CN116030009A - Retina laser photocoagulation operation quality evaluation system - Google Patents
Retina laser photocoagulation operation quality evaluation system Download PDFInfo
- Publication number
- CN116030009A CN116030009A CN202310006480.1A CN202310006480A CN116030009A CN 116030009 A CN116030009 A CN 116030009A CN 202310006480 A CN202310006480 A CN 202310006480A CN 116030009 A CN116030009 A CN 116030009A
- Authority
- CN
- China
- Prior art keywords
- light spot
- spot
- retinal
- laser photocoagulation
- rectangular block
- 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.)
- Pending
Links
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Eye Examination Apparatus (AREA)
Abstract
The invention discloses a retina laser photocoagulation operation quality evaluation system, electronic equipment and a computer readable storage medium, and belongs to the technical field of laser spot detection. According to the invention, an evaluation system is established from three dimensions of laser spot level, laser spot distance and whether a blood vessel is injured or not, and an artificial intelligent algorithm is utilized for automatic evaluation, so that the subjectivity and instability existing in large workload and manual evaluation are solved; the quality of the retina laser photocoagulation operation can be accurately and rapidly evaluated, and the influence on the dependence and subjective factors of manpower is reduced; solves the problems of large workload and subjectivity and instability in the manual evaluation of the quality of the retina laser photocoagulation operation in the prior art.
Description
Technical Field
The application relates to the technical field of laser spot detection, in particular to a retina laser photocoagulation operation quality evaluation system.
Background
The statements in this section merely provide background information related to the present application and may not necessarily constitute prior art.
The incidence rate and blindness rate of common retina blindness caused by diabetic retinopathy, retinal vein occlusion, retinopathy of prematurity and the like are high, and heavy medical and economic burden is brought to society. Retinal laser photocoagulation is one of the important means for treating the common blinding fundus diseases. High quality retinal laser photocoagulation can effectively slow down disease progression, save patient's vision, while irregular retinal laser photocoagulation cannot slow down disease progression and present serious complications. Therefore, there is a need to evaluate the quality of retinal laser photocoagulation.
Parameters of laser photocoagulation include laser power, spot size, duration, laser spot spacing, etc. Research shows that various diseases can be treated by a certain laser energy, and the laser energy acts on the retina and is expressed as a laser spot formed after laser irradiates on the retina. Therefore, researchers divide laser spots into four levels according to the colors and the forms of the laser spots formed by different energies, and clinically judge whether the retinal laser surgery reaches the standard or not through the levels of the laser spots, for example, the diabetic retinopathy, retinal vein occlusion and retinopathy of premature infants all need to reach 3 levels of laser spots to exert the treatment effect of laser photocoagulation. But the problems existing at present are:
(1) The evaluation workload of the retinal laser photocoagulation operation is large, subjectivity and instability exist in manual evaluation, and doctors with low annual resources are inexperienced, so that the retinal laser photocoagulation operation is difficult to evaluate effectively and efficiently;
(2) There is no intelligent system that can automatically evaluate the quality of retinal laser photocoagulation.
Disclosure of Invention
In order to solve the defects of the prior art, the application provides a retina laser photocoagulation operation quality evaluation system, electronic equipment and a computer readable storage medium, wherein the evaluation system is built from three dimensions of laser spot level, laser spot distance and whether blood vessels are injured, and the evaluation is automatically performed by using an artificial intelligence algorithm, so that the problems of large workload and subjectivity and instability in manual evaluation are solved.
In a first aspect, the present application provides a retinal laser photocoagulation procedure quality assessment system;
a retinal laser photocoagulation quality assessment system comprising:
an image acquisition module configured to: obtaining fundus color illumination of a patient after a retina laser photocoagulation operation;
a spot target detection module configured to: inputting fundus color photographs of a patient after a retina laser photocoagulation operation into a facula target detection model to obtain facula rectangular block images and facula levels;
a retinal vascular injury prediction module configured to: inputting the rectangular block image of the light spot into a relation model of the light spot and the retinal blood vessel, and predicting whether the light spot hurts the retinal blood vessel;
the light spot distribution detection module is configured to: according to the spot rectangular block image, judging whether the spot distribution is uniform or not according to a 3sigma anomaly detection principle by using an analysis of variance method based on the spot distance.
Further, inputting the fundus color photograph of the patient after the retinal laser photocoagulation operation into a facula target detection model, and acquiring the facula rectangular block image and the facula level specifically comprises the following steps:
receiving fundus color illumination of a patient after the retinal laser photocoagulation operation, and preprocessing the fundus color illumination of the patient after the retinal laser photocoagulation operation;
according to the preprocessed fundus color photograph, extracting light spot characteristics in the image and fusing the light spot characteristics;
and determining the position, the size and the level of the light spot according to the fused light spot characteristics, acquiring a rectangular block image of the light spot and outputting the category confidence.
Further, the preprocessing of fundus color illumination after retinal laser photocoagulation operation includes:
sequentially executing cutting operation and zooming operation on fundus color photographs after retina laser photocoagulation operation;
the size of the fundus color photograph after clipping and scaling is adjusted to 608x608x3.
Further, the light spot target detection model is a YOLO V5 model, and comprises an input end, a backup network, a rock network and a Prediction network which are sequentially connected;
the input end is used for receiving fundus color photographs of a patient after the retinal laser photocoagulation operation, preprocessing the fundus color photographs of the patient after the retinal laser photocoagulation operation, the back one network is used for extracting light spot characteristics of the preprocessed fundus color photographs, the Neck network is used for fusing the light spot characteristics, and the Prediction network is used for determining the position, the size and the level of the light spots according to the light spot characteristics and outputting category confidence and light spot rectangular block images.
Further, inputting the rectangular block image of the light spot into a relation model of the light spot and the retinal blood vessel, and predicting whether the light spot hurts the retinal blood vessel specifically comprises:
preprocessing the rectangular block image of the light spot;
extracting a high-dimensional feature map of the preprocessed facula rectangular block image, and adjusting the number of the high-dimensional feature maps;
and (3) reducing the dimension of the high-dimension feature map, finishing classification by the output feature map, and outputting a classification result, wherein the classification result is whether retinal blood vessels are contained or not.
Further, the light spot and retina blood vessel relation model is an EfficientNet network model.
Further, the EfficientNet network model comprises a Conv3x3 layer, an MBConv module, a Conv 1x1 layer and an FC full-connection module which are sequentially connected, wherein the Conv3x3 layer is used for preprocessing a facula rectangular block image, and the MBConv module is used for acquiring an image high-dimensional feature map according to the preprocessed facula rectangular block image; the Conv 1x1 layer is used for adjusting the number of the high-dimensional feature images of the image; the FC full-connection module is used for reducing the dimension of the high-dimension feature map, classifying the high-dimension feature map through the output feature map, and outputting a classification result.
Further, the implementation process of the light spot distribution detection module is as follows:
extracting the centers of light spot rectangular frames in the light spot rectangular block image, and calculating Euclidean distances among the light spot centers;
and (3) counting the mean value and variance of the light spot distance, comparing the mean value and variance with the parameters of standard light spot distribution, and judging whether the light spot distribution is uniform or not according to a 3sigma anomaly detection principle.
In a second aspect, the present application provides an electronic device;
an electronic device comprising a memory and a processor and computer instructions stored on the memory and running on the processor, which when executed by the processor, perform the steps of:
obtaining fundus color illumination of a patient after a retina laser photocoagulation operation;
inputting fundus color photographs of a patient after a retina laser photocoagulation operation into a facula target detection model to obtain facula rectangular block images and categories of facula;
inputting the rectangular block image of the light spot into a relation model of the light spot and the retinal blood vessel, and predicting whether the light spot hurts the retinal blood vessel;
according to the spot rectangular block image, judging whether the spot distribution is uniform or not according to a 3sigma anomaly detection principle by using an analysis of variance method based on the spot distance.
In a third aspect, the present application provides a computer-readable storage medium;
a computer readable storage medium storing computer instructions that, when executed by a processor, perform the steps of:
obtaining fundus color illumination of a patient after a retina laser photocoagulation operation;
inputting fundus color photographs of a patient after a retina laser photocoagulation operation into a facula target detection model to obtain facula rectangular block images and facula levels;
inputting the rectangular block image of the light spot into a relation model of the light spot and the retinal blood vessel, and predicting whether the light spot hurts the retinal blood vessel;
according to the spot rectangular block image, judging whether the spot distribution is uniform or not according to a 3sigma anomaly detection principle by using an analysis of variance method based on the spot distance.
Compared with the prior art, the beneficial effects of this application are:
(1) The method focuses on the difficult problem that the quality of the retinal laser photocoagulation operation is difficult to evaluate, an evaluation system is built from three dimensions of laser spot level, laser spot distance and whether blood vessels are injured, and the evaluation system is automatically evaluated by utilizing an artificial intelligent algorithm, so that the problems of large workload of manually evaluating the quality of the retinal laser photocoagulation operation and subjectivity and instability existing in manual evaluation are solved, and the accuracy and the efficiency of evaluating the quality of the retinal laser photocoagulation operation are improved.
(2) According to the technical scheme, the quality of the retinal laser photocoagulation operation can be efficiently and accurately evaluated, adverse reactions of the retinal laser photocoagulation operation are reduced, and the curative effect is improved. Simultaneously, the quality of the retinal laser photocoagulation operation of the low-annual doctor is improved rapidly, and the occurrence of adverse medical events is reduced.
Drawings
The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiments of the application and together with the description serve to explain the application and do not constitute an undue limitation to the application.
FIG. 1 is a schematic diagram of a system framework provided in an embodiment of the present application;
fig. 2 is a schematic diagram of a YOLO v5 network structure according to an embodiment of the present application;
fig. 3 is an architecture of an afflicientnet network provided in an embodiment of the present application;
fig. 4 is a schematic diagram of 3sigma anomaly detection provided in an embodiment of the present application.
Detailed Description
It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the present application. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of example embodiments in accordance with the present application. As used herein, unless the context clearly indicates otherwise, the singular forms also are intended to include the plural forms, and furthermore, it is to be understood that the terms "comprises" and "comprising" and any variations thereof are intended to cover non-exclusive inclusions, such as, for example, processes, methods, systems, products or devices that comprise a series of steps or units, are not necessarily limited to those steps or units that are expressly listed, but may include other steps or units that are not expressly listed or inherent to such processes, methods, products or devices.
Embodiments of the invention and features of the embodiments may be combined with each other without conflict.
Example 1
In the prior art, the manual evaluation of retinal laser photocoagulation operation has great workload and subjectivity and instability; therefore, the application provides a retina laser photocoagulation operation quality evaluation system, which is established from three dimensions of laser spot level, laser spot distance and whether blood vessels are injured or not by utilizing an artificial intelligent algorithm.
Next, a retinal laser photocoagulation operation quality evaluation system disclosed in this embodiment will be described in detail with reference to fig. 1 to 4.
A retinal laser photocoagulation operation quality evaluation system comprises an image acquisition module, a light spot target detection module, a retinal vascular injury prediction module and a light spot distribution detection module; the image acquisition module is used for acquiring fundus color photographs of the patient after the retinal laser photocoagulation operation, the spot target detection module is used for inputting the fundus color photographs of the patient after the retinal laser photocoagulation operation into the spot target detection model, and acquiring a spot rectangular block image and the level of the spot (namely grading the spot); the retinal vascular injury prediction module is used for inputting the rectangular block image of the light spot into a relation model of the light spot and the retinal vascular to predict whether the light spot hurts the retinal vascular; the light spot distribution detection module is used for judging whether light spot distribution is uniform according to a light spot rectangular block image and a variance analysis method based on light spot distance and a 3sigma anomaly detection principle.
The image acquisition module is realized as follows:
and obtaining the instant fundus color illumination of the patient after the retinal laser photocoagulation operation.
The light spot target detection module is realized as follows:
inputting fundus color photographs of a patient after a retina laser photocoagulation operation into a facula target detection model to obtain facula rectangular block images and categories of facula; the method comprises the following specific steps:
(1) Receiving fundus color illumination of a patient after the retinal laser photocoagulation operation, and preprocessing the fundus color illumination of the patient after the retinal laser photocoagulation operation; specifically, a cropping operation and a zooming operation are sequentially carried out on fundus color photographs after retina laser photocoagulation operation; the size of the fundus color photograph after clipping and scaling is adjusted to 608x608x3.
(2) According to the preprocessed fundus color photograph, extracting light spot characteristics in the image and fusing the light spot characteristics;
(3) According to the fused light spot characteristics, determining the position, the size and the category of the light spot, acquiring a rectangular block image of the light spot and outputting a category confidence level; the classification of the light spots is the classification result of the light spot target detection model on the light spots according to the classification standard of the light spots.
The target detection model of the light spot adopts a YOLO v5 model, and the network structure of the target detection model is shown in fig. 2, and specifically includes: input, backhaul network, neg network and Prediction me take four parts. The input end receives fundus color illumination pretreatment images after retina laser photocoagulation operation, and the size of the fundus color illumination pretreatment images is 608x608x3 after a series of conversion such as cutting, zooming and the like; the Backone network is used as a backbone network and comprises a Focus and CSP structure, and the aim is to extract the facula characteristics of the preprocessed fundus color photograph; the Neck network realizes further fusion of the light spot characteristics by means of more complex module combination; the Prediction network is a classifier and a regressor of the YOLO v5 model, and is used for determining the position, the size and the level of the light spot, and outputting the category confidence level thereof. Once the YOLO v5 model training is completed, the light spots and light spot levels therein can be detected from the input post-operative retinal image.
The implementation of the retinal vascular injury prediction module is as follows:
preprocessing a rectangular block image of the light spot, inputting a relation model of the light spot and retinal blood vessels, and predicting whether the light spot hurts the retinal blood vessels; the method specifically comprises the following steps:
(1) Preprocessing the rectangular block image of the light spot;
(2) Extracting a high-dimensional feature map of the preprocessed facula rectangular block image, and adjusting the number of the high-dimensional feature maps;
(3) And (3) reducing the dimension of the high-dimension feature map, finishing classification by the output feature map, and outputting a classification result, wherein the classification result is whether retinal blood vessels are contained or not.
Illustratively, the light spot and retina vascular relation model adopts an advanced convolutional neural network EfficientNet, and adjusts the convolutional input and FC output parts of the network result according to the task characteristics, as shown in FIG. 3. The EfficientNet network model balances multiple dimensions of resolution, depth and width by means of a composite model scaling method, and achieves optimization of the network in efficiency and accuracy.
The workflow of the EfficientNet network model is as follows: 1. preprocessing a spot rectangular block image detected by a YOLO v5 model to obtain an image with the size of 32x32x3, wherein the image is used as an input of a network; 2. converting an input image into an input dimension required by an MBConv module through a first Conv3x3 layer; 3. the image is subjected to a series of MBConv modules to obtain an image high-dimensional feature map (feature maps) to realize image high-dimensional feature extraction, wherein parameters of each MBConv module are subjected to fine adjustment to adapt to a current task, and a combined scale optimization method can enable a network to obtain a better receptive field; 4. adjusting the number of the high-dimensional feature graphs by using a Conv 1x1 layer; 5. the method comprises the steps of (1) reducing dimension of a feature by using a series of FC full-connection modules; 5. and finally, finishing classification by the output characteristic diagram, accurately predicting whether the light spot hurts the retinal blood vessel, if the classification result is that the light spot contains the retinal blood vessel, then the light spot hurts the retinal blood vessel, and if the classification result is that the light spot does not contain the retinal blood vessel, then the light spot does not hurt the retinal blood vessel.
The light spot distribution detection module is realized as follows:
(1) Extracting the centers of light spot rectangular frames in the light spot rectangular block image, and calculating Euclidean distances among the light spot centers; assuming that the light spot distribution mean values z are mutually independent and accord with normal distribution:
(2) And (3) counting the mean value and variance of the light spot distance, comparing the mean value and variance with the parameters of standard light spot distribution, and judging whether the light spot distribution is uniform or not according to a 3sigma anomaly detection principle.
The anomaly detection based on the 3sigma statistical method does not need an empirical threshold, adopts a 3sigma principle, assumes that the detected data only contains random errors, calculates the original data to obtain standard deviation, and then determines a section according to a certain probability, and considers that the error exceeds the section to be an anomaly value. FIG. 4 is a schematic diagram of 3sigma anomaly detection with values distributed over intervals (μ-3σ，μ+3σ) Up to 99.7%, and an outlier, i.e. a non-uniform spot distribution, is present before this interval.
And finally, evaluating the quality of the retinal laser photocoagulation operation according to the light spot level, whether the retinal blood vessels are injured by the light spots and whether the light spot distribution is uniform.
Besides the Yolo series, the spot target detection model can also be replaced by other deep learning target detection networks; the spot and retinal vascular relationship model may be replaced with other deep learning classification networks, such as VGG, resNet, googleNet and ViT, in addition to the Efficient Net.
In some embodiments, the system further comprises a surgical quality assessment module for assessing the quality of the retinal laser photocoagulation procedure based on the spot level, whether the spots are damaging to retinal blood vessels, and whether the spot distribution is uniform.
Example two
The second embodiment of the invention provides an electronic device, which comprises a memory, a processor and computer instructions stored on the memory and running on the processor, wherein when the computer instructions are run by the processor, the following steps are completed: obtaining fundus color illumination of a patient after a retina laser photocoagulation operation;
inputting fundus color photographs of a patient after a retina laser photocoagulation operation into a facula target detection model to obtain facula rectangular block images and categories of facula;
inputting the rectangular block image of the light spot into a relation model of the light spot and the retinal blood vessel, and predicting whether the light spot hurts the retinal blood vessel;
according to the spot rectangular block image, judging whether the spot distribution is uniform or not according to a 3sigma anomaly detection principle by using an analysis of variance method based on the spot distance.
Example III
The third embodiment of the invention provides a computer readable storage medium for storing computer instructions, which when executed by a processor, complete the steps of;
obtaining fundus color illumination of a patient after a retina laser photocoagulation operation;
inputting fundus color photographs of a patient after a retina laser photocoagulation operation into a facula target detection model to obtain facula rectangular block images and categories of facula;
inputting the rectangular block image of the light spot into a relation model of the light spot and the retinal blood vessel, and predicting whether the light spot hurts the retinal blood vessel;
according to the spot rectangular block image, judging whether the spot distribution is uniform or not according to a 3sigma anomaly detection principle by using an analysis of variance method based on the spot distance.
The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The foregoing embodiments are directed to various embodiments, and details of one embodiment may be found in the related description of another embodiment.
The foregoing description is only of the preferred embodiments of the present application and is not intended to limit the same, but rather, various modifications and variations may be made by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.
Claims (10)
1. A retinal laser photocoagulation quality assessment system, comprising:
an image acquisition module configured to: obtaining fundus color illumination of a patient after a retina laser photocoagulation operation;
a spot target detection module configured to: inputting fundus color photographs of a patient after a retina laser photocoagulation operation into a facula target detection model to obtain facula rectangular block images and facula levels;
a retinal vascular injury prediction module configured to: inputting the rectangular block image of the light spot into a relation model of the light spot and the retinal blood vessel, and predicting whether the light spot hurts the retinal blood vessel;
the light spot distribution detection module is configured to: according to the spot rectangular block image, judging whether the spot distribution is uniform or not according to a 3sigma anomaly detection principle by using an analysis of variance method based on the spot distance.
2. The retinal laser photocoagulation operation quality assessment system according to claim 1, wherein inputting the fundus color photograph of the patient after the retinal laser photocoagulation operation into the spot target detection model, obtaining the spot rectangular block image and the categories of the spots specifically comprises:
receiving fundus color illumination of a patient after the retinal laser photocoagulation operation, and preprocessing the fundus color illumination of the patient after the retinal laser photocoagulation operation;
according to the preprocessed fundus color photograph, extracting light spot characteristics in the image and fusing the light spot characteristics;
and determining the position, the size and the category of the light spot according to the fused light spot characteristics, acquiring a rectangular block image of the light spot and outputting the category confidence.
3. The retinal laser photocoagulation quality assessment system according to claim 2, wherein the preprocessing of fundus illumination after retinal laser photocoagulation comprises:
sequentially executing cutting operation and zooming operation on fundus color photographs after retina laser photocoagulation operation;
the size of the fundus color photograph after clipping and scaling is adjusted to 608x608x3.
4. The retinal laser photocoagulation surgical quality assessment system according to claim 1, wherein the spot target detection model is a YOLO V5 model, and the spot target detection model includes an input terminal, a backup network, a neg network, and a Prediction network connected in sequence;
the input end is used for receiving fundus color photographs of a patient after the retinal laser photocoagulation operation, preprocessing the fundus color photographs of the patient after the retinal laser photocoagulation operation, the back one network is used for extracting light spot characteristics of the preprocessed fundus color photographs, the Neck network is used for fusing the light spot characteristics, and the Prediction network is used for determining positions, sizes and types of light spots according to the light spot characteristics and outputting category confidence and light spot rectangular block images.
5. The retinal laser photocoagulation operation quality assessment system according to claim 1, wherein the inputting the rectangular block image of the light spot into the model of the relationship between the light spot and the retinal blood vessel, and predicting whether the light spot damages the retinal blood vessel specifically comprises:
preprocessing the rectangular block image of the light spot;
extracting a high-dimensional feature map of the preprocessed facula rectangular block image, and adjusting the number of the high-dimensional feature maps;
and (3) reducing the dimension of the high-dimension feature map, finishing classification by the output feature map, and outputting a classification result, wherein the classification result is whether retinal blood vessels are contained or not.
6. The retinal laser photocoagulation quality assessment system according to claim 1, wherein the spot to retinal vessel relationship model is an afflicientnet network model.
7. The retinal laser photocoagulation operation quality assessment system according to claim 6, wherein the afflicientnet network model comprises a Conv3x3 layer, an MBConv module, a Conv 1x1 layer and an FC full connection module which are sequentially connected, wherein the Conv3x3 layer is used for preprocessing a facula rectangular block image, and the MBConv module is used for acquiring an image high-dimensional feature map according to the preprocessed facula rectangular block image; the Conv 1x1 layer is used for adjusting the number of the high-dimensional feature images of the image; the FC full-connection module is used for reducing the dimension of the high-dimension feature map, classifying the high-dimension feature map through the output feature map, and outputting a classification result.
8. The retinal laser photocoagulation operation quality assessment system according to claim 1, wherein the spot morphology distribution detection module is implemented as follows:
extracting the centers of light spot rectangular frames in the light spot rectangular block image, and calculating Euclidean distances among the light spot centers;
and (3) counting the mean value and variance of the light spot distance, comparing the mean value and variance with the parameters of standard light spot distribution, and judging whether the light spot distribution is uniform or not according to a 3sigma anomaly detection principle.
9. An electronic device comprising a memory and a processor, and computer instructions stored on the memory and running on the processor, which when executed by the processor, perform the steps of:
acquiring fundus color photographs after retinal laser photocoagulation surgery and preprocessing;
inputting the preprocessed fundus color illumination into a facula target detection model to obtain facula rectangular block images and facula categories;
preprocessing a rectangular block image of the light spot, inputting a relation model of the light spot and retinal blood vessels, and predicting whether the light spot hurts the retinal blood vessels;
according to the spot rectangular block image, judging whether the spot distribution is uniform or not according to a 3sigma anomaly detection principle by using an analysis of variance method based on the spot distance.
10. A computer readable storage medium storing computer instructions which, when executed by a processor, perform the steps of:
acquiring fundus color photographs after retinal laser photocoagulation surgery and preprocessing;
inputting the preprocessed fundus color illumination into a facula target detection model to obtain facula rectangular block images and facula categories;
preprocessing a rectangular block image of the light spot, inputting a relation model of the light spot and retinal blood vessels, and predicting whether the light spot hurts the retinal blood vessels;
according to the spot rectangular block image, judging whether the spot distribution is uniform or not according to a 3sigma anomaly detection principle by using an analysis of variance method based on the spot distance.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310006480.1A CN116030009A (en) | 2023-01-04 | 2023-01-04 | Retina laser photocoagulation operation quality evaluation system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310006480.1A CN116030009A (en) | 2023-01-04 | 2023-01-04 | Retina laser photocoagulation operation quality evaluation system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116030009A true CN116030009A (en) | 2023-04-28 |
Family
ID=86075488
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310006480.1A Pending CN116030009A (en) | 2023-01-04 | 2023-01-04 | Retina laser photocoagulation operation quality evaluation system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116030009A (en) |
-
2023
- 2023-01-04 CN CN202310006480.1A patent/CN116030009A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108876775B (en) | Method for rapidly detecting diabetic retinopathy | |
| CN109087302A (en) | A kind of eye fundus image blood vessel segmentation method and apparatus | |
| CN110400289A (en) | Eye fundus image recognition methods, device, equipment and storage medium | |
| Arora et al. | Deep neural network for diabetic retinopathy detection | |
| Thota et al. | Improving the accuracy of diabetic retinopathy severity classification with transfer learning | |
| CN105513077A (en) | System for screening diabetic retinopathy | |
| CN110428421A (en) | Macula lutea image region segmentation method and apparatus | |
| CN109447962A (en) | A kind of eye fundus image hard exudate lesion detection method based on convolutional neural networks | |
| CN111986183B (en) | Chromosome scattered image automatic segmentation and identification system and device | |
| CN110428410A (en) | Eyeground medical image processing method, device, equipment and storage medium | |
| Fan et al. | Principal component analysis based cataract grading and classification | |
| CN112016626A (en) | Diabetic retinopathy classification system based on uncertainty | |
| CN110555875A (en) | Pupil radius detection method and device, computer equipment and storage medium | |
| CN109961848A (en) | Macula lutea image classification method and equipment | |
| CN114240978A (en) | Cell edge segmentation method and device based on adaptive morphology | |
| CN114445666A (en) | Deep learning-based method and system for classifying left eye, right eye and visual field positions of fundus images | |
| CN116030009A (en) | Retina laser photocoagulation operation quality evaluation system | |
| CN103955941B (en) | A kind of specularmicroscope and the method for analysis | |
| CN113053517A (en) | Facial paralysis grade evaluation method based on dynamic region quantitative indexes | |
| CN116269744A (en) | Laser energy control system for crushing stones | |
| CN115701868A (en) | Domain self-adaptive enhancement method suitable for various visual tasks | |
| CN112991289B (en) | Processing method and device for standard section of image | |
| CN111374632B (en) | Retinopathy detection method, device and computer readable storage medium | |
| CN114972148A (en) | Fundus image quality evaluation method, system and device | |
| Kumara et al. | Active contour-based segmentation and removal of optic disk from retinal images |
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 |