression tree, and the model of the classification and regression tree can be expressed by expression (4):
  • I is an exponential function, which can be expressed by expression (5):
  • the method for determining the grayscale threshold and the method for determining the classification and regression tree provided in the embodiment of the present application are only examples. In practical applications, based on the grayscale value of the pixel in the sample image and the actual concentration of the sample image, other methods can be used. Determine the gray threshold and classification regression tree, which is not limited in this embodiment of the present application.
  • the concentration of the low-density lipoprotein reagent is related to the gray value of the pixel in the image containing the low-density lipoprotein reagent, it can be based on the sample image
  • the gray value of the pixels in the image and the concentration corresponding to the sample image are used to train the classification regression tree in the preset random forest model.
  • the target image to be detected can be input into the pre-trained random forest model, based on the gray value of the pixel in the target image and each of the multiple classification nodes of the classification regression tree in the preset random forest model
  • the grayscale threshold of the node determines the concentration of the LDL reagent.
  • the preset random forest model can replace the manual detection of the concentration of the low-density lipoprotein reagent. detection efficiency.
  • the embodiment of the present application also provides a training method for a preset random forest model, including S701-S7010:
  • the device for determining the concentration of the low-density lipoprotein reagent selects sample images of the first preset value from the training set, and determines the sample images of the first preset value as the sample images of the root node in the classification nodes of the classification regression tree.
  • the device for determining the concentration of the low-density lipoprotein reagent traverses the grayscale values of the pixels of all the sample images of the root node, and classifies all the sample images of the root node with M preset grayscale values as initial thresholds respectively, and obtains M classification results.
  • the device for determining the concentration of the low-density lipoprotein reagent determines the concentration value of each of the two sub-nodes corresponding to each classification result according to the concentration values of all sample images of the root node, and determines according to the concentration values corresponding to the two sub-nodes respectively The classification error of this classification result.
  • the device for determining the concentration of the low-density lipoprotein reagent determines, among the M classification results, a preset gray value corresponding to a classification result whose classification error satisfies a preset condition as the gray threshold of the root node.
  • the device for determining the concentration of the low-density lipoprotein reagent divides all sample images of the root node into two child nodes of the root node based on the gray threshold of the root node.
  • the device for determining the concentration of the low-density lipoprotein reagent determines whether each child node of the root node satisfies the termination condition.
  • step S707 the device for determining the low-density lipoprotein reagent concentration executes step S707; when the child node of the root node does not satisfy the termination condition, executes step S708.
  • the device for determining the concentration of the low-density lipoprotein reagent acquires the concentration value of the child node of the root node.
  • step S709 is executed.
  • the device for determining the concentration of the low-density lipoprotein reagent selects the sample image of the third preset value from all the sample images of the sub-node of the root node, uses the sub-node as a new root node, and sets the third preset value
  • the sample images are all used as the sample images of the new root node.
  • step S708 return to re-execute step S702.
  • the device for determining the concentration of the low-density lipoprotein reagent trains a classification and regression tree based on the gray thresholds of the root node, all child nodes, and the root node obtained during model training.
  • the device for determining the concentration of the low-density lipoprotein reagent trains N classification and regression trees based on the training set, the test set and the fourth preset value N, and obtains a preset random forest model according to the N classification and regression trees.
  • S101 in Figure 1 can be replaced with S1011-S1014:
  • the device for determining the concentration of the low-density lipoprotein reagent acquires an original scan image, and cuts the original scan image to obtain at least one region image.
  • the device for determining the concentration of the low-density lipoprotein reagent determines the first boundary line and the second boundary line of each region image.
  • the device for determining the concentration of the low-density lipoprotein reagent cuts the region image based on the first boundary line and the second boundary line, and fills the cropped region image based on the second boundary line, so that the cropped region The image is filled with an area image of a preset size.
  • the device for determining the concentration of the low-density lipoprotein reagent determines the filled area image with a preset size as the target image.
  • the embodiment of the present application also provides a device for determining the concentration of a low-density lipoprotein reagent, which may include: an acquisition module 11 and a determination module 12 .
  • the acquisition module 11 executes S101 in the above method embodiment, and the determination module 12 executes S102 in the above method embodiment.
  • the acquisition module 11 is configured to acquire the target image of the low-density lipoprotein reagent; the determination module 12 is configured to use the preset random forest model, based on the gray value of the pixel in the target image acquired by the acquisition module 11 and the preset random forest model
  • the gray threshold value of each classification node in the multiple classification nodes of the classification regression tree determines the target classification node associated with the target image in the multiple classification nodes, and determines the concentration of the low-density lipoprotein reagent according to the concentration value of the target classification node. concentration.
  • the device for determining the concentration of the low-density lipoprotein reagent may also include a classification module; the classification module is set to traverse the gray value of the pixels of all sample images of the classification node, Classify all sample images of the classification node with M preset gray values as initial thresholds respectively, and obtain M classification results; one classification result corresponds to two classification sub-nodes, and M is a positive integer; the determination module 12 is also set to Determine the concentration value of each classification sub-node in the two classification sub-nodes corresponding to each classification result according to the concentration values of all sample images of the classification node, and determine the concentration value of the classification result according to the concentration values corresponding to the two classification sub-nodes respectively Classification error; the determination module 12 is further configured to determine, among the M classification results, the preset gray value corresponding to the classification result whose classification error satisfies the preset condition as the gray threshold of the classification node.
  • the determination module 12 is set to:
  • the concentration values of all sample images of the classification node calculate the first average concentration value of the sample image of the first node and the second average concentration value of the sample image of the second node; the first node and the second node are, with the target preset gray
  • the degree value is used as the initial threshold to classify all sample images of the classification node, and the obtained classification results correspond to two classification sub-nodes; the target preset gray value is any one of the M preset gray values; according to the first Determine the first mean error of the concentration value of the sample image of the first node by the density mean value, and determine the second mean value error of the density value of the sample image of the second node according to the second density mean value; according to the first mean value error and the second mean value error, Determine the classification error of the classification result with the target preset gray value as the initial threshold.
  • the determination module 12 is also set to:
  • step A based on the gray threshold of the root node, divide all the sample images of the root node into the root node two sub-nodes of the root node; step B: determine whether each sub-node of the root node satisfies the termination condition; the termination condition is determined according to the first preset value and the depth of the classification regression tree; step C: when the sub-node of the root node satisfies the termination condition In this case, obtain the concentration value of the child node of the root node; the concentration value of the child node of the root node is determined according to the concentration values of all sample images of the child node of the root
  • the acquisition module 11 is set to:
  • the original scanned image includes at least one area where the reagent container is located; the reagent container is set to hold the low-density lipoprotein reagent; cutting the original scanned image to obtain at least one area image; one area image corresponds to one reagent container;
  • the determining module 12 is further configured to: process each area image to obtain a target image of the area image.
  • the determination module 12 is set to:
  • the first demarcation line corresponds to the liquid level of the low-density lipoprotein reagent
  • the second demarcation line corresponds to the stratification level of the low-density lipoprotein reagent
  • the second dividing line is used to crop the area image; the cropped area image is filled based on the second dividing line, so that the cropped area image is filled to the size of the area image of the preset size; the filled preset size is The area image is determined as the target image.
  • the device for determining the concentration of the low-density lipoprotein reagent provided by the present application may also include a training module; the training module is set to be based on the training set, the test set and the fourth preset value N, N classification and regression trees are trained, and a preset random forest model is obtained according to the N classification and regression trees; the determination module 12 is set to: in the N classification and regression trees, respectively determine the N target classification nodes associated with the target image, and divide the N target The average of the concentration values of the classification nodes was determined as the concentration of the LDL reagent.
  • a training module is set to be based on the training set, the test set and the fourth preset value N, N classification and regression trees are trained, and a preset random forest model is obtained according to the N classification and regression trees
  • the determination module 12 is set to: in the N classification and regression trees, respectively determine the N target classification nodes associated with the target image, and divide the N target The average of the concentration values of the classification nodes was determined as the concentration of the L
  • the device for determining the concentration of the low-density lipoprotein reagent may also include a storage module configured to store the program code of the device for determining the concentration of the low-density lipoprotein reagent.
  • the embodiment of the present application also provides a device for determining the concentration of low-density lipoprotein reagents, including memory 41, processor 42 (42-1 and 42-2), bus 43 and communication interface 44; memory 41 Set to store computer-executed instructions, the processor 42 is connected to the memory 41 by a bus 43; when the device for determining the concentration of the low-density lipoprotein reagent runs, the processor 42 executes the computer-executed instructions stored in the memory 41, so that the low-density lipoprotein reagent
  • the device for determining the concentration implements the method for determining the concentration of the low-density lipoprotein reagent provided in the above-mentioned embodiments.
  • the processors 42 (42-1 and 42-2) may include one or more central processing units (Central Processing Unit, CPU), such as CPU0 and CPU1 shown in FIG. 9 .
  • the device for determining the concentration of the low-density lipoprotein reagent may include multiple processors 42, such as the processor 42-1 and the processor 42-2 shown in FIG. 9 .
  • Each CPU in these processors 42 may be a single-core processor (single-CPU), or a multi-core processor (multi-CPU).
  • Processor 42 herein may refer to one or more devices, circuits, and/or processing cores configured to process data (eg, computer program instructions).
  • Memory 41 can be read-only memory 41 (Read-Only Memory, ROM) or other types of static storage devices that can store static information and instructions, random access memory (Random Access Memory, RAM) or other types that can store information and instructions
  • Type of dynamic storage device also can be Electrically Erasable Programmable Read-Only Memory (EEPROM), CD-ROM (Compact Disc Read-Only Memory, CD-ROM) or other CD-ROM storage, CD-ROM storage (including compact discs, laser discs, optical discs, digital versatile discs, blu-ray discs, etc.), magnetic disk storage media, or other magnetic storage devices, or can be arranged to carry or store desired program code in the form of instructions or data structures and can be programmed by Any other medium accessed by a computer, but not limited to.
  • the memory 41 may exist independently, and is connected to the processor 42 through the bus 43 .
  • the memory 41 can also be integrated with the processor 42 .
  • the memory 41 is set to store the data in this application and the computer-executed instructions corresponding to executing the software program of this application.
  • the processor 42 can run or execute the software program stored in the memory 41 and call the data stored in the memory 41 to perform multiple functions of the device for determining the concentration of the low-density lipoprotein reagent.
  • Communication interface 44 using any device such as a transceiver, configured to communicate with other devices or communication networks, such as control systems, radio access networks (Radio Access Network, RAN), wireless local area networks (Wireless Local Area Networks, WLAN), etc. .
  • the communication interface 44 may include a receiving unit to implement a receiving function, and a sending unit to implement a sending function.
  • the bus 43 may be an Industry Standard Architecture (Industry Standard Architecture, ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (Extended Industry Standard Architecture, EISA) bus, etc.
  • ISA Industry Standard Architecture
  • PCI Peripheral Component Interconnect
  • EISA Extended Industry Standard Architecture
  • the bus 43 can be divided into an address bus, a data bus, a control bus, and the like. For ease of representation, only one thick line is used in FIG. 9 , but it does not mean that there is only one bus or one type of bus.
  • the function realized by the acquisition module 11 in the determining device of low-density lipoprotein reagent concentration is the same as the function realized by the receiving unit in Fig. 9, the determining module in the determining device of low-density lipoprotein reagent concentration
  • the functions realized by 12 are the same as those realized by the processor in FIG. 9
  • the functions realized by the storage module in the device for determining the concentration of the low-density lipoprotein reagent are the same as those realized by the memory in FIG. 9 .
  • the embodiment of the present application also provides a computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the computer executes the instructions, the computer executes the method for determining the concentration of the low-density lipoprotein reagent provided in the above-mentioned embodiments.
  • the computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or any combination thereof.
  • Examples (non-exhaustive list) of computer readable storage media include: electrical connection with one or more conductors, portable computer disk, hard disk, RAM, ROM, Erasable Programmable Read Only Memory (Erasable Programmable Read Only Memory) , EPROM), registers, hard disks, optical fibers, CD-ROMs, optical storage devices, magnetic storage devices, or any suitable combination of the above, or any other form of computer-readable storage medium known in the art.
  • An exemplary storage medium is coupled to the processor such the processor can read information from, and write information to, the storage medium.
  • the storage medium may also be a component of the processor.
  • the processor and the storage medium may be located in an Application Specific Integrated Circuit (ASIC).
  • ASIC Application Specific Integrated Circuit
  • a computer-readable storage medium may be any tangible medium containing or storing a program, and the program may be used by or in combination with an instruction execution system, device or device.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • General Health & Medical Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Biomedical Technology (AREA)
  • Computing Systems (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Molecular Biology (AREA)
  • Evolutionary Biology (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Medical Informatics (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Radiology & Medical Imaging (AREA)
  • Quality & Reliability (AREA)
  • Investigating Or Analysing Biological Materials (AREA)
  • Image Analysis (AREA)

Abstract

本文公开了一种低密度脂蛋白试剂浓度的确定方法、装置及存储介质。该低密度脂蛋白试剂浓度的确定方法包括：获取低密度脂蛋白试剂的目标图像；通过预设随机森林模型，基于目标图像中像素点的灰度值和预设随机森林模型中分类回归树的多个分类节点中每个分类节点的灰度阈值，确定多个分类节点中与目标图像关联的目标分类节点，且根据目标分类节点的浓度值，确定低密度脂蛋白试剂的浓度。

Description

低密度脂蛋白试剂浓度的确定方法、装置及存储介质
本申请要求在2021年05月11日提交中国专利局、申请号为202110510569.2的中国专利申请的优先权，该申请的全部内容通过引用结合在本申请中。
技术领域
本申请涉及医药试剂检测技术领域，例如涉及一种低密度脂蛋白试剂浓度的确定方法、装置及存储介质。
背景技术
低密度脂蛋白试剂的浓度可以表征人的身体健康水平，随着医疗水平的提高，对于低密度脂蛋白试剂的浓度检测的需求也越来越多。然而，在需要同时对大量低密度脂蛋白试剂进行浓度检测时，检测方法检测效率较低。
发明内容
本申请提供一种低密度脂蛋白试剂浓度的确定方法、装置及存储介质，可以提高低密度脂蛋白试剂浓度的检测效率。
第一方面，本申请提供一种低密度脂蛋白试剂浓度的确定方法，包括：获取低密度脂蛋白试剂的目标图像；通过预设随机森林模型，基于目标图像中像素点的灰度值和预设随机森林模型中分类回归树的多个分类节点中每个分类节点的灰度阈值，确定多个分类节点中与目标图像关联的目标分类节点，且根据目标分类节点的浓度值，确定低密度脂蛋白试剂的浓度。
第二方面，本申请提供一种低密度脂蛋白试剂浓度的确定装置，包括获取模块和确定模块；
获取模块，设置为获取低密度脂蛋白试剂的目标图像；
确定模块，设置为通过预设随机森林模型，基于获取模块获取的目标图像中像素点的灰度值和预设随机森林模型中分类回归树的多个分类节点中每个分类节点的灰度阈值，确定多个分类节点中与目标图像关联的目标分类节点，且根据目标分类节点的浓度值，确定低密度脂蛋白试剂的浓度。
第三方面，本申请提供一种低密度脂蛋白试剂浓度的确定装置，包括存储器、处理器、总线和通信接口；存储器设置为存储计算机执行指令，处理器与存储器通过总线连接；当低密度脂蛋白试剂浓度的确定装置运行时，处理器设置为执行存储器存储的计算机执行指令，以使低密度脂蛋白试剂浓度的确定装 置执行如上述第一方面提供的低密度脂蛋白试剂浓度的确定方法。
一实现方式中，该低密度脂蛋白试剂浓度的确定装置还可以包括收发器，该收发器设置为在低密度脂蛋白试剂浓度的确定装置的处理器的控制下，执行收发数据、信令或者信息的步骤，例如，获取低密度脂蛋白试剂的目标图像。
一实现方式中，该低密度脂蛋白试剂浓度的确定装置可以是设置为实现低密度脂蛋白试剂浓度的确定的物理机，也可以是物理机中的一部分装置，例如可以是物理机中的芯片系统。该芯片系统设置为支持低密度脂蛋白试剂浓度的确定装置实现第一方面中所涉及的功能，例如，接收，发送或处理上述低密度脂蛋白试剂浓度的确定方法中所涉及的数据和/或信息。该芯片系统包括芯片，也可以包括其他分立器件或电路结构。
第四方面，本申请提供一种计算机可读存储介质，计算机可读存储介质中存储有指令，当计算机执行指令时，使得计算机执行如第一方面提供的低密度脂蛋白试剂浓度的确定方法。
第五方面，本申请提供一种计算机程序产品，该计算机程序产品包括计算机指令，当计算机指令在计算机上运行时，使得计算机执行如第一方面提供的低密度脂蛋白试剂浓度的确定方法。
上述计算机指令可以全部或者部分存储在计算机可读存储介质上。其中，计算机可读存储介质可以与低密度脂蛋白试剂浓度的确定装置的处理器封装在一起的，也可以与低密度脂蛋白试剂浓度的确定装置的处理器单独封装，本申请对此不做限定。
本申请中第二方面、第三方面、第四方面以及第五方面的描述，可以参考第一方面的描述；并且，第二方面、第三方面、第四方面、以及第五方面的描述的效果，可以参考第一方面的效果分析。
在本申请中，上述低密度脂蛋白试剂浓度的确定装置的名字对设备或功能模块本身不构成限定，在实际实现中，这些设备或功能模块可以以其他名称出现。只要每个设备或功能模块的功能和本申请中的设备或功能模块的功能类似。
附图说明
图1为本申请实施例提供的一种低密度脂蛋白试剂浓度的确定方法的流程示意图；
图2为本申请实施例提供的一种原始扫描图像的示意图；
图3为本申请实施例提供的一种区域图像的示意图；
图4为本申请实施例提供的一种根据第一分界线和第二分界线，对区域图像进行裁剪得到的图像的示意图；
图5为本申请实施例提供的一种多个目标图像的示意图；
图6为本申请实施例提供的另一种低密度脂蛋白试剂浓度的确定方法的流程示意图；
图7为本申请实施例提供的又一种低密度脂蛋白试剂浓度的确定方法的流程示意图；
图8为本申请实施例提供的一种低密度脂蛋白试剂浓度的确定装置的结构示意图；
图9为本申请实施例提供的另一种低密度脂蛋白试剂浓度的确定装置的结构示意图。
具体实施方式
下面结合附图对本申请实施例提供的低密度脂蛋白试剂浓度的确定方法、装置及存储介质进行描述。
本文中术语“和/或”，仅仅是一种描述关联对象的关联关系，表示可以存在三种关系，例如，A和/或B，可以表示：单独存在A，同时存在A和B，单独存在B这三种情况。
本申请中的术语“第一”和“第二”等是用于区别不同的对象，或者用于区别对同一对象的不同处理，而不是用于描述对象的特定顺序。
此外，本申请的描述中所提到的术语“包括”和“具有”以及它们的任何变形，意图在于覆盖不排他的包含。例如包含了一系列步骤或单元的过程、方法、系统、产品或设备没有限定于已列出的步骤或单元，而是可选的还包括其他没有列出的步骤或单元，或可选的还包括对于这些过程、方法、产品或设备固有的其它步骤或单元。
本申请实施例中，“示例性的”或者“例如”等词用于表示作例子、例证或说明。本申请实施例中被描述为“示例性的”或者“例如”的任何实施例或设计方案不应被解释为比其它实施例或设计方案更优选或更具优势。确切而言，使用“示例性的”或者“例如”等词旨在以方式呈现相关概念。
在本申请的描述中，除非另有说明，“多个”的含义是指两个或两个以上。
低密度脂蛋白试剂的浓度可以表征人的身体健康水平，随着医疗水平的提高，对于低密度脂蛋白试剂的浓度检测的需求也越来越多。然而，在需要同时 对大量低密度脂蛋白试剂进行浓度检测时，检测方法检测效率较低。因此，亟待提出一种低密度脂蛋白试剂浓度的确定方法，提高低密度脂蛋白试剂浓度的检测效率。
针对上述相关技术中存在的问题，本申请实施例提供了一种低密度脂蛋白试剂浓度的确定方法、装置及存储介质。该方案可以将待检测的目标图像输入预先训练好的预设随机森林模型，基于目标图像中像素点的灰度值和预设随机森林模型中分类回归树的多个分类节点中每个分类节点的灰度阈值确定出低密度脂蛋白试剂的浓度。通过预设随机森林模型替代人工对低密度脂蛋白试剂的浓度进行检测，因此，可以提高低密度脂蛋白试剂浓度的检测效率。
本申请提供的低密度脂蛋白试剂浓度的确定方法可以应用于低密度脂蛋白试剂浓度的确定装置。其中，低密度脂蛋白试剂浓度的确定装置可以是设置为对低密度脂蛋白试剂的浓度进行检测的服务器。
服务器可以是一台服务器，也可以是由多台服务器组成的服务器集群，本申请实施例对此不做限定。
下面对本申请提供的低密度脂蛋白试剂浓度的确定方法进行说明。
参照图1，本申请实施例提供的低密度脂蛋白试剂浓度的确定方法包括S101-S102：
S101、低密度脂蛋白试剂浓度的确定装置获取低密度脂蛋白试剂的目标图像。
目标图像可以是经过预处理后的图像。
可选的，在一种实现方式中，低密度脂蛋白试剂浓度的确定装置可以先获取原始扫描图像，然后对原始扫描图像进行切割，得到至少一个区域图像，之后对每个区域图像进行处理，得到每个区域图像的目标图像。
原始扫描图像中包括至少一个试剂容器所在区域，试剂容器设置为盛放低密度脂蛋白试剂。至少一个区域图像中的每个区域图像对应一个试剂容器。
示例性的，参照图2，图2提供了一种原始扫描图像，如图2所示，原始扫描图像中包括有多个试剂容器，低密度脂蛋白试剂浓度的确定装置可以对图2的原始扫描图像进行切割，得到多个如图3所示的区域图像，之后对区域图像进行处理，得到区域图像的目标图像。其中，图3中区域图像的横坐标为试剂容器宽度，纵坐标表示试剂容器的长度，单位为mm。
可选的，原始扫描图像一般为标签图像文件格式(Tag Image File Format，TIFF)文件，而图像处理软件一般不支持对TIFF文件的图像进行处理，所以， 低密度脂蛋白试剂浓度的确定装置在获取到原始扫描图像之后，可以对其进行格式转换，转换为图像处理软件支持的便携式网络图形(Portable Network Graphics，PNG)或联合图像专家组(Joint Photographic Experts Group，JPEG)等格式。
可选的，如图2所示，原始扫描图像中的一些试剂容器可能并未盛放试剂，从图2中可以看出，未盛放试剂的试剂容器的多个像素点的灰度值均相同，而盛放试剂的试剂容器的多个像素点的灰度值呈规律分布。所以，可选的，低密度脂蛋白试剂浓度的确定装置在对原始扫描图像进行切割时，可以基于像素点的灰度值裁剪掉原始扫描图像中未盛放试剂的试剂容器所在的区域。
另外，原始扫描图像中还可以包括盛放试剂容器的试管架(比如，图2中试剂容器底部的黑色区域)，可以基于多个像素点的灰度值裁剪掉该部分。
可选的，在一种实现方式中，低密度脂蛋白试剂浓度的确定装置可以确定出每个区域图像的第一分界线和第二分界线；然后基于第一分界线和第二分界线，对该区域图像进行裁剪，之后基于第二分界线对裁剪后的区域图像进行填充，以将裁剪后的区域图像填充为预设尺寸的区域图像，将填充后的预设尺寸的区域图像确定为目标图像。
第一分界线对应低密度脂蛋白试剂的液面，第二分界线对应低密度脂蛋白试剂的分层面。预设尺寸可以是人为事先确定的尺寸大小，比如预设尺寸可以是20mm*300mm。在实际应用中，预设尺寸还可以是其他尺寸大小，本申请实施例对此不做限定。
示例性的，如图4所示，图4为一种根据第一分界线和第二分界线，对区域图像进行裁剪得到的图像。在第二分界线尾部对裁剪后的区域图像进行填充，将填充后的预设尺寸的区域图像确定为目标图像。示例性的，如图5所示，图5为在第二分界线尾部对裁剪后的区域图像进行填充得到的目标图像。
在实际应用中，还可以通过其他方式获取目标图像，本申请实施例对此不做限定。
S102、低密度脂蛋白试剂浓度的确定装置通过预设随机森林模型，基于目标图像中像素点的灰度值和预设随机森林模型中分类回归树的多个分类节点中每个分类节点的灰度阈值，确定多个分类节点中与目标图像关联的目标分类节点，且根据目标分类节点的浓度值，确定低密度脂蛋白试剂的浓度。
可选的，在一种实现方式中，可以通过以下方式确定预设随机森林模型：基于训练集、测试集以及第四预设数值N，训练N个分类回归树，根据N个分类回归树得到预设随机森林模型。确定出预设随机森林模型后，在预设随机森 林模型的N个分类回归树中，分别确定目标图像关联的N个目标分类节点，之后可以将N个目标分类节点的浓度值的平均值，确定为低密度脂蛋白试剂的浓度。
第四预设数值N可以是人为事先确定的预设随机森林模型中分类回归树的数量，比如，N可以为5，在实际应用中，N也可以为其他数值，本申请实施例对此不做限定。
训练集和测试集中包括有样本图像和样本图像对应的浓度，训练集中的样本图像用于训练预设随机森林模型，测试集中的样本图像用于对预设随机森林模型中的灰度阈值进行调优，优化预设随机森林模型，提高检测的准确率。
样本图像也是经过预处理后的预设尺寸的图像，处理过程与目标图像相同，本申请实施例在此不再赘述。
可选的，在一种实现方式中，可以通过以下方式确定分类回归树：首先，低密度脂蛋白试剂浓度的确定装置从训练集中有放回的随机选取第一预设数值的样本图像，将第一预设数值的样本图像均确定为分类回归树的分类节点中的根节点的样本图像；然后，执行步骤A：基于根节点的灰度阈值，将根节点的所有样本图像划分至根节点的两个子节点；执行步骤B：判断根节点的每个子节点是否满足终止条件；执行步骤C：在根节点的该子节点满足终止条件的情况下，获取根节点的该子节点的浓度值；执行步骤D：在根节点的该子节点不满足终止条件的情况下，从根节点的该子节点的所有样本图像中无放回的随机选取第三预设数值的样本图像，将根节点的该子节点作为新的根节点，将第三预设数值的样本图像均作为新的根节点的样本图像，之后针对新的根节点重新执行步骤A和步骤B，直至满足终止条件，并在满足终止条件的情况下，执行步骤C；最后，基于步骤A至步骤D中得到的根节点、多个子节点以及根节点的灰度阈值，训练分类回归树。
训练集中可以包括第二预设数值的样本图像和第二预设数值的样本图像对应的浓度值。
第一预设数值、第二预设数值和第三预设数值可以是人为事先确定的数值，第一预设数值小于或等于第二预设数值，且第三预设数值小于或等于第一预设数值。
终止条件可以根据第一预设数值和分类回归树的深度确定。示例性的，若分类回归树的深度为5，则终止条件可以为，判断根节点的子节点的深度是否为5，在根节点的子节点的深度未达到5的情况下，从根节点的子节点的所有样本图像中选取第三预设数值的样本图像，将该子节点作为新的根节点，将第三预 设数值的样本图像均作为新的根节点的样本图像，之后针对新的根节点重新执行步骤A和步骤B，直至根节点的子节点的深度达到5。另外，终止条件还需要结合第一预设数值确定，低密度脂蛋白试剂浓度的确定装置可以根据第一预设数值和灰度阈值确定划分到子节点的样本图像的数量，在根节点的子节点的样本图像的数量为1，也即是不能继续划分的情况下，确定为达到终止条件。
本申请实施例中，分类节点的浓度值是根据该节点的样本图像的浓度值确定。比如，根节点的子节点的浓度值根据根节点的子节点的样本图像的浓度值确定。示例性的，根节点的子节点的浓度值可以为根节点的子节点的所有样本图像的浓度值的平均值。
可选的，在一种实现方式中，可以通过以下方式确定灰度阈值：
首先，低密度脂蛋白试剂浓度的确定装置遍历分类节点的所有样本图像的像素点的灰度值，以M个预设灰度值分别作为初始阈值对分类节点的所有样本图像进行分类，得到M个分类结果；一个分类结果对应两个分类子节点，M为正整数；根据分类节点的所有样本图像的浓度值确定每个分类结果对应的两个分类子节点中每个分类子节点的浓度值，根据该两个分类子节点分别对应的浓度值，确定该分类结果的分类误差；然后将M个分类结果中，分类误差满足预设条件的分类结果对应的预设灰度值确定为该分类节点的灰度阈值。
可选的，低密度脂蛋白试剂浓度的确定装置可以根据分类节点的所有样本图像的浓度值计算第一节点的样本图像的第一浓度均值和第二节点的样本图像的第二浓度均值。其中，第一节点和第二节点为，以目标预设灰度值作为初始阈值对分类节点的所有样本图像进行分类，得到的分类结果对应的两个分类子节点；目标预设灰度值为M个预设灰度值中的任意一个。然后，低密度脂蛋白试剂浓度的确定装置根据第一浓度均值确定第一节点的样本图像的浓度值的第一均值误差，根据第二浓度均值确定第二节点的样本图像的浓度值的第二均值误差；之后根据第一均值误差与第二均值误差，确定以目标预设灰度值作为初始阈值的分类结果的分类误差。
示例性的，在一种实现的方式中，M个预设灰度值可以为灰度图像的灰度值的取值，分别为0-255中的M个值。
可选的，预设条件可以为：将M个分类结果中，分类误差最小的分类结果对应的预设灰度值确定为灰度阈值。
第一节点的样本图像的第一浓度均值可以为第一节点的所有样本图像的浓度值的平均值，第二节点的样本图像的第二浓度均值可以为第二节点的所有样本图像的浓度值的平均值。
可选的，以目标预设灰度值作为初始阈值的分类结果的分类误差可以为第一均值误差与第二均值误差之和。
示例性的，若用j表示样本图像中的像素点，用x (j)表示像素点的灰度值，用s表示目标预设灰度值，可以将像素点的灰度值小于或等于s的样本图像划分至第一节点，将像素点的灰度值大于s的样本图像划分至第二节点。用R1表示第一节点的样本图像的集合，R2表示第二节点的样本图像的集合，则可以得到表达式(1)：
R 1(j,s)＝{x|x (j)≤s},R 2(j,s)＝{x|x (j)＞s}           (1)
用N m表示划分至子节点的样本图像的数量，则第一节点的样本图像的第一浓度均值和第二节点的样本图像的第二浓度均值可以用表达式(2)表示：
Figure PCTCN2022091834-appb-000001
其中，y i表示样本图像i的实际浓度值。
低密度脂蛋白试剂浓度的确定装置确定出M个分类结果的分类子节点的样本图像的第一浓度均值和第二浓度均值之后，可以通过表达式(3)确定出分类误差最小的分类结果对应的划分点(j，s)：
Figure PCTCN2022091834-appb-000002
之后，可以将分类误差最小的分类结果对应的划分点(j，s)的中的预设灰度值确定为灰度阈值。
低密度脂蛋白试剂浓度的确定装置可以对分类回归树的根节点的样本图像进行节点划分，不断的将节点划分为两个子节点，直至满足终止条件，将根节点的所有样本图像划分为T个区域(R 1、R 2…R T)，生成分类回归树，分类回归树的模型可以用表达式(4)表示：
Figure PCTCN2022091834-appb-000003
其中，I为指数函数，可以用表达式(5)表示：
Figure PCTCN2022091834-appb-000004
本申请实施例提供的确定灰度阈值的方法和确定分类回归树的方法仅作为示例，在实际应用中，还可以基于样本图像中像素点的灰度值以及样本图像的实际浓度，通过其他方式确定灰度阈值和分类回归树，本申请实施例对此不做限定。
本申请实施例提供的低密度脂蛋白试剂浓度的确定方法中，由于低密度脂蛋白试剂的浓度是与包含低密度脂蛋白试剂的图像中像素点的灰度值有关的， 所以可以基于样本图像中像素点的灰度值和样本图像对应的浓度训练出预设随机森林模型中的分类回归树。之后，即可以将待检测的目标图像输入预先训练好的预设随机森林模型，基于目标图像中像素点的灰度值和预设随机森林模型中分类回归树的多个分类节点中每个分类节点的灰度阈值确定出低密度脂蛋白试剂的浓度。本实施例中，预设随机森林模型可以替代人工对低密度脂蛋白试剂的浓度进行检测，因此，本申请实施例提供的低密度脂蛋白试剂浓度的确定方法可以提高低密度脂蛋白试剂浓度的检测效率。
综合以上描述，如图6所示，本申请实施例还提供了一种预设随机森林模型的训练方法，包括S701-S7010：
S701、低密度脂蛋白试剂浓度的确定装置从训练集中选取第一预设数值的样本图像，将第一预设数值的样本图均确定为分类回归树的分类节点中的根节点的样本图像。
S702、低密度脂蛋白试剂浓度的确定装置遍历根节点的所有样本图像的像素点的灰度值，以M个预设灰度值分别作为初始阈值对根节点的所有样本图像进行分类，得到M个分类结果。
S703、低密度脂蛋白试剂浓度的确定装置根据根节点的所有样本图像的浓度值确定每个分类结果对应的两个子节点中每个子节点的浓度值，根据该两个子节点分别对应的浓度值确定该分类结果的分类误差。
S704、低密度脂蛋白试剂浓度的确定装置将M个分类结果中，分类误差满足预设条件的分类结果对应的预设灰度值确定为根节点的灰度阈值。
S705、低密度脂蛋白试剂浓度的确定装置基于根节点的灰度阈值，将根节点的所有样本图像划分至根节点的两个子节点。
S706、低密度脂蛋白试剂浓度的确定装置判断根节点的每个子节点是否满足终止条件。
在根节点的该子节点满足终止条件的情况下，低密度脂蛋白试剂浓度的确定装置执行步骤S707；在根节点的该子节点不满足终止条件的情况下，执行步骤S708。
S707、低密度脂蛋白试剂浓度的确定装置获取根节点的该子节点的浓度值。
在步骤S707之后，执行步骤S709。
S708、低密度脂蛋白试剂浓度的确定装置从根节点的该子节点的所有样本图像中选取第三预设数值的样本图像，将该子节点作为新的根节点，将第三预设数值的样本图像均作为新的根节点的样本图像。
在步骤S708之后，返回重新执行步骤S702。
S709、低密度脂蛋白试剂浓度的确定装置基于模型训练过程中得到的根节点、所有子节点以及根节点的灰度阈值，训练分类回归树。
S7010、低密度脂蛋白试剂浓度的确定装置基于训练集、测试集以及第四预设数值N，训练N个分类回归树，根据N个分类回归树得到预设随机森林模型。
可选的，如图7所示，图1中的S101可以替换为S1011-S1014：
S1011、低密度脂蛋白试剂浓度的确定装置获取原始扫描图像，并对原始扫描图像进行切割，得到至少一个区域图像。
S1012、低密度脂蛋白试剂浓度的确定装置确定每个区域图像的第一分界线和第二分界线。
S1013、低密度脂蛋白试剂浓度的确定装置基于第一分界线和第二分界线，对该区域图像进行裁剪，且基于第二分界线对裁剪后的区域图像进行填充，以将裁剪后的区域图像填充为预设尺寸的区域图像。
S1014、低密度脂蛋白试剂浓度的确定装置将填充后的预设尺寸的区域图像确定为目标图像。
如图8所示，本申请实施例还提供了一种低密度脂蛋白试剂浓度的确定装置，该装置可以包括：获取模块11和确定模块12。
获取模块11执行上述方法实施例中的S101，确定模块12执行上述方法实施例中的S102。
获取模块11，设置为获取低密度脂蛋白试剂的目标图像；确定模块12，设置为通过预设随机森林模型，基于获取模块11获取的目标图像中像素点的灰度值和预设随机森林模型中分类回归树的多个分类节点中每个分类节点的灰度阈值，确定多个分类节点中与目标图像关联的目标分类节点，且根据目标分类节点的浓度值，确定低密度脂蛋白试剂的浓度。
可选的，在一种实现方式中，本申请提供的低密度脂蛋白试剂浓度的确定装置还可以包括分类模块；分类模块，设置为遍历分类节点的所有样本图像的像素点的灰度值，以M个预设灰度值分别作为初始阈值对分类节点的所有样本图像进行分类，得到M个分类结果；一个分类结果对应两个分类子节点，M为正整数；确定模块12，还设置为根据分类节点的所有样本图像的浓度值确定每个分类结果对应的两个分类子节点中每个分类子节点的浓度值，根据该两个分类子节点分别对应的浓度值，确定该分类结果的分类误差；确定模块12，还设置为将M个分类结果中，分类误差满足预设条件的分类结果对应的预设灰度值 确定为该分类节点的灰度阈值。
可选的，在另一种实现方式中，确定模块12设置为：
根据该分类节点的所有样本图像的浓度值计算第一节点的样本图像的第一浓度均值和第二节点的样本图像的第二浓度均值；第一节点和第二节点为，以目标预设灰度值作为初始阈值对该分类节点的所有样本图像进行分类，得到的分类结果对应的两个分类子节点；目标预设灰度值为M个预设灰度值中的任意一个；根据第一浓度均值确定第一节点的样本图像的浓度值的第一均值误差，根据第二浓度均值确定第二节点的样本图像的浓度值的第二均值误差；根据第一均值误差与第二均值误差，确定以目标预设灰度值作为初始阈值的分类结果的分类误差。
可选的，在另一种实现方式中，确定模块12还设置为：
从训练集中选取第一预设数值的样本图像，将第一预设数值的样本图均确定为分类回归树的分类节点中的根节点的样本图像；训练集中包括第二预设数值的样本图像和第二预设数值的样本图像对应的浓度值；第一预设数值小于或等于第二预设数值；步骤A：基于根节点的灰度阈值，将根节点的所有样本图像划分至根节点的两个子节点；步骤B：判断根节点的每个子节点是否满足终止条件；终止条件根据第一预设数值和分类回归树的深度确定；步骤C：在根节点的该子节点满足终止条件的情况下，获取根节点的该子节点的浓度值；根节点的该子节点的浓度值根据根节点的该子节点的所有样本图像的浓度值确定；步骤D：在根节点的该子节点不满足终止条件的情况下，从根节点的该子节点的所有样本图像中选取第三预设数值的样本图像，将该子节点作为新的根节点，将第三预设数值的样本图像均作为新的根节点的样本图像，之后针对新的根节点重新执行步骤A和步骤B，直至满足终止条件，并在满足终止条件的情况下，执行步骤C；第三预设数值小于或等于第一预设数值；基于步骤A至步骤D中得到的根节点、所有子节点以及根节点的灰度阈值，训练分类回归树。
可选的，在另一种实现方式中，获取模块11设置为：
获取原始扫描图像；原始扫描图像中包括至少一个试剂容器所在区域；试剂容器设置为盛放低密度脂蛋白试剂；对原始扫描图像进行切割，得到至少一个区域图像；一个区域图像对应一个试剂容器；确定模块12还设置为：对每个区域图像进行处理，得到该区域图像的目标图像。
可选的，在另一种实现方式中，确定模块12设置为：
确定该区域图像的第一分界线和第二分界线；第一分界线对应低密度脂蛋白试剂的液面；第二分界线对应低密度脂蛋白试剂的分层面；基于第一分界线 和第二分界线，对该区域图像进行裁剪；基于第二分界线对裁剪后的区域图像进行填充，以将裁剪后的区域图像填充为预设尺的区域图像寸；将填充后的预设尺寸的区域图像确定为目标图像。
可选的，在另一种实现方式中，本申请提供的低密度脂蛋白试剂浓度的确定装置还可以包括训练模块；训练模块，设置为基于训练集、测试集以及第四预设数值N，训练N个分类回归树，根据N个分类回归树得到预设随机森林模型；确定模块12设置为：在N个分类回归树中，分别确定目标图像关联的N个目标分类节点，将N个目标分类节点的浓度值的平均值，确定为低密度脂蛋白试剂的浓度。
可选的，低密度脂蛋白试剂浓度的确定装置还可以包括存储模块，存储模块设置为存储该低密度脂蛋白试剂浓度的确定装置的程序代码等。
如图9所示，本申请实施例还提供一种低密度脂蛋白试剂浓度的确定装置，包括存储器41、处理器42(42-1和42-2)、总线43和通信接口44；存储器41设置为存储计算机执行指令，处理器42与存储器41通过总线43连接；当低密度脂蛋白试剂浓度的确定装置运行时，处理器42执行存储器41存储的计算机执行指令，以使低密度脂蛋白试剂浓度的确定装置执行如上述实施例提供的低密度脂蛋白试剂浓度的确定方法。
在实现中，作为一种实施例，处理器42(42-1和42-2)可以包括一个或多个中央处理器(Central Processing Unit，CPU)，例如图9中所示的CPU0和CPU1。且作为一种实施例，低密度脂蛋白试剂浓度的确定装置可以包括多个处理器42，例如图9中所示的处理器42-1和处理器42-2。这些处理器42中的每一个CPU可以是一个单核处理器(single-CPU)，也可以是一个多核处理器(multi-CPU)。这里的处理器42可以指一个或多个设备、电路、和/或设置为处理数据(例如计算机程序指令)的处理核。
存储器41可以是只读存储器41(Read-Only Memory，ROM)或可存储静态信息和指令的其他类型的静态存储设备，随机存取存储器(Random Access Memory，RAM)或者可存储信息和指令的其他类型的动态存储设备，也可以是电可擦可编程只读存储器(Electrically Erasable Programmable Read-Only Memory，EEPROM)、只读光盘(Compact Disc Read-Only Memory，CD-ROM)或其他光盘存储、光碟存储(包括压缩光碟、激光碟、光碟、数字通用光碟、蓝光光碟等)、磁盘存储介质或者其他磁存储设备、或者能够设置为携带或存储具有指令或数据结构形式的期望的程序代码并能够由计算机存取的任何其他介质，但不限于此。存储器41可以是独立存在，通过总线43与处理器42相连接。存储器41也可以和处理器42集成在一起。
在实现中，存储器41，设置为存储本申请中的数据和执行本申请的软件程序对应的计算机执行指令。处理器42可以通过运行或执行存储在存储器41内的软件程序，以及调用存储在存储器41内的数据，低密度脂蛋白试剂浓度的确定装置的多种功能。
通信接口44，使用任何收发器一类的装置，设置为与其他设备或通信网络通信，如控制系统、无线接入网(Radio Access Network，RAN)，无线局域网(Wireless Local Area Networks，WLAN)等。通信接口44可以包括接收单元实现接收功能，以及发送单元实现发送功能。
总线43，可以是工业标准体系结构(Industry Standard Architecture，ISA)总线、外部设备互连(Peripheral Component Interconnect，PCI)总线或扩展工业标准体系结构(Extended Industry Standard Architecture，EISA)总线等。该总线43可以分为地址总线、数据总线、控制总线等。为便于表示，图9中仅用一条粗线表示，但并不表示仅有一根总线或一种类型的总线。
作为一个示例，结合图8，低密度脂蛋白试剂浓度的确定装置中的获取模块11实现的功能与图9中的接收单元实现的功能相同，低密度脂蛋白试剂浓度的确定装置中的确定模块12实现的功能与图9中的处理器实现的功能相同，低密度脂蛋白试剂浓度的确定装置中的存储模块实现的功能与图9中的存储器实现的功能相同。
本实施例中相关内容的解释可参考上述方法实施例，此处不再赘述。
通过以上的实施方式的描述，为描述的方便和简洁，仅以上述多个功能模块的划分进行举例说明，实际应用中，可以根据需要而将上述功能分配由不同的功能模块完成，即将装置的内部结构划分成不同的功能模块，以完成以上描述的全部或者部分功能。上述描述的系统，装置和单元的工作过程，可以参考前述方法实施例中的对应过程，在此不再赘述。
本申请实施例还提供一种计算机可读存储介质，计算机可读存储介质中存储有指令，当计算机执行该指令时，使得计算机执行上述实施例提供的低密度脂蛋白试剂浓度的确定方法。
计算机可读存储介质，例如可以是但不限于电、磁、光、电磁、红外线、或半导体的系统、装置或器件，或者任意以上的组合。计算机可读存储介质的例子(非穷举的列表)包括：具有一个或多个导线的电连接、便携式计算机磁盘、硬盘、RAM、ROM、可擦式可编程只读存储器(Erasable Programmable Read Only Memory，EPROM)、寄存器、硬盘、光纤、CD-ROM、光存储器件、磁存储器件、或者上述的任意合适的组合、或者本领域熟知的任何其它形式的计 算机可读存储介质。一种示例性的存储介质耦合至处理器，从而使处理器能够从该存储介质读取信息，且可向该存储介质写入信息。当然，存储介质也可以是处理器的组成部分。处理器和存储介质可以位于特定用途集成电路(Application Specific Integrated Circuit，ASIC)中。在本申请实施例中，计算机可读存储介质可以是任何包含或存储程序的有形介质，该程序可以被指令执行系统、装置或者器件使用或者与其结合使用。

Claims (10)

  1. 一种低密度脂蛋白试剂浓度的确定方法，包括：
    获取低密度脂蛋白试剂的目标图像；
    通过预设随机森林模型，基于所述目标图像中像素点的灰度值和所述预设随机森林模型中分类回归树的多个分类节点中每个分类节点的灰度阈值，确定所述多个分类节点中与所述目标图像关联的目标分类节点，且根据所述目标分类节点的浓度值，确定所述低密度脂蛋白试剂的浓度。
  2. 根据权利要求1所述的方法，还包括，通过以下方式确定所述分类节点的灰度阈值：
    遍历所述分类节点的所有样本图像的像素点的灰度值，以M个预设灰度值分别作为初始阈值对所述分类节点的所有样本图像进行分类，得到M个分类结果；其中，一个分类结果对应两个分类子节点，M为正整数；
    根据所述分类节点的所有样本图像的浓度值确定每个分类结果对应的两个分类子节点中每个分类子节点的浓度值，根据所述两个分类子节点分别对应的浓度值，确定所述每个分类结果的分类误差；
    将所述M个分类结果中，分类误差满足预设条件的分类结果对应的预设灰度值确定为所述分类节点的灰度阈值。
  3. 根据权利要求2所述的方法，其中，所述根据所述分类节点的所有样本图像的浓度值确定每个分类结果对应的两个分类子节点中每个分类子节点的浓度值，根据所述两个分类子节点分别对应的浓度值，确定所述每个分类结果的分类误差，包括：
    根据所述分类节点的所有样本图像的浓度值计算第一节点的样本图像的第一浓度均值和第二节点的样本图像的第二浓度均值；其中，所述第一节点和所述第二节点为，以目标预设灰度值作为初始阈值对所述分类节点的所有样本图像进行分类，得到的分类结果对应的两个分类子节点；所述目标预设灰度值为所述M个预设灰度值中的一个；
    根据所述第一浓度均值确定所述第一节点的样本图像的浓度值的第一均值误差，根据所述第二浓度均值确定所述第二节点的样本图像的浓度值的第二均值误差；
    根据所述第一均值误差与所述第二均值误差，确定以所述目标预设灰度值作为初始阈值的分类结果的分类误差。
  4. 根据权利要求1-3中任一项所述的方法，还包括，通过以下方式确定所述分类回归树：
    从训练集中选取第一预设数值的样本图像，将所述第一预设数值的样本图均确定为所述分类回归树的分类节点中的根节点的样本图像；其中，所述训练集中包括第二预设数值的样本图像和所述第二预设数值的样本图像对应的浓度值；所述第一预设数值小于或等于所述第二预设数值；
    基于所述根节点的灰度阈值，将所述根节点的所有样本图像划分至所述根节点的两个子节点；
    判断所述根节点的每个子节点是否满足终止条件；其中，所述终止条件根据所述第一预设数值和所述分类回归树的深度确定；
    在所述每个子节点满足所述终止条件的情况下，获取所述每个子节点的浓度值；其中，所述每个子节点的浓度值根据所述每个子节点的所有样本图像的浓度值确定；
    在所述每个子节点不满足所述终止条件的情况下，从所述每个子节点的所有样本图像中选取第三预设数值的样本图像，将所述每个子节点作为新的根节点，将所述第三预设数值的样本图像均作为所述新的根节点的样本图像，针对所述新的根节点重新执行所述基于所述根节点的灰度阈值，将所述根节点的所有样本图像划分至所述根节点的两个子节点和所述判断所述根节点的每个子节点是否满足终止条件的操作，直至满足所述终止条件，并在满足所述终止条件的情况下，执行所述获取所述每个子节点的浓度值的操作；所述第三预设数值小于或等于所述第一预设数值；
    基于得到的所述根节点、所有子节点以及所述根节点的灰度阈值，训练所述分类回归树。
  5. 根据权利要求1所述的方法，其中，所述获取低密度脂蛋白试剂的目标图像，包括：
    获取原始扫描图像；其中，所述原始扫描图像中包括至少一个试剂容器所在区域；所述试剂容器设置为盛放所述低密度脂蛋白试剂；
    对所述原始扫描图像进行切割，得到至少一个区域图像；一个区域图像对应一个试剂容器；
    对每个区域图像进行处理，得到所述每个区域图像的目标图像。
  6. 根据权利要求5所述的方法，其中，所述对每个区域图像进行处理，得到所述每个区域图像的目标图像，包括：
    确定所述每个区域图像的第一分界线和第二分界线；其中，所述第一分界线对应所述低密度脂蛋白试剂的液面；所述第二分界线对应所述低密度脂蛋白 试剂的分层面；
    基于所述第一分界线和所述第二分界线，对所述每个区域图像进行裁剪；
    基于所述第二分界线对裁剪后的区域图像进行填充，以将所述裁剪后的区域图像填充为预设尺寸的区域图像；
    将填充后的所述预设尺寸的区域图像确定为所述目标图像。
  7. 根据权利要求1-6中任一项所述的方法，还包括：
    基于训练集、测试集以及第四预设数值N，训练N个分类回归树，根据所述N个分类回归树得到所述预设随机森林模型；
    所述确定所述多个分类节点中与所述目标图像关联的目标分类节点，且根据所述目标分类节点的浓度值，确定所述低密度脂蛋白试剂的浓度，包括：
    在所述N个分类回归树中，分别确定所述目标图像关联的N个目标分类节点，将所述N个目标分类节点的浓度值的平均值，确定为所述低密度脂蛋白试剂的浓度。
  8. 一种低密度脂蛋白试剂浓度的确定装置，包括：
    获取模块，设置为获取低密度脂蛋白试剂的目标图像；
    确定模块，设置为通过预设随机森林模型，基于所述获取模块获取的所述目标图像中像素点的灰度值和所述预设随机森林模型中分类回归树的多个分类节点中每个分类节点的灰度阈值，确定所述多个分类节点中与所述目标图像关联的目标分类节点，且根据所述目标分类节点的浓度值，确定所述低密度脂蛋白试剂的浓度。
  9. 一种低密度脂蛋白试剂浓度的确定装置，包括存储器、处理器、总线和通信接口；所述存储器设置为存储计算机执行指令，所述处理器与所述存储器通过所述总线连接；
    当所述低密度脂蛋白试剂浓度的确定装置运行时，所述处理器设置为执行所述存储器存储的所述计算机执行指令，以使所述低密度脂蛋白试剂浓度的确定装置执行如权利要求1-7中任一项所述的低密度脂蛋白试剂浓度的确定方法。
  10. 一种计算机可读存储介质，存储有指令，当计算机执行所述指令时，使得所述计算机执行如权利要求1-7中任一项所述的低密度脂蛋白试剂浓度的确定方法。
PCT/CN2022/091834 2021-05-11 2022-05-10 低密度脂蛋白试剂浓度的确定方法、装置及存储介质 WO2022237746A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202110510569.2A CN113096125B (zh) 2021-05-11 2021-05-11 低密度脂蛋白试剂浓度的确定方法、装置及存储介质
CN202110510569.2 2021-05-11

Publications (1)

Publication Number Publication Date
WO2022237746A1 true WO2022237746A1 (zh) 2022-11-17

Family

ID=76665078

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/091834 WO2022237746A1 (zh) 2021-05-11 2022-05-10 低密度脂蛋白试剂浓度的确定方法、装置及存储介质

Country Status (2)

Country Link
CN (1) CN113096125B (zh)
WO (1) WO2022237746A1 (zh)

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113096125B (zh) * 2021-05-11 2024-05-31 上海宝藤生物医药科技股份有限公司 低密度脂蛋白试剂浓度的确定方法、装置及存储介质
CN113674367B (zh) * 2021-08-20 2024-03-26 上海宝藤生物医药科技股份有限公司 一种电泳后脂蛋白胆固醇试剂扫描图的预处理方法
CN113960146B (zh) * 2021-09-03 2024-04-16 上海淘源生物科技有限公司 脂蛋白亚分型检测方法及系统
CN115147395A (zh) * 2022-07-20 2022-10-04 上海宝藤生物医药科技股份有限公司 一种脂蛋白亚型组分划分方法、装置、设备及存储介质

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20160328841A1 (en) * 2015-05-08 2016-11-10 Siemens Aktiengesellschaft Region Clustering Forest for Analyzing Medical Imaging Data
CN107590510A (zh) * 2017-08-29 2018-01-16 上海联影医疗科技有限公司 一种图像定位方法、装置、计算机及存储介质
CN113096125A (zh) * 2021-05-11 2021-07-09 上海宝藤生物医药科技股份有限公司 低密度脂蛋白试剂浓度的确定方法、装置及存储介质

Family Cites Families (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4920498A (en) * 1985-08-17 1990-04-24 Olympus Optical Co., Ltd. Method of processing and analyzing electrophoretic image, and method of displaying electrophoregram and a medium for recording electrophoregram
SG2014007454A (en) * 2010-08-13 2014-07-30 Somalogic Inc Pancreatic cancer biomarkers and uses thereof
CN106068449A (zh) * 2014-03-05 2016-11-02 思勘度股份有限公司 通过量化和解释颜色的分析物浓度
CN105606608B (zh) * 2016-03-28 2018-12-07 广州博厚医疗技术有限公司 基于图像灰度处理的数据计算方法及其在检测领域的应用
CN113607719B (zh) * 2016-11-29 2024-10-22 普诺森公司 用于同时检测大范围蛋白质浓度的方法和装置
CN109030480B (zh) * 2018-08-16 2021-03-19 湖南友哲科技有限公司 样品分析方法、装置、可读存储介质及计算机设备
CN109919054B (zh) * 2019-02-25 2023-04-07 电子科技大学 一种基于机器视觉的试剂卡自动分类的检测方法
CN110738641B (zh) * 2019-10-07 2022-07-01 福州大学 基于图像处理及kelm的医药试剂浓度定性检测方法
CN111812091B (zh) * 2020-06-28 2023-09-05 上海交通大学 芯片凝胶电泳及其在线uv-vis成像检测装置
CN112700859A (zh) * 2020-12-15 2021-04-23 贵州小宝健康科技有限公司 一种基于医学影像的医疗诊断辅助方法及系统

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20160328841A1 (en) * 2015-05-08 2016-11-10 Siemens Aktiengesellschaft Region Clustering Forest for Analyzing Medical Imaging Data
CN107590510A (zh) * 2017-08-29 2018-01-16 上海联影医疗科技有限公司 一种图像定位方法、装置、计算机及存储介质
CN113096125A (zh) * 2021-05-11 2021-07-09 上海宝藤生物医药科技股份有限公司 低密度脂蛋白试剂浓度的确定方法、装置及存储介质

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
FANG XINRUI, WEN ZHAOFEI;CHEN JILONG;WU SHENGJUN;HUANG YUANYANG;MA MAOHUA: "Remote Sensing Estimation of Suspended Sediment Concentration based on Random Forest Regression Model", JOURNAL OF REMOTE SENSING, vol. 23, no. 4, 14 August 2019 (2019-08-14), pages 756 - 772, XP093004502, ISSN: 1007-4619, DOI: 10.11834/jrs.20197498] *
ZHANG YING, QIANQIAN GAO: "Chinese Journal of Environmental Engineering Water quality evaluation of Chaohu Lake based on random forest method", HUANJING GONGCHENG XUEBAO - CHINESE JOURNAL OF ENVIRONMENTAL ENGINEERING, ZHONGGUO KEXUEYUAN SHENGTAI HUANJING YANJIU ZHONGXIN, CN, vol. 10, no. 2, 28 February 2016 (2016-02-28), CN , pages 992 - 998, XP093004504, ISSN: 1673-9108 *

Also Published As

Publication number Publication date
CN113096125A (zh) 2021-07-09
CN113096125B (zh) 2024-05-31

Similar Documents

Publication Publication Date Title
WO2022237746A1 (zh) 低密度脂蛋白试剂浓度的确定方法、装置及存储介质
CN109933802B (zh) 图文匹配方法、装置及存储介质
CN110532968B (zh) 表格识别方法、装置和存储介质
US7995821B2 (en) Image processing system and image processing method
US20140195891A1 (en) System and method for automatically extracting multi-format data from documents and converting into xml
CN111461243B (zh) 分类方法、装置、电子设备及计算机可读存储介质
CN110807362B (zh) 一种图像检测方法、装置和计算机可读存储介质
RU2579899C1 (ru) Обработка документа с использованием нескольких потоков обработки
JP6876914B2 (ja) 情報処理装置
CN113592807B (zh) 一种训练方法、图像质量确定方法及装置、电子设备
CN105577763A (zh) 一种动态副本一致性维护系统、方法及云存储平台
CN113191235B (zh) 杂物检测方法、装置、设备及存储介质
CN115311216A (zh) 抗核抗体的荧光图片的判读方法及装置、设备及存储介质
AU2017288044A1 (en) Method and system for flexible, high performance structured data processing
JP2023510667A (ja) キャラクタ取得、ページ処理と知識グラフ構築方法及び装置、媒体
CN111597336B (zh) 训练文本的处理方法、装置、电子设备及可读存储介质
CN112070144B (zh) 图像聚类方法、装置、电子设备以及存储介质
CN107346312A (zh) 一种大数据处理方法及系统
CN113139581A (zh) 一种基于多图融合的图像分类方法和系统
CN113256622A (zh) 基于三维图像的目标检测方法、装置及电子设备
CN113361584B (zh) 模型训练的方法及装置、肺动脉高压的测量方法及装置
JP7037180B2 (ja) 学習用データ判別装置および学習用データ判別プログラム
CN117271645A (zh) 一种测试数据的处理方法、装置及计算机可读存储介质
CN116468970A (zh) 模型训练方法、图像处理方法、装置、设备及介质
CN104933055B (zh) 网页识别方法及网页识别装置

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: 22806709

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 22806709

Country of ref document: EP

Kind code of ref document: A1