CN113962975B - System for carrying out quality evaluation on pathological slide digital image based on gradient information - Google Patents
System for carrying out quality evaluation on pathological slide digital image based on gradient information Download PDFInfo
- Publication number
- CN113962975B CN113962975B CN202111264905.6A CN202111264905A CN113962975B CN 113962975 B CN113962975 B CN 113962975B CN 202111264905 A CN202111264905 A CN 202111264905A CN 113962975 B CN113962975 B CN 113962975B
- Authority
- CN
- China
- Prior art keywords
- target
- quality
- image
- digital image
- examples
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 230000001575 pathological effect Effects 0.000 title claims abstract description 45
- 238000013441 quality evaluation Methods 0.000 title claims abstract description 13
- 238000001514 detection method Methods 0.000 claims abstract description 149
- 230000007170 pathology Effects 0.000 claims abstract description 115
- 238000012545 processing Methods 0.000 claims abstract description 31
- 238000011156 evaluation Methods 0.000 claims abstract description 18
- 238000001303 quality assessment method Methods 0.000 claims abstract description 5
- 238000000605 extraction Methods 0.000 claims description 15
- 238000000034 method Methods 0.000 description 34
- 238000010586 diagram Methods 0.000 description 12
- 238000004590 computer program Methods 0.000 description 8
- 238000003745 diagnosis Methods 0.000 description 6
- 230000001413 cellular effect Effects 0.000 description 4
- 238000007781 pre-processing Methods 0.000 description 4
- 230000011218 segmentation Effects 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 230000003628 erosive effect Effects 0.000 description 3
- 230000010339 dilation Effects 0.000 description 2
- 238000007689 inspection Methods 0.000 description 2
- 239000007788 liquid Substances 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 206010008342 Cervix carcinoma Diseases 0.000 description 1
- 208000006105 Uterine Cervical Neoplasms Diseases 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 238000004422 calculation algorithm Methods 0.000 description 1
- 201000010881 cervical cancer Diseases 0.000 description 1
- 239000002131 composite material Substances 0.000 description 1
- 238000004195 computer-aided diagnosis Methods 0.000 description 1
- 238000012790 confirmation Methods 0.000 description 1
- 230000002950 deficient Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 208000037765 diseases and disorders Diseases 0.000 description 1
- 239000000834 fixative Substances 0.000 description 1
- 239000011521 glass Substances 0.000 description 1
- 239000012535 impurity Substances 0.000 description 1
- 230000003902 lesion Effects 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 238000010186 staining Methods 0.000 description 1
- 238000003786 synthesis reaction Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/20—Image enhancement or restoration using local operators
- G06T5/30—Erosion or dilatation, e.g. thinning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/187—Segmentation; Edge detection involving region growing; involving region merging; involving connected component labelling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/194—Segmentation; Edge detection involving foreground-background segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10056—Microscopic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30024—Cell structures in vitro; Tissue sections in vitro
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30004—Biomedical image processing
- G06T2207/30096—Tumor; Lesion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30168—Image quality inspection
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- General Health & Medical Sciences (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Radiology & Medical Imaging (AREA)
- Quality & Reliability (AREA)
- Health & Medical Sciences (AREA)
- Investigating Or Analysing Biological Materials (AREA)
- Image Analysis (AREA)
Abstract
The present disclosure describes a system for quality assessment of pathology slide digital images based on gradient information, comprising: the device comprises an acquisition module, a detection processing module, a quality detection module and a comprehensive evaluation module; the detection processing module is configured to acquire a plurality of first target areas and characteristic information thereof from the pathological slide digital image; the quality detection module is configured to select a plurality of first target areas as second target areas, expand the second target areas based on the characteristic information, divide the expanded second target areas into a plurality of image blocks, acquire gradient information and quality grades of the image blocks, and respectively acquire the quality grades of the expanded second target areas based on the quality grades of the image blocks; the comprehensive evaluation module is configured to perform a quality evaluation on the pathology slide digital image based on the quality rating. Under the condition, the computing resources can be saved, and the digital image quality of the pathological slide can be detected more efficiently and more comprehensively.
Description
The application is filed as20 days 1 month in 2021Application No. is202110078230XThe invention is named asFor diseases and disorders System and method for detecting digital image quality of glass sheetDivisional application of the patent application.
Technical Field
The present disclosure relates specifically to a system for quality assessment of pathology slide digital images based on gradient information.
Background
Liquid-based cytology, a branch of cytopathology, is the collection of cell samples into liquid fixative solutions that, after staining, can be used for observation and diagnosis. A common application scenario includes cervical sectioning for screening for pre-cancerous cervical lesions that may lead to cervical cancer.
The development of computer-aided diagnosis in recent years reduces the workload of pathologists and also improves the efficiency. However, due to the influence of the focusing and synthesis of the scanner, the scanned digital image of the pathological slide often has poor image quality, such as the cells cannot be clearly focused in the whole or single visual field. For such digital images of pathological slides, the accuracy of computer diagnostic results is significantly reduced. At present, the method for detecting whether a digital image of a pathological slide is qualified is mainly based on the analysis of the whole digital image of the pathological slide or the sampling analysis of fixed orientation points of the digital image of the pathological slide.
However, the existing analysis of the whole pathological slide digital image is often very computationally expensive and inefficient; the sampling analysis of fixed orientation point taking often has the problem of false detection or missing detection.
Disclosure of Invention
The present disclosure is made in view of the above-mentioned situation, and an object of the present disclosure is to provide a system and a method for detecting digital image quality of a pathological slide, which can save computing resources and can detect digital image quality of a pathological slide more efficiently and more comprehensively.
To this end, a first aspect of the present disclosure provides a system for digital image quality detection of a pathological slide, characterized by comprising: the device comprises an acquisition module, a detection processing module, a quality detection module and a comprehensive evaluation module; wherein the acquisition module is used for acquiring the pathology slide digital image, and the pathology slide digital image is provided with an effective area containing contents; the detection processing module is configured to intercept a plurality of target images from the pathological slide digital image, and perform target detection on the plurality of target images, wherein the target detection comprises respectively acquiring a plurality of first target regions in each target image and characteristic information corresponding to each first target region on the basis of a characteristic extraction model, and the characteristic information at least comprises position information of the first target region, and the category and the confidence coefficient of contents in the first target region; the quality detection module is configured to sequentially select a plurality of first target areas from the plurality of target images as second target areas according to a descending order based on the confidence degrees in the feature information, and perform quality detection on each second target area respectively, wherein the quality detection includes expanding the second target areas based on the position information and a preset target proportion in the feature information, dividing the expanded second target areas into a plurality of image blocks and obtaining gradient information of each image block, dividing the quality levels of the image blocks based on the gradient information, and obtaining the quality levels of each expanded second target area based on the quality levels of each image block; the comprehensive evaluation module is configured to perform a quality evaluation on the pathology slide digital image based on the quality level of each enlarged second target region.
In the disclosure, an acquisition module is used for acquiring a pathological slide digital image, a detection processing module is used for intercepting a plurality of target images from the pathological slide digital image, the target detection is performed on the plurality of target images to acquire a first target area and characteristic information thereof, a quality detection module is used for selecting a plurality of first target areas from the plurality of target images to be used as second target areas, the quality detection is respectively performed on each second target area to acquire a quality grade corresponding to each second target area, and a comprehensive evaluation module is used for evaluating the quality of the whole pathological slide digital image. Under the condition, the computing resources can be saved, and the digital image quality of the pathological slide can be detected more efficiently and more comprehensively.
Additionally, in the system according to the first aspect of the present disclosure, optionally, the acquisition module acquires a plurality of pathology slide digital images of different resolutions, and the acquisition module preprocesses the plurality of pathology slide digital images to identify a valid region within any one of the pathology slide digital images. Thereby, the effective area within the pathology slide digital image can be determined.
In addition, in the system according to the first aspect of the present disclosure, optionally, in the preprocessing, a pathology slide digital image having a reference resolution, which may be smaller than the target resolution, is selected from a plurality of pathology slide digital images as a reference image and a pathology slide digital image having a target resolution is selected as a target slice image, an effective area of the reference image is acquired based on the reference image, and the effective area of the reference image is mapped to the target slice image to determine an effective area of the target slice image. In this case, the effective region of the target slice image can be confirmed, and the amount of calculation can be effectively reduced.
In addition, in the system according to the first aspect of the present disclosure, optionally, the detection processing module processes the pathology slide digital image according to a preset size based on an effective area of the pathology slide digital image by using a sliding window method to cut out a plurality of target images from the pathology slide digital image, wherein a window of the sliding window method has a half of the preset size as a sliding distance of the window. Thereby, a plurality of target images can be cut out from the pathology slide digital image.
In addition, in the system according to the first aspect of the present disclosure, optionally, a combination of image areas corresponding to the plurality of target images covers at least the effective area. Thereby, the effective area of the pathology slide digital image can be detected.
Further, in the system according to the first aspect of the present disclosure, optionally, the gradient information includes a gradient standard deviation of the image blocks, and the quality detection module evaluates quality levels of the respective image blocks based on the gradient standard deviation, where the quality levels include 3 levels, blank, blurred, and sharp, respectively. This makes it possible to confirm the quality level of each image block more specifically.
In addition, in the system according to the first aspect of the present disclosure, optionally, if the gradient standard deviation of an image block is not less than a first preset threshold, the quality detection module divides the quality level of the image block into clear, if the gradient standard deviation of the image block is less than a third preset threshold, the quality detection module calculates a saturation corresponding to the image block, if the saturation of the image block is less than a second preset threshold, the quality detection module divides the quality level of the image block into blank, where the third preset threshold is not greater than the first preset threshold. Thus, the quality level of each image block can be clearly confirmed.
A second aspect of the present disclosure provides a method for pathological slide digital image quality detection, comprising: acquiring the pathology slide digital image, wherein the pathology slide digital image is provided with an effective area containing contents; intercepting a plurality of target images from the pathological slide digital image, and carrying out target detection on the plurality of target images, wherein the target detection comprises respectively obtaining a plurality of first target areas in each target image and characteristic information corresponding to each first target area on the basis of a characteristic extraction model, and the characteristic information at least comprises position information of the first target area, and the category and the confidence coefficient of contents in the first target area; and sequentially selecting a plurality of first target areas from the plurality of target images as second target areas according to a descending order based on the confidence degrees in the feature information, and respectively performing quality detection on each second target area, wherein the quality detection comprises expanding the second target areas based on the position information and a preset target proportion in the feature information, dividing the expanded second target areas into a plurality of image blocks and acquiring gradient information of each image block, dividing the quality grades of the image blocks based on the gradient information, respectively acquiring the quality grades of each expanded second target area based on the quality grades of each image block, and performing quality evaluation on the pathological slide digital image based on the quality grades of each expanded second target area.
In the disclosure, a digital image of a pathological slide may be acquired, a plurality of target images may be captured from the digital image of the pathological slide, target detection may be performed on the plurality of target images to acquire a first target region and characteristic information thereof, a plurality of first target regions may be selected from the plurality of target images as second target regions, quality detection may be performed on each of the second target regions to acquire a quality grade corresponding to each of the second target regions, and quality evaluation may be performed on the entire digital image of the pathological slide. Under the condition, the computing resources can be saved, and the digital image quality of the pathological slide can be detected more efficiently and more comprehensively.
A third aspect of the present disclosure provides a computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the above-mentioned method when executing the computer program.
A fourth aspect of the present disclosure provides a computer-readable storage medium, on which a computer program is stored, wherein the computer program, when executed by a processor, implements the steps of the method described above.
According to the present disclosure, a system for pathological slide digital image quality detection, a method thereof, a computer device and a computer-readable storage medium thereof are provided, which can save computing resources and can detect pathological slide digital image quality more efficiently and more comprehensively.
Drawings
Embodiments of the present disclosure will now be explained in further detail by way of example only with reference to the accompanying drawings, in which:
fig. 1 is a diagram illustrating an application scenario of a system for pathology slide digital image quality detection according to an example of the present disclosure.
Fig. 2 is a block diagram showing a system according to an embodiment of the present disclosure.
Fig. 3 is a schematic diagram illustrating a digital image of a pathology slide according to an example of the present disclosure.
Fig. 4 is a schematic diagram illustrating target image acquisition in accordance with an example of the present disclosure.
Fig. 5 is a schematic diagram illustrating a first target region within a target image in accordance with an example of the present disclosure.
Fig. 6 is a flow diagram illustrating a method for pathology slide digital image quality detection in accordance with an example of the present disclosure.
Detailed Description
Hereinafter, preferred embodiments of the present disclosure will be described in detail with reference to the accompanying drawings. In the following description, the same components are denoted by the same reference numerals, and redundant description thereof is omitted. The drawings are schematic and the ratio of the dimensions of the components and the shapes of the components may be different from the actual ones.
It is noted that the terms "comprises," "comprising," and "having," and any variations thereof, in this disclosure, for example, a process, method, system, article, or apparatus that comprises or has a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include or have other steps or elements not expressly listed or inherent to such process, method, article, or apparatus. All methods described in this disclosure can be performed in any suitable order unless otherwise indicated herein or otherwise clearly contradicted by context.
The present disclosure discloses a system 10 for pathology slide digital image 40 quality detection. Quality inspection of the pathology slide digital image 40 may be performed by the system 10 of the present disclosure, for example, the degree of clarity of the pathology slide digital image 40 may be inspected, and the like. Therefore, the follow-up medical staff can conveniently diagnose the examinee. In some examples, the pathology slide digital image 40 may be a cellular image acquired by scanning a cellular slide made of cells of a subject by a scanner (see fig. 1 and 3). For example, the pathology Slide digital image 40 may be a Whole Slide Image (WSI), and the WSI image is generally very large, for example, the size of the WSI image may be 600Mb to 10Gb, so that the conventional quality detection method is generally not suitable for processing the WSI image. By the system 10 for detecting the quality of the pathological slide digital image 40, computing resources can be saved, and the quality of the pathological slide digital image 40 can be detected more efficiently and more comprehensively.
The system 10 according to the present disclosure is described in detail below with reference to the drawings.
Fig. 1 is a diagram illustrating an application scenario of a system 10 for quality inspection of a digital image 40 of a pathological slide according to an embodiment of the present disclosure. Fig. 2 is a block diagram showing the structure of the system 10 according to the embodiment of the present disclosure.
In some examples, system 10 may include an acquisition module 110, a detection processing module 120, and a quality detection module 130 (see fig. 1 and 2). The acquisition module 110 can be used to acquire a digital image 40 of a pathology slide, among other things. The detection processing module 120 and the quality detection module 130 can process and detect the pathology slide digital image 40, etc. For example, referring to fig. 1, the acquisition device 20 (e.g., a scanner) can perform a high resolution scan of the slide 30 to acquire a pathology slide digital image 40. The pathology slide digital image 40 may be uploaded to the system 10 after the scan is complete. The system 10 may implement the processing and detection of the pathology slide digital image 40 by executing computer program instructions, for example, the system 10 may be used to detect the quality of the pathology slide digital image 40, and the like.
In some examples, the acquisition module 110 may be used to acquire a pathology slide digital image 40 (see fig. 2), as described above. In some examples, the acquisition module 110 may be used to acquire a plurality of pathology slide digital images 40 of different resolutions. In some examples, the acquisition module 110 may acquire a slice image that may contain a plurality of pathology slide digital images 40 of different resolutions. For example, referring to fig. 1, the acquisition device 20 (e.g., a scanner) may separately perform high resolution scans of the entire slide 30 based on different multiples to acquire a plurality of pathology slide digital images 40. The plurality of pathology slide digital images 40 may be sorted by resolution to form a pyramid-form slice image. In general, the resolution of the bottom most pathology slide digital image of the pyramid is the greatest and the resolution of the top most pathology slide digital image of the pyramid is the least, e.g., the top most pathology slide digital image may correspond to a thumbnail of a cell image of the slide 30. The slice images may be uploaded to the system 10 after the scan is complete. The slice images may be acquired by an acquisition module 110 of the system 10.
Fig. 3 is a schematic diagram illustrating a pathology slide digital image 40 according to an example of the present disclosure.
In some examples, referring to fig. 3, the pathology slide digital image 40 may have an active area containing contents. In some examples, the pathology slide digital image 40 may also have a background region that is distinct from the active region. In some examples, the contents may be various types of cells (e.g., contents 410 in fig. 3, etc.). In some examples, the pathology slide digital image 40 may be a grayscale image. In other examples, the pathology slide digital image 40 may be a color image.
In some examples, the acquisition module 110 may pre-process the pathology slide digital images 40 to confirm a valid region within any of the pathology slide digital images 40. Thereby, the effective area within the pathology slide digital image 40 can be determined. In some examples, in the preprocessing, the acquisition module 110 may acquire a plurality of pathology slide digital images 40 of different resolutions. In some examples, the acquisition module 110 may select a pathology slide digital image having a reference resolution as a reference image and a pathology slide digital image having a target resolution as a target slice image from the plurality of pathology slide digital images 40. In some examples, the reference resolution may be less than the target resolution. In some examples, the reference resolution may be a corresponding minimum resolution of the plurality of pathology slide digital images 40. The target resolution may be the highest corresponding resolution of the plurality of pathology slide digital images 40. For example, in the preprocessing, the acquisition module 110 may select a digital image of a pathological slide having a reference resolution as a reference image and a digital image of a pathological slide having a target resolution as a target slice image from the slice images. Wherein the pathology slide digital image having the reference resolution may be a thumbnail contained in the slice image. The pathology slide digital image with the target resolution may be the pathology slide digital image with the highest resolution in the slice images.
In some examples, a valid region corresponding to a reference picture may be identified from the reference picture based on the reference picture. In some examples, the active area of the reference image may be mapped to the target slice image to determine the active area of the target slice image. In this case, the effective region of the target slice image can be confirmed, and the amount of calculation can be effectively reduced.
In some examples, in acquiring the active area of the reference image, the reference image may be converted into a reference gray-scale image in a gray-scale mode, the reference gray-scale image may be adaptively threshold-divided and color-inverted using a binarization threshold segmentation algorithm (e.g., the large attorney approach (OTSU)) to acquire a reference slice image, the reference slice image may be subjected to an expansion and erosion process to acquire a white area, and the white area may be used as the active area of the reference image. Thereby, the effective area of the reference image can be determined. In some examples, the dilation and erosion process on the reference slice image may obtain a binary segmentation image containing white and black regions. For example, the reference slice image may be subjected to 2 dilation and 2 erosion operations to acquire a binary segmentation image containing white and black regions. In some examples, the black region may be a background region of the reference image. In some examples, the reference grayscale image may be denoised (e.g., median blurred) prior to adaptive threshold segmentation of the reference grayscale image.
In some examples, the active area of the reference image may be mapped to the target slice image to determine the active area of the target image. Specifically, a circumscribed rectangle of the effective region of the reference image may be acquired, and a circumscribed rectangle corresponding to the effective region of the target slice image may be acquired based on a reduction factor of the reference image with respect to the target slice image. In some examples, a circumscribed rectangle corresponding to the effective region of the target slice image may be taken as the effective region of the target slice image. In some examples, the circumscribed rectangle may be increased by 5% to 10% as the effective area of the target slice image. This enables effective confirmation of the effective region corresponding to the target slice image.
Examples of the present disclosure are not limited thereto, and in some examples, the acquisition module 110 may acquire the effective area of the pathology slide digital image 40 in other ways in the preprocessing.
In some examples, if the acquisition module 110 acquires multiple digital images 40 of a pathology slide at different resolutions, the acquisition module 110 may acquire a target region corresponding to a digital image of a pathology slide (i.e., a target slice image) having a target resolution. In some examples, subsequent modules of the system 10 (e.g., the detection processing module 120) may process the pathology slide digital image having the target resolution. In some examples, the target resolution may be confirmed according to actual circumstances. In some examples, if the acquisition module 110 acquires only a single pathology slide digital image 40, the acquisition module 110 may acquire a target region of the pathology slide digital image 40. In this case, the acquisition module 110 can treat the pathology slide digital image 40 as a pathology slide digital image having the target resolution.
In some examples, as described above with reference to fig. 2, system 10 may also include detection processing module 120. In some examples, the detection processing module 120 may be used to process the pathology slide digital image 40. For example, the detection processing module 120 may receive the digital image 40 of the pathological slide processed by the acquisition module 110 for processing. In some examples, the detection processing module 120 may know the effective area within the pathology slide digital image 40 via the acquisition module 110.
Fig. 4 is a schematic diagram illustrating target image acquisition in accordance with an example of the present disclosure.
In some examples, the detection processing module 120 may intercept a plurality of target images (see fig. 4, e.g., target image a and target image B, etc.) from the pathology slide digital image 40. In some examples, the combination of image regions corresponding to the plurality of target images may cover at least the effective area of the pathology slide digital image 40. This enables detection of the effective region of the pathology slide digital image 40. In some examples, the detection processing module 120 may process the pathology slide digital image 40 to a preset size based on an effective area of the pathology slide digital image 40 using a sliding window method to cut a plurality of target images from the pathology slide digital image 40. Specifically, the sliding window may intercept the target image from the pathological slide digital image 40 according to a preset size, that is, the intercepted target image may be a preset size; one half of the preset size can be used as the sliding distance of the window, and the window is slid along the transverse direction and the longitudinal direction of the effective area of the target image according to the sliding distance; the corresponding image of the slid window on the pathology slide digital image 40 may be used as the target image. For example, the sliding window may capture the target image from the pathology slide digital image 40 in a preset size of 1024 × 1024, and the sliding distance of the window in the transverse direction and the longitudinal direction may be 512, respectively. Thereby, a plurality of target images can be cut out from the pathology slide digital image 40.
However, the examples of the present disclosure are not limited thereto, and in other examples, the plurality of target images may be directly cut out from the pathology slide digital image 40 for the pathology slide digital image 40 without acquiring the effective area of the pathology slide digital image 40.
Fig. 5 is a schematic diagram illustrating a first target region within a target image in accordance with an example of the present disclosure.
In some examples, the detection processing module 120 may also be used for object detection on multiple object images. In some examples, the target detection may include obtaining a plurality of first target regions within each target image (see fig. 5, for example, a first target region a may be obtained from within the target image a, etc.), and feature information corresponding to each first target region, respectively, based on the feature extraction model. In some examples, the feature extraction model may be an object detection network. In some examples, the feature extraction model may be an Object detection network based on an Efficient and Efficient Object detection architecture. In some examples, the characteristic information may include at least location information of the corresponding first target region, a category and confidence of contents within the first target region, and the like. In some examples, the detection processing module 120 may obtain feature information corresponding to a plurality of first target regions. The detection processing module 120 may know the first target region according to the position information in the feature information.
In some examples, the feature extraction model may be trained. In some examples, the feature extraction model may be configured to screen out regions that may contain content (i.e., the first target region) from within the respective target images.
In some examples, the number of the plurality of first target regions (simply "first number") within each target image may be determined by the number of contents in the target image. In some examples, one target image may correspond to a plurality of first target regions. In some examples, the first number may be confirmed by the precision of the feature extraction model. In some examples, the number of cells, impurities, etc. in the target image may affect the feature extraction model to screen the first target region.
In some examples, as described above with reference to fig. 2, system 10 may also include a quality detection module 130. In some examples, the quality detection module 130 may obtain a plurality of first target regions corresponding to respective target images and feature information corresponding to the respective first target regions from the detection processing module 120. In some examples, the quality detection module 130 may process the first target region based on the characteristic information.
In some examples, the quality detection module 130 may select several first target regions from the plurality of target images as the second target regions based on the feature information corresponding to the respective first target regions. In some examples, the quality detection module 130 may select several first target regions from a plurality of target images (e.g., all target images) in order of increasing confidence in the feature information as the second target regions, respectively.
In some examples, the number of second target regions may be confirmed via the number of the plurality of first target regions within each target image and a preset number. In some examples, the number of second target areas may be not less than the preset number. This enables the quality of the pathology slide digital image 40 to be detected relatively comprehensively. In some examples, if the number of the plurality of first target regions is not less than the preset number, the quality detection module 130 may select a preset number of first target regions from the plurality of target images as the second target regions respectively. In some examples, if the number of the plurality of first target regions is less than the preset number, the quality detection module 130 may select the first number of first target regions from the plurality of target images as the second target regions, respectively. In some examples, the preset number may be determined at the discretion of the skilled artisan. In some examples, the predetermined number may be 0 to 200. For example, the predetermined number may be 50, 75, 100, 125, 150, 175, or 200, etc.
In some examples, the quality detection module 130 may perform quality detection on the respective second target regions. In this case, the quality detection module 130 can obtain the quality level corresponding to each second target area. For example, the quality detection module 130 may obtain a quality level (described in detail later) of each of the enlarged second target regions.
In some examples, the quality detection module 130 may further expand each second target area based on the position information in the feature information and a preset target ratio. Thus, the quality detection module 130 can process the enlarged second target region conveniently. Specifically, the quality detection module 130 may enlarge each second target region according to a preset target ratio based on the position information of the second target region. That is, the area size of each second target area can be enlarged respectivelyAnd obtaining the second expanded target area from the preset target proportion of the area. In some examples, the preset target proportion may be 0-50%. For example, the preset target ratio may be 10%, 12%, 14%, 16%, 18%, 20%, 22%, 24%, 26%, 28%, 30%, 40%, or 50%, etc. For example, the position information of a certain second target region may be (X) min ,Y min ,X max ,Y max ) (ii) a The quality detection module 130 may expand the second target region by 20%, and may obtain the position information (X) of the expanded second target region min ’,Y min ’,X max ’,Y max '). For example, the position information (X) of the enlarged second target region min ’,Y min ’,X max ’,Y max ') can satisfy: x min ’＝X min -20％*(X max -X min )；Y min ’＝Y min ；X max ’＝X max ；Y max ’＝Y max ；
In some examples, the quality detection module 130 may divide each enlarged second target area into a plurality of image blocks, respectively. For example, the quality detection module 130 may divide each of the expanded second target regions into a plurality of 40 × 40 image blocks. Examples of the present disclosure are not limited thereto, and in some examples, the quality detection module 130 may divide the second target area into a plurality of image blocks, respectively. In this case, the quality detection module 130 may perform subsequent processing on the plurality of image blocks obtained from the second target area that is not enlarged. Reference may be made specifically to the following description of the processing of the plurality of image blocks obtained from the enlarged second target area by the subsequent quality detection module 130.
In some examples, the quality detection module 130 may also obtain relevant information for the individual image blocks to perform quality assessment on the individual image blocks.
In general, a sharply focused image may have richer gradient information, i.e., its corresponding gradient standard deviation may be higher, than an unfocused image. An image block with less gradient information (i.e., a smaller gradient standard deviation) may be due to being out of focus or due to the image block being a blank area. In the embodiment according to the present disclosure, the saturation of the image block may be obtained, and if the saturation is close to 0, the image block may be a blank area.
In some examples, the quality detection module 130 may obtain gradient information corresponding to each tile, and the quality detection module 130 may evaluate the sharpness (i.e., the quality level of each tile) of each tile based on the gradient information (e.g., the standard deviation of the gradient) of each tile.
In some examples, the quality level may be divided into 3 levels, blank, fuzzy, and clear, respectively. This makes it possible to confirm the quality level of each image block more specifically. For example, if the quality level is clear, it indicates that the partial image is clear. If the quality level is fuzzy, it indicates that the partial image may be fuzzy due to factors such as unclear focusing. If the quality grade is blank, the partial image is a blank area. Examples of the present disclosure are not limited thereto, and in some examples, the quality level may be divided in more detail by a person skilled in the relevant art.
In some examples, the quality detection module 130 may obtain gradient information corresponding to each image patch. In some examples, the gradient information may include a gradient standard deviation. In this case, the quality detection module 130 may evaluate the quality level of the image patch according to the gradient standard deviation. In some examples, the quality detection module 130 may rank the quality of the respective image blocks according to the gradient standard deviation and a preset threshold. In some examples, the quality detection module 130 may evaluate whether the quality level of the image block is clear according to the gradient standard deviation and a preset first preset threshold. For example, if the standard deviation of the gradient is not less than the first preset threshold, the quality level of the image block is clear. And if the gradient standard deviation is smaller than a first preset threshold, the quality grade of the image block is fuzzy or blank.
In some examples, the quality detection module 130 may also classify a quality level of an image block according to its saturation. In some examples, the quality detection module 130 may also obtain the saturation corresponding to each image block. The quality detection module 130 may treat the image blocks with the saturation less than the second preset threshold as the blank area. That is, if the saturation is smaller than the second preset threshold, the quality detection module 130 may evaluate the quality level of the image block as blank. That is, if the saturation is smaller than the second preset threshold, the quality detection module 130 may divide the quality level of the image block into blanks.
In other examples, the quality detection module 130 may classify the quality level of the image patch according to the gradient information and the saturation of the image patch. In some examples, the quality detection module 130 may obtain gradient information (e.g., a standard deviation of gradients) corresponding to each image patch. The quality detection module 130 may obtain the saturation corresponding to the image block whose gradient standard deviation is smaller than the third preset threshold. In some examples, the quality detection module 130 may evaluate whether the quality level of the image block is blank based on the saturation and a second preset threshold. In some examples, the quality levels of other image blocks having a gradient standard deviation not greater than the third preset threshold may be blurred except for image blocks having a blank quality level. In some examples, the third preset threshold may not be greater than the first preset threshold. For example, the third preset threshold may be equal to the first preset threshold. In some examples, the quality level of an image block with a gradient standard deviation between the third preset threshold and the first preset threshold may be blurred. Thus, the quality level of each image block can be clearly confirmed.
In some examples, the first preset threshold, the second preset threshold, and the third preset threshold may be set by a person skilled in the relevant art. In some examples, the first preset threshold may be selected from 40-60. For example, the first preset threshold may be 40, 45, 50, 55, 60, or the like. In some examples, the second preset threshold may be selected from 5 to 15. For example, the second preset threshold may be 10. In this case, if the saturation of an image block is less than 10, the quality level of the image block may be blank. In some examples, the third preset threshold may be selected from 10-60.
In some examples, the quality detection module 130 may obtain each enlarged second target region separately, as described above. The quality detection module 130 may further obtain a plurality of image blocks corresponding to each expanded second target area, and the quality detection module 130 may further obtain quality levels of each image block. That is, the quality detection module 130 may obtain the quality levels of the image blocks included in each of the expanded second target areas.
In some examples, the quality detection module 130 may obtain the quality level of each enlarged second target area based on the quality levels of the plurality of image blocks within the enlarged second target area. In some examples, the quality detection module 130 may evaluate the quality level of the enlarged second target region based on a ratio of the number of image blocks of each quality level within the enlarged second target region.
In some examples, the quality detection module 130 may obtain the quality levels of the respective image blocks in any expanded second target area, and the quality detection module 130 may count the number of image blocks with blank quality levels in the expanded second target area to account for a first percentage of the total number of image blocks.
In some examples, the quality detection module 130 may screen out image blocks with blank quality levels in the enlarged second target area. The number of the image blocks other than the image block whose quality level is blank may be the target number. The quality detection module 130 may count a second percentage of the number of image blocks with quality levels that are blurred to the target number.
In some examples, the quality detection module 130 may obtain whether the quality level of the enlarged second target area is blank based on the first percentage and a preset first preset ratio.
In some examples, the quality detection module 130 may obtain whether the quality level of the enlarged second target region is fuzzy based on the second percentage and a second preset ratio.
In some examples, the quality level of the enlarged second target region is clear if the quality detection module 130 evaluates that the quality level of the enlarged second target region is not blank and fuzzy.
For example, the quality detection module 130 may first obtain a first percentage, compare the first percentage with a first preset ratio, and if the first percentage is greater than the first preset ratio, indicate that the quality level of the enlarged second image area is blank; otherwise, the quality grade of the enlarged second image area is not blank. If the second image area is not blank, the quality detection module 130 may screen the enlarged image block with the blank quality level in the second image area to obtain a second percentage. If the second percentage is larger than a second preset proportion, the quality grade of the enlarged second image area is fuzzy; otherwise, the quality grade of the enlarged second image area is clear.
In some examples, the first preset ratio and the second preset ratio may be set by a person skilled in the art. In some examples, the first preset proportion may be 40-60%. In some examples, the second predetermined proportion may be 20-50%. For example, the first preset proportion may be 45%, 50%, 55%, or the like. The second predetermined proportion may be 25%, 30%, 35%, 40%, 45% or 50% etc.
In some examples, as described above, the quality detection module 130 may obtain the quality level corresponding to each enlarged second target region.
In some examples, referring to fig. 2, system 10 may also include a composite evaluation module 140. In some examples, the comprehensive evaluation module 140 may be used to perform a quality evaluation of the pathology slide digital image 40. In some examples, the integrated evaluation module 140 may only know the quality level corresponding to each enlarged second target region within the pathology slide digital image 40 by the quality detection module 130. In some examples, the comprehensive evaluation module 140 may perform a quality evaluation of the pathology slide digital image 40 based on the quality rating of each enlarged second target region. Therefore, the quality of the pathology slide digital image 40 can be detected more efficiently and more comprehensively.
In some examples, the comprehensive evaluation module 140 may count the number of enlarged second target regions with clear quality levels within the pathology slide digital image 40. In some examples, the comprehensive evaluation module 140 may count a third percentage of the number of enlarged second target regions with clear quality ratings to the total number of second target regions within the pathology slide digital image 40. In some examples, the comprehensive assessment module 140 can perform a quality assessment on the pathology slide digital image 40 based on a third percentage and a third preset ratio that is preset. In some examples, the comprehensive evaluation module 140 may know the quality level of each enlarged second target region within the pathology slide digital image 40 and may statistically obtain the third percentage. The comprehensive evaluation module 140 may evaluate the quality of the pathology slide digital image 40 based on a comparison of the third percentage and a third preset ratio. If the third percentage is not less than the third predetermined ratio, the pathology slide digital image 40 is qualified, i.e., the pathology slide digital image 40 is clear. In this case, the medical staff can diagnose the subject using the pathology slide digital image 40. If the third percentage is less than the third predetermined percentage, the digital image 40 of the pathological slide is not qualified. In this case, a qualified digital image 40 of the pathology slide is acquired again. Thereby, the quality evaluation can be performed on the pathology slide digital image 40.
In embodiments of the present disclosure, system 10 may be directly embedded in an existing diagnostic assistance system. The existing auxiliary diagnosis system can carry out target detection on the cell slice system. Specifically, the existing auxiliary diagnosis system can perform target detection on the digital image of the pathological slide, the quality detection module 130 and the comprehensive evaluation module 140 can be embedded into the auxiliary diagnosis system, and after the auxiliary diagnosis system performs target detection on the digital image of the pathological slide, the quality of the digital image of the pathological slide can be evaluated through the embedded quality detection module 130 and the comprehensive evaluation module 140. In this case, the auxiliary diagnosis system can perform quality evaluation on the pathology slide digital image. This can effectively reduce the possibility of wasting the calculation resource on the defective image.
The method for detecting digital image quality of pathological slide according to the present disclosure is described in detail below with reference to the accompanying drawings. Fig. 6 is a flow diagram illustrating a method for pathology slide digital image quality detection in accordance with an example of the present disclosure.
In the present embodiment, referring to fig. 6, the method for digital image quality detection of a pathology slide may include the steps of: acquiring a pathology slide digital image 40 (step S10); cutting a plurality of target images from the pathology slide digital image 40, and performing target detection on the plurality of target images (step S20); selecting a plurality of first target areas from the plurality of target images as second target areas based on the feature information obtained by the target detection, and performing quality detection on each second target respectively (step S30); the pathology slide digital images 40 are quality evaluated based on the quality levels of the respective enlarged second target regions obtained via the quality detection (step S40). According to the method for detecting the digital image quality of the pathological slide, the computing resources can be saved, and the digital image quality of the pathological slide can be detected more efficiently and more comprehensively.
In the present embodiment, the acquisition or processing of the pathology slide digital image 40, the target image, the target detection, the feature information, the first target region, the second target region, the quality detection, the quality grade and the quality evaluation in the method can be referred to the above-mentioned description of the pathology slide digital image, the target detection, the feature information, the first target region, the second target region, the quality detection, the quality grade and the quality evaluation.
In step S10, the pathology slide digital image 40 may be acquired, as described above.
In some examples, the pathology slide digital image 40 may be a cellular image acquired by a scanner scanning a cellular slide made of cells of a subject. In some examples, the pathology slide digital image 40 may have an active area containing contents. The obtaining and processing of the effective region may refer to the above description of the effective region. In some examples, the contents may be various types of cells.
In some examples, in step S10, a plurality of pathology slide digital images, such as slice images or the like, of different resolutions may be acquired. The acquisition and processing of the slice image can be referred to the above description of the slice image. In some examples, step S10 may be implemented by acquisition module 110 in system 10.
In step S20, as described above, a plurality of target images may be cut out from the pathology slide digital image 40, and target detection may be performed on the plurality of target images.
In some examples, a plurality of target images may be cut from the pathology slide digital image 40. Wherein, the combination of the target areas corresponding to the plurality of target images can cover at least the effective area of the pathology slide digital image 40. In some examples, the target detection may include obtaining a plurality of first target regions within each target image and feature information corresponding to each first target region, respectively, based on a feature extraction model. In some examples, the characteristic information may include at least location information of the corresponding first target region, a category and confidence of contents within the first target region, and the like. In some examples, the feature extraction model may be trained. In some examples, the feature extraction model may be configured to screen out regions that may contain content (i.e., the first target region) from within the respective target images. In the present embodiment, the obtaining and processing of the feature extraction model in the method can be referred to the above description of the feature extraction model. In some examples, step S20 may be implemented by detection processing module 120 in system 10.
In step S30, as described above, several first target regions may be selected from the plurality of target images as second target regions based on the feature information obtained through the target detection, and the quality detection may be performed on the respective second targets, respectively.
In some examples, several first target regions may be selected from the plurality of target images as the second target regions based on the feature information corresponding to the respective first target regions obtained in step S20. In some examples, several first target regions may be sequentially selected from a plurality of target images (e.g., all target images) in order from large to small based on the confidence in the feature information obtained in step S20 as the second target regions, respectively.
In some examples, the number of second target regions may be confirmed via the number of the plurality of first target regions within each target image and a preset number. In some examples, the number of second target areas may be not less than the preset number.
In some examples, the quality detection may be based on individual second target regions. Thereby, the quality level corresponding to each second target region can be obtained. In some examples, the quality levels may include 3 levels, blank, fuzzy, and clear, respectively. In some examples, step S30 may be implemented by quality detection module 130 in system 10.
In step S40, as described above, the pathological slide digital image 40 may be quality-evaluated based on the quality level of each enlarged second target region obtained via the quality detection of step S30.
In some examples, the quality level of each enlarged second target region processed in step S30 may be obtained. In some examples, the pathology slide digital image 40 may be quality evaluated based on the quality rating of the respective enlarged second target region. Therefore, the quality of the pathological slide digital image 40 can be detected more efficiently and more comprehensively. In some examples, step S40 may be implemented by the comprehensive evaluation module 140 in the system 10.
In some examples, the present disclosure also provides a computer device, which may include a memory storing a computer program and a processor implementing the steps S10-S40 of the above method when the processor executes the computer program. In some examples, the present disclosure also provides a computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements steps S10-S40 of the above method.
While the present disclosure has been described in detail above with reference to the drawings and examples, it should be understood that the above description is not intended to limit the disclosure in any way. Variations and changes may be made as necessary by those skilled in the art without departing from the true spirit and scope of the disclosure, which fall within the scope of the disclosure.
Claims (8)
1. A system for quality assessment of a pathology slide digital image based on gradient information, comprising: the device comprises an acquisition module, a detection processing module, a quality detection module and a comprehensive evaluation module; wherein the acquisition module is used for acquiring the pathological slide digital image; the detection processing module is configured to intercept a plurality of target images from the pathological slide digital image, and perform target detection on the plurality of target images, wherein the target detection comprises respectively acquiring a plurality of first target regions in each target image and characteristic information corresponding to each first target region on the basis of a characteristic extraction model, and the characteristic information at least comprises position information of the first target region, and the category and the confidence coefficient of contents in the first target region;
the quality detection module is configured to sequentially select a plurality of first target areas from the plurality of target images as second target areas based on the confidence degrees in the feature information in a descending order, expand the second target areas based on the feature information, divide the expanded second target areas into a plurality of image blocks and acquire gradient information of each image block, divide the quality levels of the image blocks based on the gradient information, and respectively acquire the quality levels of the expanded second target areas based on the quality levels of the image blocks; the comprehensive evaluation module is configured to perform a quality evaluation on the pathology slide digital image based on the quality level of each enlarged second target region.
2. The system of claim 1,
the pathology slide digital image has an effective area containing contents and a background area distinguished from the effective area.
3. The system of claim 1,
and expanding the second target area based on the position information in the characteristic information and a preset target proportion, wherein the preset target proportion is 0-50%.
4. The system of claim 1,
the gradient information comprises a gradient standard deviation of the image blocks, and the quality detection module evaluates quality grades of each image block based on the gradient standard deviation, wherein the quality grades comprise 3 grades, namely blank, fuzzy and clear.
5. The system of claim 4,
if the gradient standard deviation of the image block is not smaller than a first preset threshold, the quality detection module divides the quality grade of the image block into clear grades, if the gradient standard deviation of the image block is smaller than a third preset threshold, the quality detection module calculates the saturation corresponding to the image block, if the saturation of the image block is smaller than a second preset threshold, the quality detection module divides the quality grade of the image block into blank grades, wherein the third preset threshold is not larger than the first preset threshold.
6. The system of claim 4,
the quality detection module is configured to obtain the quality grade of each image block in any expanded second target area, count the number of image blocks with blank quality grades in the expanded second target area to account for a first percentage of the total number of the image blocks, and judge whether the quality grade of the expanded second target area is blank or not based on the first percentage and a preset first preset proportion.
7. The system of claim 6,
the quality detection module is configured to screen out image blocks with blank quality levels in the expanded second target area, the number of the image blocks except the image blocks with blank quality levels is used as a target number, the quality detection module counts a second percentage of the fuzzy quality level image block number in the target number, and the quality detection module judges whether the quality level of the expanded second target area is fuzzy or not based on the second percentage and a preset second preset proportion.
8. The system of claim 7,
and if the quality grade of the expanded second target area is not blank or fuzzy, the quality detection module judges that the quality grade of the expanded second target area is clear.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111264905.6A CN113962975B (en) | 2021-01-20 | 2021-01-20 | System for carrying out quality evaluation on pathological slide digital image based on gradient information |
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111264905.6A CN113962975B (en) | 2021-01-20 | 2021-01-20 | System for carrying out quality evaluation on pathological slide digital image based on gradient information |
| CN202110078230.XA CN112750121B (en) | 2021-01-20 | 2021-01-20 | System and method for detecting digital image quality of pathological slide |
Related Parent Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110078230.XA Division CN112750121B (en) | 2021-01-20 | 2021-01-20 | System and method for detecting digital image quality of pathological slide |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113962975A CN113962975A (en) | 2022-01-21 |
| CN113962975B true CN113962975B (en) | 2022-09-13 |
Family
ID=75652715
Family Applications (3)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111264905.6A Active CN113962975B (en) | 2021-01-20 | 2021-01-20 | System for carrying out quality evaluation on pathological slide digital image based on gradient information |
| CN202110078230.XA Active CN112750121B (en) | 2021-01-20 | 2021-01-20 | System and method for detecting digital image quality of pathological slide |
| CN202111267162.8A Active CN113962976B (en) | 2021-01-20 | 2021-01-20 | Quality evaluation method for pathological slide digital image |
Family Applications After (2)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110078230.XA Active CN112750121B (en) | 2021-01-20 | 2021-01-20 | System and method for detecting digital image quality of pathological slide |
| CN202111267162.8A Active CN113962976B (en) | 2021-01-20 | 2021-01-20 | Quality evaluation method for pathological slide digital image |
Country Status (1)
| Country | Link |
|---|---|
| CN (3) | CN113962975B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113658209B (en) * | 2021-08-12 | 2023-06-16 | 山东省计算中心（国家超级计算济南中心） | Morphology-based excellent discriminating method for mounting position of slice tissue |
| CN114187277B (en) * | 2021-12-14 | 2023-09-15 | 赛维森(广州)医疗科技服务有限公司 | Detection method for thyroid cytology multiple cell types based on deep learning |
| CN114092479B (en) * | 2022-01-21 | 2022-05-03 | 武汉大学 | Medical image evaluation method and device |
| CN114219803B (en) * | 2022-02-21 | 2022-07-15 | 浙江大学 | Detection method and system for three-stage image quality evaluation |
| CN114842000A (en) * | 2022-07-01 | 2022-08-02 | 杭州同花顺数据开发有限公司 | Endoscope image quality evaluation method and system |
Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102947858A (en) * | 2010-04-08 | 2013-02-27 | 帕斯科数码有限责任公司（商业用为奥姆尼克斯有限责任公司） | Image quality assessment including comparison of overlapped margins |
| CN103175834A (en) * | 2013-01-28 | 2013-06-26 | 宁波江丰生物信息技术有限公司 | Method and system for evaluating quality of digital pathological section |
| CN106600538A (en) * | 2016-12-15 | 2017-04-26 | 武汉工程大学 | Human face super-resolution algorithm based on regional depth convolution neural network |
| WO2018209057A1 (en) * | 2017-05-11 | 2018-11-15 | The Research Foundation For The State University Of New York | System and method associated with predicting segmentation quality of objects in analysis of copious image data |
| CN110738658A (en) * | 2019-12-21 | 2020-01-31 | 杭州迪英加科技有限公司 | Image quality evaluation method |
| CN110738160A (en) * | 2019-10-12 | 2020-01-31 | 成都考拉悠然科技有限公司 | human face quality evaluation method combining with human face detection |
| CN110853006A (en) * | 2019-11-05 | 2020-02-28 | 华南理工大学 | Method for evaluating quality of digital pathological image acquired by scanner |
| CN111179242A (en) * | 2019-12-25 | 2020-05-19 | Tcl华星光电技术有限公司 | Image processing method and device |
| CN111337495A (en) * | 2020-04-09 | 2020-06-26 | 江苏迪赛特医疗科技有限公司 | Pathological section image whole-course quality control device and pathological section analysis device |
| CN111768405A (en) * | 2020-08-05 | 2020-10-13 | 赛维森(广州)医疗科技服务有限公司 | Method, device, equipment and storage medium for processing annotated image |
| CN111899246A (en) * | 2020-07-30 | 2020-11-06 | 平安科技（深圳）有限公司 | Slide digital information quality detection method, device, equipment and medium |
| CN111986157A (en) * | 2020-07-21 | 2020-11-24 | 万达信息股份有限公司 | Digital pathological image quality evaluation system |
| CN112102247A (en) * | 2020-08-18 | 2020-12-18 | 广州金域医学检验中心有限公司 | Pathological section quality evaluation method based on machine learning and related equipment |
Family Cites Families (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6751364B2 (en) * | 2001-10-15 | 2004-06-15 | Tyson Fresh Meats, Inc. | Image analysis systems for grading of meat, predicting quality of meat and/or predicting meat yield of an animal carcass |
| EP2093697B1 (en) * | 2008-02-25 | 2017-08-23 | Telefonaktiebolaget LM Ericsson (publ) | Method and arrangement for retrieving information comprised in a barcode |
| JP5387146B2 (en) * | 2009-06-03 | 2014-01-15 | 日本電気株式会社 | Pathological tissue image analysis apparatus, pathological tissue image analysis method, pathological tissue image analysis program |
| US10083340B2 (en) * | 2016-01-26 | 2018-09-25 | Ge Healthcare Bio-Sciences Corp. | Automated cell segmentation quality control |
| US10713785B2 (en) * | 2017-02-13 | 2020-07-14 | Siemens Healthcare Gmbh | Image quality assessment system and method |
| CN107507173B (en) * | 2017-08-15 | 2021-07-27 | 上海交通大学 | No-reference definition evaluation method and system for full-slice image |
| CN107945156A (en) * | 2017-11-14 | 2018-04-20 | 宁波江丰生物信息技术有限公司 | A kind of method of automatic Evaluation numeral pathology scan image image quality |
| CN111368837B (en) * | 2018-12-25 | 2023-12-05 | 中移（杭州）信息技术有限公司 | Image quality evaluation method and device, electronic equipment and storage medium |
| CN109697719B (en) * | 2019-03-05 | 2021-12-24 | 北京康夫子健康技术有限公司 | Image quality evaluation method and device and computer readable storage medium |
| CN110110799B (en) * | 2019-05-13 | 2021-11-16 | 广州锟元方青医疗科技有限公司 | Cell sorting method, cell sorting device, computer equipment and storage medium |
| CN110376198B (en) * | 2019-08-29 | 2021-08-10 | 广州锟元方青医疗科技有限公司 | Cervical liquid-based cell slice quality detection system |
| CN111462076B (en) * | 2020-03-31 | 2023-05-16 | 湖南国科智瞳科技有限公司 | Full-slice digital pathological image fuzzy region detection method and system |
| CN111340807B (en) * | 2020-05-15 | 2020-09-11 | 中南大学 | Nidus positioning core data extraction method, system, electronic equipment and storage medium |
| CN112150446A (en) * | 2020-09-27 | 2020-12-29 | 郑州中普医疗器械有限公司 | Microscope examination and film reading scanning method, computer equipment and system |
-
2021
- 2021-01-20 CN CN202111264905.6A patent/CN113962975B/en active Active
- 2021-01-20 CN CN202110078230.XA patent/CN112750121B/en active Active
- 2021-01-20 CN CN202111267162.8A patent/CN113962976B/en active Active
Patent Citations (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102947858A (en) * | 2010-04-08 | 2013-02-27 | 帕斯科数码有限责任公司（商业用为奥姆尼克斯有限责任公司） | Image quality assessment including comparison of overlapped margins |
| CN103175834A (en) * | 2013-01-28 | 2013-06-26 | 宁波江丰生物信息技术有限公司 | Method and system for evaluating quality of digital pathological section |
| CN106600538A (en) * | 2016-12-15 | 2017-04-26 | 武汉工程大学 | Human face super-resolution algorithm based on regional depth convolution neural network |
| WO2018209057A1 (en) * | 2017-05-11 | 2018-11-15 | The Research Foundation For The State University Of New York | System and method associated with predicting segmentation quality of objects in analysis of copious image data |
| CN110738160A (en) * | 2019-10-12 | 2020-01-31 | 成都考拉悠然科技有限公司 | human face quality evaluation method combining with human face detection |
| CN110853006A (en) * | 2019-11-05 | 2020-02-28 | 华南理工大学 | Method for evaluating quality of digital pathological image acquired by scanner |
| CN110738658A (en) * | 2019-12-21 | 2020-01-31 | 杭州迪英加科技有限公司 | Image quality evaluation method |
| CN111179242A (en) * | 2019-12-25 | 2020-05-19 | Tcl华星光电技术有限公司 | Image processing method and device |
| CN111337495A (en) * | 2020-04-09 | 2020-06-26 | 江苏迪赛特医疗科技有限公司 | Pathological section image whole-course quality control device and pathological section analysis device |
| CN111986157A (en) * | 2020-07-21 | 2020-11-24 | 万达信息股份有限公司 | Digital pathological image quality evaluation system |
| CN111899246A (en) * | 2020-07-30 | 2020-11-06 | 平安科技（深圳）有限公司 | Slide digital information quality detection method, device, equipment and medium |
| CN111768405A (en) * | 2020-08-05 | 2020-10-13 | 赛维森(广州)医疗科技服务有限公司 | Method, device, equipment and storage medium for processing annotated image |
| CN112102247A (en) * | 2020-08-18 | 2020-12-18 | 广州金域医学检验中心有限公司 | Pathological section quality evaluation method based on machine learning and related equipment |
Non-Patent Citations (1)
| Title |
|---|
| 基于眼优势的非对称失真立体图像质量评价;唐祎玲;《自动化学报》;20190807;第45卷(第11期);第2092-2106页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113962976B (en) | 2022-09-16 |
| CN112750121B (en) | 2021-11-26 |
| CN113962976A (en) | 2022-01-21 |
| CN113962975A (en) | 2022-01-21 |
| CN112750121A (en) | 2021-05-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113962975B (en) | System for carrying out quality evaluation on pathological slide digital image based on gradient information | |
| CN110376198B (en) | Cervical liquid-based cell slice quality detection system | |
| US9489562B2 (en) | Image processing method and apparatus | |
| JP5920994B2 (en) | Method and system for identifying well wall boundaries of microplates | |
| US5796861A (en) | Mosaic construction, processing, and review of very large electronic micrograph composites | |
| CN111753692B (en) | Target object extraction method, product detection method, device, computer and medium | |
| CN110378313B (en) | Cell cluster identification method and device and electronic equipment | |
| CN112380900A (en) | Deep learning-based cervical fluid-based cell digital image classification method and system | |
| CN111524137A (en) | Cell identification counting method and device based on image identification and computer equipment | |
| CN114187277B (en) | Detection method for thyroid cytology multiple cell types based on deep learning | |
| CN111462076A (en) | Method and system for detecting fuzzy area of full-slice digital pathological image | |
| CN112464802B (en) | Automatic identification method and device for slide sample information and computer equipment | |
| CN115170518A (en) | Cell detection method and system based on deep learning and machine vision | |
| CN111899246A (en) | Slide digital information quality detection method, device, equipment and medium | |
| US20220318979A1 (en) | Image Quality Analysis for Artifact Detection in Pathology Slide Images | |
| CN109147932B (en) | Cancer cell HER2 gene amplification analysis method and system | |
| JP4864709B2 (en) | A system for determining the staining quality of slides using a scatter plot distribution | |
| CN118115496B (en) | Wafer defect detection method and device | |
| CN110837809A (en) | Blood automatic analysis method, blood automatic analysis system, blood cell analyzer, and storage medium | |
| CN112330613A (en) | Method and system for evaluating quality of cytopathology digital image | |
| CN114399764A (en) | Pathological section scanning method and system | |
| CN110246139B (en) | Method for rapidly extracting plankton in-situ image ROI (region of interest) based on double thresholds | |
| CN109978849B (en) | Baseline cell determination method, device and readable medium based on digital pathology image | |
| JP4897488B2 (en) | A system for classifying slides using a scatter plot distribution | |
| CN117218672A (en) | Deep learning-based medical records text recognition method and system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |