Blood smear edge detection method, device, equipment, storage medium and product

Abstract

The invention discloses a blood smear edge detection method and device, equipment, a storage medium and a product, and relates to the technical field of medical diagnos.The blood smear edge detection method is based on dense residual features, and the blood smear edge detection method comprises the steps that a blood smear picture is preprocessed, and the preprocessed blood smear picture is obtained; loading the preprocessed blood smear picture to a lightweight segmentation network to obtain a segmented picture; and when the area ratio of the target region in the segmented picture is greater than a preset threshold value, judging that the segmented picture in the visual field is located at the edge of the blood smear. Through the lightweight segmentation network and dense residual feature extraction, the real-time detection requirement of the blood smear is met, and the detection capability of the tiny edge of the blood smear and the edge identification accuracy of the blood smear are improved.

Description

Technical Field

[0001] This application relates to the field of medical diagnosis technologies, and particularly to a method, device, equipment, storage medium and product for blood smear edge detection. Background Art

[0002] The commonly used medical image enhancement methods at present mainly include: (1) Spatial domain methods such as contrast enhancement and sharpening enhancement. Each method has its obvious deficiencies. The contrast enhancement method cannot completely solve the problems existing in histogram equalization, while the sharpening method is only applicable to the sharpening of images with good contrast but blurred details, and the enhancement effect is not obvious when used alone on medical images. (2) Transform domain methods. Wavelet transform is the most common transform domain method. Transform domain methods usually require post-processing in the spatial domain. (3) Combined with statistical methods. Statistical learning methods generally can obtain better enhancement effects, but their processing speed cannot meet the real-time requirements. (4) Algorithms using fuzzy theory and combinatorial optimization algorithms. The image is regarded as a fuzzy set, and the fuzzy theory is used to process the image, and the "fuzzy contrast enhancement operator (INT)" is used to modify pixels to achieve the purpose of contrast enhancement.

[0003] In the field of medical diagnosis, the recognition of regions of interest in blood cell inspection technology is particularly important. However, in the imaging process, there are often some degradation factors, resulting in blurred boundaries and degraded quality of the final image. Summary of the Invention

[0004] The main purpose of this application is to provide a method, device, equipment, storage medium and product for blood smear edge detection, aiming to solve the technical problems in the prior art that the recognition of the edge contour of the segmented blood smear is inaccurate, and the recognition network training model is too large to meet the lightweight requirement.

[0005] To achieve the above purpose, this application proposes a blood smear edge detection method, and the blood smear edge detection method is based on dense residual features;

[0006] The blood smear edge detection method includes:

[0007] Preprocess the blood smear picture to obtain a preprocessed blood smear picture;

[0008] Load the preprocessed blood smear picture into a lightweight segmentation network to obtain a segmented picture;

[0009] When the area ratio of the target region in the segmented picture is greater than a preset threshold, it is determined that the segmented picture within the field of view is at the blood smear edge.

[0010] Optionally, the step of preprocessing the blood smear picture through an algorithm to obtain a preprocessed blood smear picture specifically includes:

[0011] Perform color space separation on the blood smear image;

[0012] Enhance the edge of the separated blood smear image through an algorithm;

[0013] Filter the blood smear image after edge enhancement;

[0014] Perform binary segmentation on the filtered blood smear image to obtain the preprocessed blood smear image.

[0015] Optionally, the step of loading the preprocessed blood smear image into a lightweight segmentation network to obtain a segmented image specifically includes:

[0016] Load the preprocessed blood smear image into a lightweight segmentation network;

[0017] Extract image features of the preprocessed blood smear image using a neural network;

[0018] Obtain a segmented image based on the extracted image features.

[0019] Optionally, after the step of loading the preprocessed blood smear image into a lightweight segmentation network to obtain a segmented image, it further includes:

[0020] Determine the total area of the blood smear image within the field of view and the area of each region of the blood smear image based on the segmented image;

[0021] Select a target region area from each of the region areas;

[0022] Determine the ratio of the target region area in the segmented image based on the target region area and the total area of the blood smear image within the field of view.

[0023] Optionally, the step of selecting a target region area from each of the region areas specifically includes:

[0024] Sort the area of each region of the blood smear image;

[0025] Determine the target region area based on the sorting result.

[0026] Optionally, after the step of determining that the segmented image within the field of view is at the edge of the blood smear when the ratio of the target region area in the segmented image is greater than a preset threshold, it further includes:

[0027] Determine that the segmented image within the field of view is inside the blood smear when the ratio of the target region area in the segmented image is less than the preset threshold.

[0028] In addition, to achieve the above object, the present application also proposes a blood smear edge detection device, and the device includes:

[0029] A processing module for preprocessing a blood smear image to obtain a preprocessed blood smear image;

[0030] A segmentation module for loading the preprocessed blood smear image into a lightweight segmentation network to obtain a segmented image;

[0031] A determination module for determining that the segmented image within the field of view is at the edge of the blood smear when the area ratio of the target region in the segmented image is greater than a preset threshold.

[0032] In addition, to achieve the above object, the present application also provides a blood smear edge detection device, which includes: a memory, a processor, and a computer program stored on the memory and executable on the processor, and the computer program is configured to implement the steps of the blood smear edge detection method.

[0033] In addition, to achieve the above object, the present application also provides a storage medium, which is a computer-readable storage medium, and a computer program is stored on the storage medium, and when the computer program is executed by a processor, the steps of the blood smear edge detection method are implemented.

[0034] In addition, to achieve the above object, the present application also provides a computer program product, which includes a computer program, and when the computer program is executed by a processor, the steps of the blood smear edge detection method are implemented.

[0035] One or more technical solutions proposed by the present application have at least the following effects:

[0036] The present application discloses a blood smear edge detection method, device, equipment, storage medium and product. The blood smear edge detection method is based on dense residual features, and the blood smear edge detection method includes: preprocessing a blood smear image to obtain a preprocessed blood smear image; loading the preprocessed blood smear image into a lightweight segmentation network to obtain a segmented image; when the area ratio of the target region in the segmented image is greater than a preset threshold, determining that the segmented image within the field of view is at the edge of the blood smear. Through the lightweight segmentation network and dense residual feature extraction, the real-time detection requirement of the blood smear is met, and the detection ability of the tiny edge of the blood smear and the recognition accuracy of the blood smear edge are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, for those of ordinary skill in the art, other drawings can also be obtained based on these drawings without creative efforts.

[0038] Figure 1 Schematic flowchart of the first embodiment of the blood smear edge detection method proposed by the embodiments of the present application;

[0039] Figure 2 Schematic flowchart of the second embodiment of the blood smear edge detection method proposed by the embodiments of the present application;

[0040] Figure 3 Schematic flowchart of the third embodiment of the blood smear edge detection method proposed by the embodiments of the present application;

[0041] Figure 4 Schematic diagram of the module structure of the blood smear edge detection device according to the embodiments of the present application;

[0042] Figure 5 Schematic diagram of the device structure of the hardware operating environment involved in the blood smear edge detection method according to the embodiments of the present application;

[0043] Figure 6 Detection effect diagram of the blood smear edge of the present application;

[0044] Figure 7 Dense residual feature extraction diagram of the present application;

[0045] Figure 8 Depthwise separable convolution diagram of the present application;

[0046] Figure 9 Feature weight redistribution module diagram of the present application.

[0047] Explanation of the reference numerals in the drawings:

[0048] Label Name Label Name 10 Processing module 20 Segmentation module 30 Determination module 1001 Processing device 1002 ROM 1003 Storage device 1004 RAM 1005 Bus 1006 I / O interface 1007 Input device 1008 Output device 1009 Communication device

[0049] The realization of the purpose, functional characteristics and advantages of the present application will be further described with reference to the embodiments and the accompanying drawings. Detailed implementation manners

[0050] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of the present application, and are not used to limit the present application.

[0051] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts shall fall within the protection scope of the present application.

[0052] It should be noted that all directional indications (such as up, down, left, right, front, back...) in the embodiments of the present application are only used to explain the relative positional relationship, movement conditions, etc. between components in a specific posture (as shown in the attached drawings). If the specific posture changes, the directional indications will also change accordingly.

[0053] In addition, the descriptions involving "first", "second", etc. in the present application are only for descriptive purposes, and cannot be understood as indicating or implying their relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include at least one such feature. In addition, the technical solutions between various embodiments can be combined with each other, but it must be based on the ability of those of ordinary skill in the art to implement. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the protection scope required by the present application.

[0054] The main solution of the embodiments of the present application is: through a lightweight segmentation network and dense residual feature extraction, the real-time detection requirements of blood smears are met, and the detection ability of the tiny edges of blood smears and the recognition accuracy of blood smear edges are improved.

[0055] In the embodiment, for the convenience of description, the following will be described with the processor of the blood smear edge detection method as the execution subject.

[0056] The present application provides a solution. The present application proposes a blood smear edge detection method, device, equipment, storage medium and product. The blood smear edge detection method is based on dense residual features. The blood smear edge detection method includes: preprocessing a blood smear picture to obtain a preprocessed blood smear picture; loading the preprocessed blood smear picture into a lightweight segmentation network to obtain a segmented picture; when the area ratio of the target area in the segmented picture is greater than a preset threshold, it is determined that the segmented picture within the field of view is at the blood smear edge. Through the lightweight segmentation network and dense residual feature extraction, the real-time detection requirements of blood smears are met, and the detection ability of the tiny edges of blood smears and the recognition accuracy of blood smear edges are improved.

[0057] Based on this, the embodiments of the present application provide a blood smear edge detection method.

[0058] Reference Figure 1 , Figure 1 is a schematic flowchart of the first embodiment of the blood smear edge detection method proposed in the embodiments of the present application.

[0059] Considering that in the prior art, the recognition of the edge contour of the segmented blood smear is inaccurate, and the recognition network training model is large and cannot meet the lightweight requirement. As Figure 1As shown, the blood smear edge detection method in this embodiment is based on dense residual features;

[0060] The blood smear edge detection method includes steps S10 to S30:

[0061] Step S10: Preprocess the blood smear image to obtain a preprocessed blood smear image.

[0062] It should be noted that a blood smear is a commonly used clinical examination method, which involves making a smear of the blood of the person to be examined and then observing indicators such as the morphology, quantity, distribution, and staining characteristics of blood cells under a microscope, so as to provide a basis for disease screening and diagnosis. Preprocessing the blood smear image is to enhance the edge of the blood smear to enhance the robustness of the overall image contour recognition and clarify the boundaries of each region.

[0063] It can be understood that, as Figure 6 shown, Figure 6 is the detection effect diagram of the blood smear edge of this application. Figure 6 From left to right, they are the original image, the binary image, and the result image output by the neural network. The original image is an RGB image with clear and uncomplicated texture. The binary image is the result after obtaining the global optimal threshold through OTSU. The result image is the result of boundary refinement using the neural network algorithm.

[0064] It should be noted that OTSU is a threshold segmentation algorithm widely used in the field of image processing. The basic idea of the OTSU algorithm is to use a certain criterion function to judge the effect of the segmentation threshold. The threshold when the criterion function obtains the optimal value (i.e., the maximum or minimum) is the image segmentation threshold we need to determine. This criterion function is defined as the between-class variance in the OTSU algorithm, that is, to maximize the variance between the foreground and background classes.

[0065] Step S20: Load the preprocessed blood smear image into a lightweight segmentation network to obtain a segmented image.

[0066] It should be noted that, as Figure 7 shown, Figure 7This is the dense residual feature extraction diagram of the present application. The lightweight segmentation network can be MobileNet, SqueezeNet and their variants. The lightweight segmentation network in this embodiment uses the backbone network of ResNet50 and fine-tunes the network, changing some of the convolution operations into depthwise separable convolutions to balance network accuracy and network size and enhance the convenience of later deployment. The feature extraction network (lightweight feature extraction module) inputs feature maps of different levels into the attention mechanism module for significant feature redistribution. The multi-scale dense residual feature extraction module adds inter-level residual connections on the basis of the feature pyramid to promote the interoperability of features; residual connections are used in all adjacent feature layers to form multi-scale dense residual blocks, enriching the output feature information, enhancing the sensitivity to high-frequency information of the picture, effectively capturing the cell boundary position, and obtaining the segmented picture.

[0067] It can be understood that MobileNet is a lightweight deep learning model designed to provide the ability for high-performance image classification and computer vision tasks while maintaining low computational resource consumption. SqueezeNet is a lightweight convolutional neural network (CNN) model that uses a small number of parameters and computations while maintaining high accuracy, making it very suitable for deployment in resource-limited environments such as mobile devices and embedded systems. ResNet50 is a deep convolutional neural network model. The full name of ResNet50 is Residual Network 50, which means residual network 50. The core innovation of ResNet50 is the introduction of residual blocks (Residual Blocks). This structure solves the gradient vanishing and degradation problems in deep networks through skip connections, enabling the network to reach deeper levels without overfitting or gradient problems.

[0068] It should be noted that as Figure 8 shown, Figure 8This is the depthwise separable convolution map of this application. "DepthwiseConvolution" means the layer-by-layer convolution part, "Pointwise Convolution" means the point-by-point convolution part, "MChannel" means that the input feature has M layers, "M Filter" means that there are M convolution kernels, and the size of the convolution kernel depends on the specific situation; "M Feature Map" means the result after layer-by-layer convolution; "NFilter" means that there are N 1*1 convolution kernels; "N FeatureMap" means that there are N output feature layers. After the input feature "M Channel" passes through M "3×3" convolution kernels, the corresponding M-dimensional feature map "M Feature Map" is generated. F "1×1" convolution kernels passing through the point-by-point convolution module can effectively reduce the number of network parameters. If ordinary convolution is used, the number of parameters is: 3*3*M*N, where M represents the number of input feature map channels and N is the number of output feature map channels; if depthwise separable convolution is used, the number of parameters is: 3*3*M+M*N*1*1. Since M and N are integers, 9M+MN must be greater than 9MN. Therefore, using depthwise separable convolution can effectively reduce the number of network model parameters and improve portability. This step is used to replace a common convolution of a lightweight neural network to reduce the number of model parameters and improve the portability of the model.

[0069] It should be noted that if Figure 9 As shown, Figure 9 This is the feature weight redistribution module diagram for this application. The feature weight redistribution module extracts channel and spatial features through the channel module and the spatial module respectively, selects the most representative feature vector, and ensures that the feature pyramid obtains the most effective information. "Channel AttentionModel" represents the channel-level weight distribution module; "Spatial Attention Model" represents the spatial-level weight distribution module. For details on application areas, please refer to Figure 7 The word "Att" is used to filter out the representative vectors from the neural network part that obtains the features and input them into the feature pyramid. The feature weight redistribution module is abbreviated as "Att".

[0070] Step S30: When the area ratio of the target region in the segmented image is greater than a preset threshold, it is determined that the segmented image in the field of view is at the edge of the blood smear.

[0071] It should be noted that if Figure 6As shown in the figure, each closed area in the figure represents an area. The largest single area is located in the left half of the field of view. The target area is the largest single area within the field of view. The target area ratio refers to the ratio of the largest single area within the field of view to all areas, where all areas refer to the field of view area, that is, the picture pixels. The preset threshold can be 0.4, or it can be set according to the actual situation, and this embodiment does not limit it. Whether the field of view of the camera reaches the edge of the picture can be judged through the preset threshold.

[0072] In specific implementation, preprocess the blood smear picture to obtain the preprocessed blood smear picture; load the preprocessed blood smear picture into a lightweight segmentation network to obtain a segmented picture; when the target area ratio in the segmented picture is greater than the preset threshold, it is determined that the segmented picture within the field of view is at the edge of the blood smear. Thus, it is judged whether the field of view of the camera reaches the edge of the picture, improving the detection ability of the tiny edge of the blood smear and the recognition accuracy of the blood smear edge.

[0073] Further, step S10 in this embodiment specifically includes:

[0074] Separate the color space of the blood smear picture;

[0075] Enhance the edge of the separated blood smear picture through an algorithm;

[0076] Filter the blood smear picture after edge enhancement;

[0077] Perform binary segmentation on the filtered blood smear picture to obtain the preprocessed blood smear picture.

[0078] It should be noted that the color space separation is the HSV space decomposition. The HSV space is a color space model that uses three channels of hue, saturation, and value to represent colors. Decomposing a color picture (blood smear picture / original picture) into the HSV space means normalizing the RGB channel information to values between 0 and 1, providing a prerequisite for the neural network training to converge faster. Mapping the RGB variables into the three spaces of the H variable obtains better perceptual information. As shown in the relevant formulas (1)(2)(3)(4). R, G, and B represent the three primary color channels, that is, Red, Green, and Blue. C max represents the maximum value of the three-channel normalized values, C min represents the minimum value of the three-channel normalized values, and δ represents the change range between the maximum and minimum values. Hue, Saturation, and V represent the variables of the three dimensions of the HSV space.

[0079]

[0080]

[0081]

[0082] V = C max (4)

[0083] It should be noted that the algorithm in this embodiment enhances the edges of the blood smear image through the Sobel operator, superimposes two directions through the directional Sobel operator, extracts features through different convolutions in the horizontal and vertical directions, obtains the approximate brightness difference values in these two directions, combines the brightness approximate values in the xy two directions to obtain the brightness value of the corresponding pixel point, better distinguishes the target from the background, and thus enhances the edge. The image after edge enhancement has higher distinguishability. The calculation formula can be expressed as shown in formula (5). dst img , α, β, ScaleAbs, filterx, filtery, src x , src y respectively represent the finally obtained result image, the image product coefficient after convolution in the x direction, the image product coefficient after convolution in the y direction, taking the absolute value, the convolution operator in the x direction, the convolution operator in the y direction, the original image of the convolution operation in the x direction, and the original image of the convolution operation in the y direction.

[0084] dst img = α * ScaleAbs(filterx(src x )) + β * ScaleAbs(filtery(src y )) (5)

[0085] It should be noted that the filtering in this embodiment is Gaussian filtering. If there is overlap between cells or the image after binarization does not form a closed region, it will cause confusion between cells and the background. Therefore, Gaussian filtering is equivalent to the role of expanding the dataset, which not only enhances the robustness of the overall picture contour recognition, but also increases the class distance between the region of interest ROI (region of interest) category and the category inside the blood smear, and reduces the intra-class distance of the category, making the algorithm more sensitive to the region of interest. The formula of the Gaussian filter is shown in formula (6), where x and y represent the pixel positions of the two-dimensional picture.

[0086]

[0087] It should be noted that the binarization segmentation of the filtered blood smear image is a grayscale operation. The grayscale operation is applied to the image after Gaussian filtering, and the image is binarized using the OTSU threshold segmentation type. The entire process is shown in Equation (7). In this patent, since the features in the grayscale image are required for the segmentation task, in order to ensure the grayscale image with more obvious features is extracted, the original image is first operated on by Gaussian filtering, and then the image after Gaussian filtering is sent into the grayscale operation. Before performing the segmentation task, a preliminary rough recognition of the target and the background is carried out, so the OTSU method is used to obtain the thresholds for the two categories of the target and the background. Using this type of threshold segmentation not only has a good effect but also has a fast segmentation speed. Speed is also a key link in the automated blood smear process. As shown in Equation (7), img_org represents the original image of the picture. Gauss represents the Gaussian filtering function. Binary represents the binarization operation of this application.

[0088] img = Binary(Gray(Gauss(img_ori))) (7)

[0089] In specific implementation, the color space of the blood smear image is separated; the edges of the separated blood smear image are enhanced through an algorithm; the blood smear image after edge enhancement is filtered; the filtered blood smear image is binarized and segmented to obtain the preprocessed blood smear image. Thus, the preprocessing of the blood smear image is realized, and the enhancement of the blood smear image is realized to enhance the robustness of the overall image contour recognition and clarify the boundaries of each region.

[0090] Furthermore, step S20 in this embodiment specifically includes:

[0091] Loading the preprocessed blood smear image into a lightweight segmentation network;

[0092] Using a neural network to extract image features from the preprocessed blood smear image;

[0093] Obtaining a segmented image according to the extracted image features.

[0094] It should be noted that the neural network is the backbone network of ResNet50 and is fine-tuned for the network. It can also be set according to the actual situation, and this embodiment does not limit it. The feature extraction of the image is carried out through a feature weight redistribution module.

[0095] In specific implementation, the preprocessed blood smear image is loaded into a lightweight segmentation network; a neural network is used to extract image features from the preprocessed blood smear image; a segmented image is obtained according to the extracted image features, thereby obtaining a blood smear image with refined edges.

[0096] Based on the first embodiment of the present application, in the second embodiment of the present application, for the same or similar content as the above first embodiment, reference can be made to the above introduction and will not be elaborated hereinafter. On this basis, please refer to Figure 2 , Figure 2 which is a schematic flowchart of the second embodiment of the blood smear edge detection method proposed in the embodiment of the present application.

[0097] Considering the recognition of the edge contour of the blood smear, and further in order to determine the total area of the blood smear picture within the field of view and the area of each region of the blood smear picture. As Figure 2 shown, after the step S20 in this embodiment, the following steps are further included:

[0098] Step S201: Determine the total area of the blood smear picture within the field of view and the area of each region of the blood smear picture according to the segmented picture.

[0099] Step S202: Select a target region area from each of the region areas.

[0100] Step S203: Determine the ratio of the target region area in the segmented picture according to the target region area and the total area of the blood smear picture within the field of view.

[0101] It should be noted that the field of view refers to the field of view where the camera detects the blood smear.

[0102] In a specific implementation, according to the segmented picture, determine the total area of the blood smear picture within the field of view and the area of each region of the blood smear picture; select a target region area from each of the region areas; determine the ratio of the target region area in the segmented picture according to the target region area and the total area of the blood smear picture within the field of view, so as to determine whether the camera field of view is at the edge of the blood smear.

[0103] Furthermore, step S202 in this embodiment specifically includes:

[0104] Sort the area of each region of the blood smear picture;

[0105] Determine the target region area according to the sorting result.

[0106] It should be noted that the sorting can be from large to small, or from small to large, or can be set according to the actual situation, and this embodiment does not limit it. The target region area is the largest area selected from each region area.

[0107] In a specific implementation, sort the area of each region of the blood smear picture; determine the target region area according to the sorting result, so as to determine whether the camera field of view is at the edge of the blood smear.

[0108] Based on the second embodiment of the present application, in the third embodiment of the present application, the same or similar content as that in the above-mentioned second embodiment can be referred to the above introduction and will not be repeated hereinafter. On this basis, please refer to Figure 3 , Figure 3 which is a schematic flowchart of the third embodiment of the blood smear edge detection method proposed in the embodiment of the present application.

[0109] Considering whether the segmented picture in the field of view is inside the blood smear, as Figure 3 shown, after the step S30 in this embodiment, the following steps are further included:

[0110] Step S301: When the area ratio of the target region in the segmented picture is less than a preset threshold, it is determined that the segmented picture in the field of view is inside the blood smear.

[0111] It should be noted that the preset threshold can be 0.4, and it can also be set according to the actual situation, which is not limited in this embodiment. The field of view refers to the field of view where the camera detects the blood smear.

[0112] In specific implementation, when the area ratio of the target region in the segmented picture is less than the preset threshold, it is determined that the segmented picture in the field of view is inside the blood smear, so as to ensure the accuracy rate of blood smear edge recognition.

[0113] In addition, to achieve the above purpose, as Figure 4 shown, the present application also proposes a blood smear edge detection device, and the device includes:

[0114] A processing module 10, configured to preprocess the blood smear picture to obtain a preprocessed blood smear picture;

[0115] A segmentation module 20, configured to load the preprocessed blood smear picture into a lightweight segmentation network to obtain a segmented picture;

[0116] A determination module 30, configured to determine that the segmented picture in the field of view is at the edge of the blood smear when the area ratio of the target region in the segmented picture is greater than a preset threshold.

[0117] The fan blade warning device provided by the present application adopts the blood smear edge detection method in the above embodiment, and can solve the technical problems that the recognition of the edge contour of the segmented blood smear in the prior art is inaccurate and the recognition network training model is too large to meet the lightweight requirement. Compared with the prior art, the beneficial effects of the blood smear edge detection device provided by the present application are the same as those of the blood smear edge detection method provided in the above embodiment, and other technical features in the blood smear edge detection device are the same as the features disclosed in the above embodiment method, which will not be repeated here.

[0118] In addition, to achieve the above object, the present application also provides a blood smear edge detection device, which includes: a memory, a processor, and a computer program stored on the memory and executable on the processor, and the computer program is configured to implement the steps of the blood smear edge detection method described above.

[0119] As Figure 5 shown, the blood smear edge detection device may include a processing device 1001 (such as a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM: Read Only Memory) 1002 or a program loaded from a storage device 1003 into a random access memory (RAM: Random Access Memory) 1004. In the RAM 1004, various programs and data required for the operation of the blood smear edge detection device are also stored. The processing device 1001, the ROM 1002, and the RAM 1004 are connected to each other through a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Generally, the following systems may be connected to the I / O interface 1006: an input device 1007 including, for example, a touch screen, a touch pad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output device 1008 including, for example, a liquid crystal display (LCD: Liquid Crystal Display), a speaker, a vibrator, etc.; a storage device 1003 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 1009. The communication device 1009 can allow the blood smear edge detection device to communicate with other devices wirelessly or wiredly to exchange data. Although the figure shows a blood smear edge detection device having various systems, it should be understood that it is not required to implement or have all the shown systems. More or fewer systems may be alternatively implemented or had.

[0120] Specifically, according to the embodiments disclosed in the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes program codes for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through the communication device, or installed from the storage device 1003, or installed from the ROM 1002. When the computer program is executed by the processing device 1001, the above functions defined in the method of the embodiments disclosed in the present application are executed.

[0121] The blood smear edge detection device provided by this application adopts the blood smear edge detection method in the above-mentioned embodiment, which can solve the technical problems in the prior art that the recognition of the edge contour of the segmented blood smear is inaccurate and the recognition network training model is too large to meet the lightweight requirement. Compared with the prior art, the beneficial effects of the blood smear edge detection device provided by this application are the same as those of the blood smear edge detection method provided by the above-mentioned embodiment, and other technical features in this blood smear edge detection device are the same as those disclosed in the method of the previous embodiment, which will not be elaborated here.

[0122] It should be understood that each part disclosed in this application can be implemented by hardware, software, firmware or a combination thereof. In the description of the above embodiments, specific features, structures, materials or characteristics can be combined in a suitable manner in any one or more embodiments or examples.

[0123] As mentioned above, it is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed in this application, and all should be covered by the protection scope of this application. Therefore, the protection scope of this application should be subject to the protection scope of the claims.

[0124] In addition, to achieve the above object, this application also proposes a storage medium. The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the blood smear edge detection method are implemented.

[0125] The computer-readable storage medium provided by the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or components, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections with one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM) or flash memory, optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the above. In this embodiment, the computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, device, or component. The program code contained on the computer-readable storage medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0126] The above computer-readable storage medium may be included in the blood smear edge detection device; or may exist separately without being assembled into the blood smear edge detection device.

[0127] The above computer-readable storage medium carries one or more programs. When the one or more programs are executed by the blood smear edge detection device, the blood smear edge detection device is caused to: during the operation of the fan blade, collect the characteristic signals of the fan blade through a sensor, where the characteristic signals include sound signals and vibration signals; extract the characteristic values of the sound signals and vibration signals in the characteristic signals through an algorithm; and output a warning signal when the difference between the characteristic value and the preset characteristic value is greater than the preset threshold.

[0128] Computer program code for performing the operations of this application can be written in one or more programming languages or combinations thereof. The above-mentioned programming languages include object-oriented programming languages such as Java, Smalltalk, C++, and also include conventional procedural programming languages such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, executed as an independent software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any kind of network, including a local area network (LAN: Local Area Network) or a wide area network (WAN: Wide Area Network), or it can be connected to an external computer (for example, by using an Internet service provider to connect through the Internet).

[0129] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in the flowchart or block diagram may represent a module, a program segment, or a part of code that contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than that marked in the accompanying drawings. For example, two consecutively represented blocks may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and the combination of blocks in the block diagram and / or flowchart, can be implemented by a dedicated hardware-based system for performing the specified functions or operations, or can be implemented by a combination of dedicated hardware and computer instructions.

[0130] The modules described in the embodiments of this application can be implemented in software or in hardware. Among them, the name of the module does not constitute a limitation on the unit itself in some cases.

[0131] The readable storage medium provided in this application is a computer-readable storage medium, and the computer-readable storage medium stores computer-readable program instructions (i.e., computer programs) for performing the above-mentioned blood smear edge detection method, which can solve the technical problems in the prior art that the recognition of the edge contour of the segmented blood smear is inaccurate and the recognition network training model is too large to meet the lightweight requirements. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the blood smear edge detection method provided in the above embodiments, and will not be elaborated here.

[0132] In addition, to achieve the above object, the present application also provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the steps of the blood smear edge detection method described above are implemented.

[0133] The above are only the preferred embodiments of the present application, and do not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made by using the content of the specification and drawings of the present application, or directly or indirectly applied to other related technical fields, shall be equally included in the patent protection scope of the present application.

Claims

1. A blood smear edge detection method, characterized in that: The blood smear edge detection method is based on dense residual features; The blood smear edge detection method comprises: Preprocessing the blood smear image to obtain a preprocessed blood smear image; Loading the preprocessed blood smear image into a lightweight segmentation network to obtain a segmented image; When the area ratio of the target region in the segmented image is greater than a preset threshold, it is determined that the segmented image in the field of view is at the edge of the blood smear.

2. The blood smear edge detection method according to claim 1, characterized in that: The step of preprocessing the blood smear image by an algorithm to obtain a preprocessed blood smear image specifically includes: Perform color space separation on blood smear images; The edge of the separated blood smear image is enhanced by an algorithm; Filter the edge-enhanced blood smear image; The filtered blood smear image is binarized and segmented to obtain a preprocessed blood smear image.

3. The blood smear edge detection method according to claim 1, characterized in that: The step of loading the pre-processed blood smear image into a lightweight segmentation network to obtain a segmented image specifically includes: Loading the preprocessed blood smear image into a lightweight segmentation network; Using a neural network to extract image features from the preprocessed blood smear image; According to the extracted image features, a segmented image is obtained.

4. The blood smear edge detection method according to claim 1, characterized in that: After the step of loading the pre-processed blood smear image into the lightweight segmentation network to obtain the segmented image, the method further includes: Determine the total area of ​​the blood smear image and the area of ​​each region of the blood smear image within the field of view according to the segmented image; Selecting a target area from each of the areas; According to the target area area and the total area of ​​the blood smear image in the field of view, the target area area ratio in the segmented image is determined.

5. The blood smear edge detection method according to claim 4, characterized in that: The step of selecting the target area from the areas of the regions specifically includes: sorting the areas of the regions of the blood smear image; According to the sorting results, the area of ​​the target region is determined.

6. The blood smear edge detection method according to claim 1, characterized in that: After the step of determining that the segmented image in the field of view is at the edge of the blood smear when the area ratio of the target region in the segmented image is greater than a preset threshold, the method further includes: When the area ratio of the target region in the segmented image is less than a preset threshold, it is determined that the segmented image in the field of view is inside the blood smear.

7. A blood smear edge detection device, characterized in that: The device comprises: A processing module, used for preprocessing the blood smear image to obtain a preprocessed blood smear image; A segmentation module, used for loading the preprocessed blood smear image into a lightweight segmentation network to obtain a segmented image; The determination module is used to determine that the segmented image in the field of view is at the edge of the blood smear when the area ratio of the target region in the segmented image is greater than a preset threshold.

8. A blood smear edge detection device, characterized in that: The device comprises: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the blood smear edge detection method according to any one of claims 1 to 6.

9. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the blood smear edge detection method according to any one of claims 1 to 6 are implemented.

10. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the steps of the blood smear edge detection method according to any one of claims 1 to 6 are implemented.

Images