WO2021139446A1 - 一种抗血管内皮生长因子vegf疗效预测装置及方法 - Google Patents
一种抗血管内皮生长因子vegf疗效预测装置及方法 Download PDFInfo
- Publication number
- WO2021139446A1 WO2021139446A1 PCT/CN2020/132473 CN2020132473W WO2021139446A1 WO 2021139446 A1 WO2021139446 A1 WO 2021139446A1 CN 2020132473 W CN2020132473 W CN 2020132473W WO 2021139446 A1 WO2021139446 A1 WO 2021139446A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- oct images
- image
- vegf
- curative effect
- features corresponding
- Prior art date
Links
- 230000000694 effects Effects 0.000 title claims abstract description 127
- 238000000034 method Methods 0.000 title claims abstract description 43
- 108010041308 Endothelial Growth Factors Proteins 0.000 title claims abstract description 30
- 230000002137 anti-vascular effect Effects 0.000 title claims abstract description 30
- 101100372758 Danio rerio vegfaa gene Proteins 0.000 title 1
- 101150030763 Vegfa gene Proteins 0.000 title 1
- 238000012014 optical coherence tomography Methods 0.000 claims abstract description 359
- 238000000605 extraction Methods 0.000 claims abstract description 82
- 102000005789 Vascular Endothelial Growth Factors Human genes 0.000 claims abstract description 36
- 108010019530 Vascular Endothelial Growth Factors Proteins 0.000 claims abstract description 36
- 230000004927 fusion Effects 0.000 claims abstract description 29
- 230000004438 eyesight Effects 0.000 claims description 29
- 238000012545 processing Methods 0.000 claims description 24
- 238000013528 artificial neural network Methods 0.000 claims description 23
- 238000012937 correction Methods 0.000 claims description 18
- 230000006403 short-term memory Effects 0.000 claims description 16
- 238000003702 image correction Methods 0.000 claims description 14
- 230000007787 long-term memory Effects 0.000 claims description 14
- 230000006866 deterioration Effects 0.000 claims description 13
- 230000015654 memory Effects 0.000 claims description 13
- 230000006872 improvement Effects 0.000 claims description 11
- 206010025421 Macule Diseases 0.000 claims description 9
- 238000004891 communication Methods 0.000 claims description 8
- 238000004590 computer program Methods 0.000 claims description 7
- 230000009286 beneficial effect Effects 0.000 abstract description 3
- 238000002347 injection Methods 0.000 description 12
- 239000007924 injection Substances 0.000 description 12
- 230000011218 segmentation Effects 0.000 description 11
- 239000000243 solution Substances 0.000 description 9
- 238000010586 diagram Methods 0.000 description 8
- 230000001965 increasing effect Effects 0.000 description 8
- 230000008569 process Effects 0.000 description 8
- 230000004304 visual acuity Effects 0.000 description 8
- 230000006870 function Effects 0.000 description 6
- 108010073929 Vascular Endothelial Growth Factor A Proteins 0.000 description 5
- 230000008034 disappearance Effects 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 5
- 230000000007 visual effect Effects 0.000 description 5
- 208000000208 Wet Macular Degeneration Diseases 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 4
- 210000004027 cell Anatomy 0.000 description 4
- 201000010099 disease Diseases 0.000 description 4
- 208000037265 diseases, disorders, signs and symptoms Diseases 0.000 description 4
- 210000001525 retina Anatomy 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 206010064930 age-related macular degeneration Diseases 0.000 description 3
- 208000002780 macular degeneration Diseases 0.000 description 3
- 238000007781 pre-processing Methods 0.000 description 3
- 230000002207 retinal effect Effects 0.000 description 3
- 210000001519 tissue Anatomy 0.000 description 3
- 208000037170 Delayed Emergence from Anesthesia Diseases 0.000 description 2
- 102000009524 Vascular Endothelial Growth Factor A Human genes 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000013527 convolutional neural network Methods 0.000 description 2
- 125000004122 cyclic group Chemical group 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 208000030533 eye disease Diseases 0.000 description 2
- 238000002372 labelling Methods 0.000 description 2
- 210000004379 membrane Anatomy 0.000 description 2
- 239000012528 membrane Substances 0.000 description 2
- 230000002093 peripheral effect Effects 0.000 description 2
- 238000002560 therapeutic procedure Methods 0.000 description 2
- 230000009466 transformation Effects 0.000 description 2
- 238000012285 ultrasound imaging Methods 0.000 description 2
- 210000003556 vascular endothelial cell Anatomy 0.000 description 2
- 102000010834 Extracellular Matrix Proteins Human genes 0.000 description 1
- 108010037362 Extracellular Matrix Proteins Proteins 0.000 description 1
- 108090000386 Fibroblast Growth Factor 1 Proteins 0.000 description 1
- 102100031706 Fibroblast growth factor 1 Human genes 0.000 description 1
- 230000032683 aging Effects 0.000 description 1
- 230000033115 angiogenesis Effects 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 210000000270 basal cell Anatomy 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 210000001775 bruch membrane Anatomy 0.000 description 1
- 210000000805 cytoplasm Anatomy 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000007850 degeneration Effects 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 210000002744 extracellular matrix Anatomy 0.000 description 1
- 230000014509 gene expression Effects 0.000 description 1
- 230000003902 lesion Effects 0.000 description 1
- 230000005012 migration Effects 0.000 description 1
- 238000013508 migration Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000035755 proliferation Effects 0.000 description 1
- 230000000717 retained effect Effects 0.000 description 1
- 210000000844 retinal pigment epithelial cell Anatomy 0.000 description 1
- 230000001225 therapeutic effect Effects 0.000 description 1
- 238000003325 tomography Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 230000008728 vascular permeability Effects 0.000 description 1
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B3/00—Apparatus for testing the eyes; Instruments for examining the eyes
- A61B3/10—Objective types, i.e. instruments for examining the eyes independent of the patients' perceptions or reactions
- A61B3/102—Objective types, i.e. instruments for examining the eyes independent of the patients' perceptions or reactions for optical coherence tomography [OCT]
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B3/00—Apparatus for testing the eyes; Instruments for examining the eyes
- A61B3/0016—Operational features thereof
- A61B3/0025—Operational features thereof characterised by electronic signal processing, e.g. eye models
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B3/00—Apparatus for testing the eyes; Instruments for examining the eyes
- A61B3/10—Objective types, i.e. instruments for examining the eyes independent of the patients' perceptions or reactions
- A61B3/12—Objective types, i.e. instruments for examining the eyes independent of the patients' perceptions or reactions for looking at the eye fundus, e.g. ophthalmoscopes
- A61B3/1225—Objective types, i.e. instruments for examining the eyes independent of the patients' perceptions or reactions for looking at the eye fundus, e.g. ophthalmoscopes using coherent radiation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/25—Fusion techniques
- G06F18/253—Fusion techniques of extracted features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- 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
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H30/00—ICT specially adapted for the handling or processing of medical images
- G16H30/20—ICT specially adapted for the handling or processing of medical images for handling medical images, e.g. DICOM, HL7 or PACS
-
- 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
- G06T2207/10072—Tomographic images
- G06T2207/10101—Optical tomography; Optical coherence tomography [OCT]
-
- 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
- G06T2207/30101—Blood vessel; Artery; Vein; Vascular
Definitions
- This application relates to the field of medical science and technology, and in particular to an anti-vascular endothelial growth factor VEGF curative effect prediction device and method.
- ATD age-related macular degeneration
- VEGF vascular endothelial growth factor
- intraocular injection is an effective treatment for wet AMD, but anti-VEGF injection therapy is expensive and has strict indications, and the efficacy of different patients varies. Due to the lack of effective anti-VEGF curative effect predictions, doctors often adopt uniform injection methods for patients, resulting in anti-VEGF injections for some patients who are not applicable. Therefore, effective anti-VEGF curative effect prediction is an urgent need of doctors.
- Optical coherence tomography is a commonly used device for diagnosing ophthalmic diseases. It uses light reflection technology similar to ultrasound imaging to provide image reference for the detection and treatment of ophthalmic diseases.
- OCT optical coherence tomography
- the inventor realized that in the currently commonly used anti-VEGF curative effect prediction method, the lesion area (such as effusion, high reflection point, etc.) is segmented through a segmentation network, and then anti-VEGF curative effect prediction is performed after segmentation.
- the segmentation network based on deep learning requires a large number of doctors' annotations during the training process, and the accuracy of the annotations and the accuracy of the segmentation network segmentation will affect the curative effect prediction results.
- a large amount of retinal tissue change information that may improve the accuracy of anti-VEGF curative effect prediction is lost, resulting in low accuracy of anti-VEGF curative effect prediction.
- the application provides an anti-vascular endothelial growth factor VEGF curative effect prediction device and method, which is beneficial to improve the accuracy of anti-VEGF curative effect prediction.
- the first aspect of this application provides an anti-vascular endothelial growth factor VEGF curative effect prediction device, which includes: an acquisition module for acquiring multiple OCT images of the optical coherence tomography acquired for the macula; a feature extraction module for evaluating the Perform feature extraction on the multiple OCT images to obtain multiple image features corresponding to the multiple OCT images, wherein the image features corresponding to each OCT image in the multiple OCT images include Multi-scale features; a spatial information fusion module for spatial information fusion of multiple image features corresponding to the multiple OCT images to obtain the features corresponding to the multiple OCT images; a determination module, used to determine the features corresponding to the multiple OCT images The characteristics corresponding to the OCT image determine the predictive results of anti-VEGF curative effect.
- the second aspect of the present application provides a method for predicting the therapeutic effect of anti-vascular endothelial growth factor VEGF, which includes: acquiring multiple OCT images of optical coherence tomography collected for the macular area; performing feature extraction on the multiple OCT images to obtain the results
- the multiple image features corresponding to the multiple OCT images wherein the image feature corresponding to each OCT image in the multiple OCT images includes the multi-scale feature in each OCT image; and the multiple OCT images Spatial information fusion is performed on the corresponding multiple image features to obtain the features corresponding to the multiple OCT images; according to the features corresponding to the multiple OCT images, the anti-VEGF curative effect prediction result is determined.
- the third aspect of the present application provides an electronic device that includes a processor, a memory, a communication interface, and one or more programs, wherein the one or more programs are stored in the memory and are
- the configuration is executed by the processor to realize the following method: acquiring multiple OCT images of optical coherence tomography collected for the macula; performing feature extraction on the multiple OCT images to obtain multiple corresponding to the multiple OCT images Image features, wherein the image feature corresponding to each OCT image in the multiple OCT images includes the multi-scale feature in each OCT image; the multiple image features corresponding to the multiple OCT images are spatially Information fusion is used to obtain the features corresponding to the multiple OCT images; according to the features corresponding to the multiple OCT images, the anti-VEGF curative effect prediction result is determined.
- the fourth aspect of the present application provides a computer-readable storage medium, the computer-readable storage medium stores a computer program, and the computer program is executed by a processor to implement the following method: Obtain a plurality of optical coherence images collected for the macular region A tomography OCT image; feature extraction is performed on the multiple OCT images to obtain multiple image features corresponding to the multiple OCT images, wherein the image features corresponding to each OCT image in the multiple OCT images include The multi-scale features in each OCT image; spatial information fusion is performed on the multiple image features corresponding to the multiple OCT images to obtain the features corresponding to the multiple OCT images; according to the corresponding features of the multiple OCT images Features to determine the predictive results of anti-VEGF curative effect.
- the application After acquiring multiple OCT images, the application does not need to be segmented, which reduces the cost of labeling in the segmentation network, and at the same time avoids the accuracy error of anti-VEGF curative effect prediction caused by inaccurate segmentation.
- the image features corresponding to each extracted OCT image include multi-scale features, which increases the richness and comprehensiveness of the extracted features, thereby improving the accuracy of anti-VEGF curative effect prediction.
- the spatial information fusion of multiple image features effectively utilizes the spatial characteristics of multiple OCT images, enriches the spatial information of the features, and also improves the accuracy of anti-VEGF curative effect prediction.
- Fig. 1 is a schematic diagram of a network structure for predicting the curative effect of anti-vascular endothelial growth factor VEGF according to an embodiment of the application.
- Figure 2 is a schematic flow chart of a method for predicting curative effect of anti-vascular endothelial growth factor VEGF provided by an embodiment of the application.
- FIG. 3 is a schematic diagram of a feature extraction network provided by an embodiment of this application.
- Fig. 4 is a schematic flow chart of another method for predicting the efficacy of anti-vascular endothelial growth factor VEGF provided by an embodiment of the application.
- Fig. 5 is a schematic diagram of an anti-vascular endothelial growth factor VEGF curative effect prediction device provided by an embodiment of the application.
- FIG. 6 is a schematic diagram of the structure of an electronic device in a hardware operating environment involved in an embodiment of the application.
- the device and method for predicting the curative effect of anti-vascular endothelial growth factor VEGF provided in the embodiments of the present application are beneficial to improve the accuracy of anti-VEGF curative effect prediction.
- At least one refers to one or more, and “multiple” refers to two or more.
- And/or describes the association relationship of the associated objects, indicating that there can be three relationships, for example, A and/or B, which can mean: A alone exists, A and B exist at the same time, and B exists alone, where A, B can be singular or plural.
- the character “/” generally indicates that the associated objects before and after are in an “or” relationship.
- "The following at least one item (a)” or similar expressions refers to any combination of these items, including any combination of a single item (a) or a plurality of items (a).
- first, second, third, and fourth in the specification and claims of this application and the above-mentioned drawings are used to distinguish different objects, rather than describing a specific order, Timing, priority, or importance.
- first information and the second information are only for distinguishing different information, but do not indicate the difference in content, priority, sending order, or importance of the two types of information.
- the terms “including” and “having” and any variations of them are intended to cover non-exclusive inclusions.
- a process, method, system, product, or device that includes a series of steps or units is not limited to the listed steps or units, but optionally includes unlisted steps or units, or optionally also includes Other steps or units inherent to these processes, methods, products or equipment.
- the technical solution of this application can be applied to the fields of artificial intelligence, smart city, digital medical and/or blockchain technology to realize smart medical treatment.
- the data involved in this application can be stored in a database, or can be stored in a blockchain, such as distributed storage through a blockchain, which is not limited in this application.
- AMD AMD is a major blinding eye disease, which is an aging change in the structure of the macular area.
- the main manifestation is that the ability of retinal pigment epithelial cells to phagocytose and digest the outer segmental membrane of the optic cell decreases.
- the residual disc membranes that have not been completely digested are retained in the basal cell protoplasm, discharged outside the cell, and deposited on Bruch membrane , The formation of drusen.
- vascular endothelial growth factor vascular endothelial growth factor
- VEGF vascular endothelial growth factor
- VPF vascular permeability factor
- Intraocular injection of anti-vascular endothelial growth factor VEGF is an effective treatment for wet AMD, but anti-VEGF injection therapy is expensive and has strict indications, and the effect is different for different patients.
- Optical coherence tomography is a commonly used device for diagnosing ophthalmic diseases. It uses light reflection technology similar to ultrasound imaging to provide image reference for the detection and treatment of ophthalmic diseases.
- FIG. 1 is a schematic diagram of a network structure for predicting the curative effect of anti-vascular endothelial growth factor VEGF according to an embodiment of the application.
- the network structure in the embodiment of the present application includes a feature extraction network for feature extraction of OCT images, and a long and short-term memory artificial neural network for fusion of spatial information of multiple OCT images, and finally achieves resistance Prediction of the efficacy of VEGF.
- the multiple OCT images may be twelve-line OCT images, that is, 12 OCT images.
- the feature extraction network is a residual network structure, and the convolutional layer in the feature extraction network is used to extract Multi-scale features in each OCT image.
- the anti-VEGF curative effect is predicted, and the anti-VEGF curative effect prediction result is obtained, where the anti-VEGF curative effect prediction result includes vision improvement or vision deterioration.
- the anti-VEGF curative effect prediction results can provide effective references for doctors' treatment plans.
- the anti-VEGF curative effect prediction probability is less than the preset probability threshold, it is determined that the anti-VEGF curative effect prediction result is vision deterioration, indicating that anti-VEGF injection is not recommended.
- the anti-VEGF curative effect prediction result is determined to be improved vision, indicating that anti-VEGF injection is recommended.
- the anti-VEGF curative effect prediction probability is equal to the preset probability threshold, it is determined that the anti-VEGF curative effect prediction result is visual acuity deterioration or visual acuity improvement. Specifically, visual acuity deterioration or visual acuity improvement can be determined according to requirements.
- an anti-VEGF curative effect prediction method provided by an embodiment of the present application may include the following steps.
- An OCT image obtained by scanning the patient's eyes with an optical coherence tomography scanner OCT, and extract the regions collected for the macular area in the multiple initial OCT images to obtain the multiple images collected for the macular area.
- each of the multiple initial OCT images includes the patient Retinal tissue information.
- extract the image parts scanned for the macular area in the multiple initial OCT images where the macular area is located in the center of the retina, and the macular area is the most concentrated part of the central visual cells of the retina of the human eye.
- preprocessing may be performed on the multiple OCT images, which specifically includes: performing image correction on each OCT image, and performing contrast enhancement processing after correction.
- the image correction includes image tilt correction and/or image brightness correction.
- the multiple OCT images must first be preprocessed, including image correction processing and contrast enhancement processing, so as to correct too bright or too dark images, and The slanted image is corrected, and at the same time, the contrast of the image is improved, thereby improving the visual effect of the image.
- the multiple OCT images may be twelve-line OCT images, that is, 12 OCT images.
- the multiple OCT images are input to a feature extraction network to obtain multiple image features corresponding to the multiple OCT images.
- FIG. 3 is a schematic diagram of a feature extraction network provided by an embodiment of this application.
- the feature extraction network is a residual network structure.
- the input is directly passed to the output as the initial result through the shortcut connection.
- the convolutional layer in the feature extraction network is used to extract multi-scale features in each OCT image.
- the multiple image features corresponding to the multiple OCT images are first extracted through the feature extraction network.
- the multiple OCT images are first input into the feature extraction network, which is a residual network structure.
- the residual network is a type of convolutional neural network, which is easy to optimize and can be increased by Depth to improve accuracy, the residual block inside the residual network uses jump connections, which alleviates the problem of gradient disappearance caused by increasing depth in the deep neural network.
- the convolutional layer in the feature extraction network is used to extract the multi-scale features in each OCT image.
- the features of different scales reflect different image features
- the features of the shallower scale reflect the features of the shallower image, for example Edges, etc.
- deeper-scale features reflect deeper-level image features such as object contours.
- Feature extraction is performed on multiple OCT images through a feature extraction network that combines residual network structure and multi-scale feature extraction functions.
- feature extraction of multiple OCT images through the feature extraction network of residual network structure can reduce computational cost.
- And can also alleviate the problem of gradient disappearance caused by increasing depth in deep neural networks.
- the multi-scale features in the OCT image can be effectively obtained, which increases the richness and comprehensiveness of the extracted features, thereby making The forecast result is more accurate.
- spatial information fusion is performed on the multiple image features, thereby enriching the spatial information of the features.
- spatial information fusion first obtain the time corresponding to each OCT image, and sort the image features corresponding to each OCT image in chronological order according to the time corresponding to each OCT image, so as to obtain the corresponding multiple OCT images Time series of multiple image features.
- the time series of the multiple image features corresponding to the multiple OCT images are input into a long short-term memory artificial neural network (LSTM) to obtain the features corresponding to the multiple OCT images.
- LSTM long short-term memory artificial neural network
- the long and short-term memory artificial neural network is a time cyclic neural network, including forgetting gates, input gates and output gates.
- the image features extracted from the spatial sequence are input into the long and short-term memory artificial neural network in the form of time series, thereby fusing multiple image features corresponding to multiple OCT images of the same patient to enrich the spatial information of the features , Making the prediction method of the network closer to the process of doctors reading the pictures, so that the prediction results are more accurate.
- the anti-VEGF curative effect is classified into two categories to obtain the anti-VEGF curative effect prediction result, wherein the anti-VEGF curative effect prediction result includes improved vision or worsened vision.
- the anti-VEGF curative effect prediction results can provide effective references for doctors' treatment plans.
- the anti-VEGF curative effect prediction probability is less than the preset probability threshold, it is determined that the anti-VEGF curative effect prediction result is vision deterioration, indicating that anti-VEGF injection is not recommended.
- the anti-VEGF curative effect prediction probability is greater than the preset probability threshold, the anti-VEGF curative effect prediction result is determined to be improved vision, indicating that anti-VEGF injection is recommended.
- the anti-VEGF curative effect prediction probability is equal to the preset probability threshold, it is determined that the anti-VEGF curative effect prediction result is visual acuity deterioration or visual acuity improvement. Specifically, visual acuity deterioration or visual acuity improvement can be determined according to requirements.
- the anti-vascular endothelial growth factor VEGF curative effect prediction method provided by the embodiments of this application, after obtaining multiple OCT images, segmentation is not required, which reduces the cost of labeling in the segmentation network, and avoids inaccurate segmentation, etc.
- the accuracy error caused by anti-VEGF curative effect prediction the image features corresponding to each extracted OCT image include multi-scale features, which increases the richness and comprehensiveness of the extracted features, thereby improving the accuracy of anti-VEGF curative effect prediction.
- the spatial information fusion of multiple image features effectively utilizes the spatial characteristics of multiple OCT images, enriches the spatial information of the features, and also improves the accuracy of anti-VEGF curative effect prediction.
- the solution of this application can also be applied to the field of smart medical care.
- the anti-VEGF curative effect prediction method provided in this application is used to determine the anti-VEGF curative effect prediction result. Since the anti-VEGF curative effect prediction method provided by the present application can determine the anti-VEGF curative effect prediction results more accurately, this can provide a more accurate basis for the doctor's treatment plan, thereby improving the doctor's treatment efficiency and accuracy.
- FIG. 4 is a schematic flow chart of another anti-vascular endothelial growth factor VEGF curative effect prediction method provided by an embodiment of the application.
- another anti-vascular endothelial growth factor VEGF curative effect prediction method provided by the embodiment of the present application may include the following steps.
- An OCT image obtained by scanning the patient's eyes with an optical coherence tomography scanner OCT, and extract the regions collected for the macular area in the multiple initial OCT images to obtain the multiple images collected for the macular area.
- each of the multiple initial OCT images includes the patient Retinal tissue information.
- extract the image parts scanned for the macular area in the multiple initial OCT images where the macular area is located in the center of the retina, and the macular area is the most concentrated part of the central visual cells of the retina of the human eye.
- the multiple OCT images may be twelve-line OCT images, that is, 12 OCT images.
- Preprocessing the multiple OCT images includes: performing image correction on each OCT image, and performing contrast enhancement processing after correction.
- the image correction includes image tilt correction and/or image brightness correction.
- the multiple OCT images must first be preprocessed, including image correction processing and contrast enhancement processing, so as to correct too bright or too dark images, and The slanted image is corrected, and at the same time, the contrast of the image is improved, thereby improving the visual effect of the image.
- V represents the R, G, and B channels of each OCT image.
- V represents the R, G, and B channels of each OCT image.
- the R, G, and B channel data of each OCT image subtract the minimum value of all the data of the channel respectively, and then divide by the maximum value of all the data of the channel minus the minimum value, and finally The operation of multiplying by 255 to restore it to the value range of [0, 255].
- the R, G, and B three-channel data of each OCT image can be more evenly distributed from 0 to 255, which improves the contrast of the image, and achieves the purpose of improving the subjective visual effect of the image and enhancing the details of the image.
- the feature extraction network has a residual network structure, and the convolutional layer in the feature extraction network is used to extract multi-scale features in each OCT image.
- the multiple image features corresponding to the multiple OCT images are first extracted through the feature extraction network.
- the multiple OCT images are first input into the feature extraction network, which is a residual network structure.
- the residual network is a type of convolutional neural network, which is easy to optimize and can be increased by Depth to improve accuracy, the residual block inside the residual network uses jump connections, which alleviates the problem of gradient disappearance caused by increasing depth in the deep neural network.
- the convolutional layer in the feature extraction network is used to extract the multi-scale features in each OCT image.
- the features of different scales reflect different image features
- the features of the shallower scale reflect the features of the shallower image, for example Edges, etc.
- deeper-scale features reflect deeper-level image features such as object contours.
- the multiple OCT images are twelve-line OCT images (that is, 12 OCT images).
- the 12 OCT images are input into the feature extraction network, and 12 image features corresponding to the 12 OCT images are obtained.
- Feature extraction is performed on multiple OCT images through a feature extraction network that combines residual network structure and multi-scale feature extraction functions.
- feature extraction of multiple OCT images through the feature extraction network of residual network structure can reduce computational cost.
- And can also alleviate the problem of gradient disappearance caused by increasing depth in deep neural networks.
- the multi-scale features in the OCT image can be effectively obtained, which increases the richness and comprehensiveness of the extracted features, thereby making The forecast result is more accurate.
- the long and short-term memory artificial neural network is a time cyclic neural network, including forgetting gates, input gates and output gates.
- the image features extracted from the spatial sequence are input into the long and short-term memory artificial neural network in the form of time series, thereby fusing multiple image features corresponding to multiple OCT images of the same patient to enrich the spatial information of the features , Making the prediction method of the network closer to the process of doctors reading the pictures, so that the prediction results are more accurate.
- the anti-VEGF curative effect prediction probability is used to determine the anti-VEGF curative effect prediction result.
- the anti-VEGF curative effect prediction results include vision improvement or vision deterioration, and the anti-VEGF curative effect prediction results can provide an effective reference for the doctor's treatment plan.
- the anti-VEGF curative effect prediction probability is less than the preset probability threshold, it is determined that the anti-VEGF curative effect prediction result is vision deterioration.
- the anti-VEGF curative effect prediction probability is not less than the preset probability threshold, it is determined that the anti-VEGF curative effect prediction result is vision improvement.
- the feature extraction network that combines the residual network structure and the multi-scale feature extraction function is used to extract the features of multiple OCT images.
- the residual network structure The feature extraction network performs feature extraction on multiple OCT images, which can reduce the computational cost, and can also alleviate the problem of gradient disappearance caused by increasing depth in the deep neural network.
- the multi-scale features in the OCT image can be effectively obtained, which increases the richness and comprehensiveness of the extracted features, thereby making The forecast result is more accurate.
- the image features extracted from the spatial sequence are input into the long and short-term memory artificial neural network in the form of time series, thereby fusing multiple image features corresponding to multiple OCT images of the same patient to enrich the spatial information of the features , Making the prediction method of the network closer to the process of doctors reading the pictures, so that the prediction results are more accurate.
- an anti-vascular endothelial growth factor VEGF curative effect prediction device provided by an embodiment of the application.
- an anti-vascular endothelial growth factor VEGF curative effect prediction device provided by an embodiment of the present application may include the following modules.
- the acquiring module 501 is configured to acquire multiple OCT images of the optical coherence tomography acquired for the macula area.
- the feature extraction module 502 is configured to perform feature extraction on the multiple OCT images to obtain multiple image features corresponding to the multiple OCT images, where each of the multiple OCT images corresponds to the image feature Including the multi-scale features in each OCT image.
- the spatial information fusion module 503 is configured to perform spatial information fusion on the multiple image features corresponding to the multiple OCT images to obtain the features corresponding to the multiple OCT images.
- the determining module 504 is configured to determine the anti-VEGF curative effect prediction result according to the characteristics corresponding to the multiple OCT images.
- the acquiring module 501 is specifically configured to: acquire multiple initial OCT images obtained by scanning the eye with an optical coherence tomography scanner OCT; and extract the multiple initial OCT images for the macula. To obtain the multiple OCT images collected for the macular area.
- the device further includes a processing module configured to: perform image correction processing on the multiple OCT images to obtain the corrected multiple OCT images, wherein The image correction processing includes image tilt correction and/or image brightness correction; performing contrast enhancement processing on the corrected multiple OCT images to obtain the multiple OCT images after contrast enhancement.
- a processing module configured to: perform image correction processing on the multiple OCT images to obtain the corrected multiple OCT images, wherein The image correction processing includes image tilt correction and/or image brightness correction; performing contrast enhancement processing on the corrected multiple OCT images to obtain the multiple OCT images after contrast enhancement.
- the feature extraction module 502 is specifically configured to: input the multiple OCT images into a feature extraction network to obtain multiple image features corresponding to the multiple OCT images, wherein the feature The extraction network is a residual network structure, and the convolutional layer in the feature extraction network is used to extract the multi-scale features in each OCT image.
- the spatial information fusion module 503 is specifically configured to: obtain the time corresponding to each OCT image in the multiple OCT images; and determine the time corresponding to each OCT image according to the time corresponding to each OCT image.
- the time sequence of the multiple image features corresponding to the multiple OCT images; the time sequence of the multiple image features corresponding to the multiple OCT images is input into a long and short-term memory artificial neural network to obtain the features corresponding to the multiple OCT images.
- the determining module 504 is specifically configured to: perform a two-classification of the anti-VEGF curative effect according to the characteristics corresponding to the multiple OCT images to obtain the anti-VEGF curative effect prediction result, wherein the anti-VEGF curative effect
- the predictive results of VEGF curative effect include improved or worsened vision.
- the determining module 504 is specifically configured to: determine the anti-VEGF curative effect prediction probability according to the characteristics corresponding to the multiple OCT images; when the anti-VEGF curative effect prediction probability is less than a preset probability threshold , Determining that the anti-VEGF curative effect prediction result is vision deterioration; when the anti-VEGF curative effect prediction probability is greater than the preset probability threshold, it is determined that the anti-VEGF curative effect prediction result is vision improvement; when the anti-VEGF curative effect prediction probability When it is equal to the preset probability threshold, it is determined that the anti-VEGF curative effect prediction result is vision deterioration or vision improvement.
- anti-vascular endothelial growth factor VEGF curative effect prediction device for the specific implementation of the anti-vascular endothelial growth factor VEGF curative effect prediction device in the embodiments of the present application, please refer to the foregoing embodiments of the anti-vascular endothelial growth factor VEGF curative effect prediction method, which will not be repeated here.
- FIG. 6 is a schematic structural diagram of an electronic device in a hardware operating environment involved in an embodiment of the application.
- the electronic device of the hardware operating environment involved in the embodiment of the present application may include: a processor 601, such as a CPU.
- the memory 602 optionally, the memory may be a high-speed RAM memory, or a stable memory, such as a disk memory.
- the communication interface 603 is used to implement connection and communication between the processor 601 and the memory 602.
- FIG. 6 does not constitute a limitation on the electronic device, and may include more or fewer components than shown in the figure, or a combination of certain components, or different component arrangements. .
- the memory 602 may include an operating system, a network communication module, and an anti-vascular endothelial growth factor VEGF curative effect prediction program.
- the operating system is a program that manages and controls the hardware and software resources of electronic equipment, supports the operation of anti-vascular endothelial growth factor VEGF curative effect prediction programs and other software or programs.
- the network communication module is used to implement communication between various components in the memory 602 and communication with other hardware and software in the electronic device.
- the processor 601 is used to execute the anti-vascular endothelial growth factor VEGF curative effect prediction program stored in the memory 602 to implement the following steps: Obtain multiple optical coherence tomography OCT images collected for the macular area Perform feature extraction on the multiple OCT images to obtain multiple image features corresponding to the multiple OCT images, wherein the image feature corresponding to each OCT image in the multiple OCT images includes the each OCT Multi-scale features in the image; spatial information fusion is performed on the multiple image features corresponding to the multiple OCT images to obtain the features corresponding to the multiple OCT images; the anti-VEGF is determined according to the features corresponding to the multiple OCT images Efficacy prediction results.
- Another embodiment of the present application provides a computer-readable storage medium, the computer-readable storage medium stores a computer program, and the computer program is executed by a processor to implement the following steps: Obtain multiple optical coherence tomography images collected from the macula Instrument OCT image; perform feature extraction on the multiple OCT images to obtain multiple image features corresponding to the multiple OCT images, wherein the image feature corresponding to each OCT image in the multiple OCT images includes the Multi-scale features in each OCT image; spatial information fusion is performed on the multiple image features corresponding to the multiple OCT images to obtain the features corresponding to the multiple OCT images; according to the features corresponding to the multiple OCT images, Determine the predictive results of anti-VEGF efficacy.
- the storage medium involved in this application such as a computer-readable storage medium, may be non-volatile or volatile.
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Theoretical Computer Science (AREA)
- Public Health (AREA)
- Veterinary Medicine (AREA)
- Biophysics (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- General Physics & Mathematics (AREA)
- Radiology & Medical Imaging (AREA)
- Business, Economics & Management (AREA)
- Animal Behavior & Ethology (AREA)
- Ophthalmology & Optometry (AREA)
- Surgery (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Molecular Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Strategic Management (AREA)
- Quality & Reliability (AREA)
- Human Resources & Organizations (AREA)
- Economics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- General Business, Economics & Management (AREA)
- Tourism & Hospitality (AREA)
- Operations Research (AREA)
- Marketing (AREA)
- Entrepreneurship & Innovation (AREA)
- Game Theory and Decision Science (AREA)
- Development Economics (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Epidemiology (AREA)
Abstract
一种抗血管内皮生长因子VEGF疗效预测装置及方法。该装置包括:获取模块(501),用于获取针对黄斑区采集的多张光学相干断层扫描仪OCT图像;特征提取模块(502),用于对多张OCT图像进行特征提取,得到多张OCT图像对应的多个图像特征,其中,多张OCT图像中的每张OCT图像对应的图像特征包括每张OCT图像中的多尺度特征;空间信息融合模块(503),用于将多张OCT图像对应的多个图像特征进行空间信息融合,得到多张OCT图像对应的特征;确定模块(504),用于根据多张OCT图像对应的特征,确定抗VEGF疗效预测结果。该装置有利于提高抗VEGF疗效预测的精确度。
Description
本申请要求于2020年9月30日提交中国专利局、申请号为202011061099.8,发明名称为“一种抗血管内皮生长因子VEGF疗效预测装置及方法”的中国专利申请的优先权,其全部内容通过引用结合在本申请中。
本申请涉及医疗科技领域,尤其涉及一种抗血管内皮生长因子VEGF疗效预测装置及方法。
湿性老年黄斑变性(age-related macular degeneration,AMD)是一种主要的致盲性眼病。抗血管内皮生长因子(vascular endothelial growth factor,VEGF)眼内注射是对湿性AMD一种有效疗法,但是抗VEGF注射疗法的费用高昂,并且有严格的适应征,对于不同的患者疗效不一。由于缺乏有效的抗VEGF疗效预测,医生往往对患者采取统一的注射手段,导致对于部分不适用的患者也同样进行抗VEGF注射。因此,有效的抗VEGF疗效预测是医生的迫切需求。
光学相干断层扫描仪(optical coherence tomography,OCT)是目前常用的用于诊断眼科疾病的设备,利用类似超声波成像的光反射技术,对眼科疾病的检测、治疗等提供图像方面的参考。发明人意识到,目前通常使用的抗VEGF疗效预测方法中,通过分割网络对病灶区域(如积液、高反射点等)进行分割,分割后再进行抗VEGF疗效预测。但是,基于深度学习的分割网络在训练过程中,需要大量医生的标注,其标注的准确性以及分割网络分割的准确性都会影响疗效预测结果。此外,基于分割网络提取的特征中,丢失了大量可能提升抗VEGF疗效预测精确度的视网膜组织变化信息,导致抗VEGF疗效预测的精确度较低。
发明内容
本申请提供了一种抗血管内皮生长因子VEGF疗效预测装置及方法,有利于提高抗VEGF疗效预测的精确度。
本申请第一方面提供了一种抗血管内皮生长因子VEGF疗效预测装置,包括:获取模块,用于获取针对黄斑区采集的多张光学相干断层扫描仪OCT图像;特征提取模块,用于对所述多张OCT图像进行特征提取,得到所述多张OCT图像对应的多个图像特征,其中,所述多张OCT图像中的每张OCT图像对应的图像特征包括所述每张OCT图像中的多尺度特征;空间信息融合模块,用于将所述多张OCT图像对应的多个图像特征进行空间信息融合,得到所述多张OCT图像对应的特征;确定模块,用于根据所述多张OCT图像对应的特征,确定抗VEGF疗效预测结果。
本申请第二方面提供了一种抗血管内皮生长因子VEGF疗效预测方法,包括:获取针对黄斑区采集的多张光学相干断层扫描仪OCT图像;对所述多张OCT图像进行特征提取,得到所述多张OCT图像对应的多个图像特征,其中,所述多张OCT图像中的每张OCT图像对应的图像特征包括所述每张OCT图像中的多尺度特征;将所述多张OCT图像对应的多个图像特征进行空间信息融合,得到所述多张OCT图像对应的特征;根据所述多张OCT图像对应的特征,确定抗VEGF疗效预测结果。
本申请第三方面提供了一种电子设备,所述电子设备包括处理器、存储器、通信接口以及一个或多个程序,其中,所述一个或多个程序被存储在所述存储器中,并且被配置由所述处理器执行,以实现以下方法:获取针对黄斑区采集的多张光学相干断层扫描仪OCT图像;对所述多张OCT图像进行特征提取,得到所述多张OCT图像对应的多个图像特征,其中,所述多张OCT图像中的每张OCT图像对应的图像特征包括所述每张OCT图像中的多尺度特征;将所述多张OCT图像对应的多个图像特征进行空间信息融合,得到所述多张 OCT图像对应的特征;根据所述多张OCT图像对应的特征,确定抗VEGF疗效预测结果。
本申请第四方面提供了一种计算机可读存储介质,所述计算机可读存储介质存储有计算机程序,所述计算机程序被处理器执行以实现以下方法:获取针对黄斑区采集的多张光学相干断层扫描仪OCT图像;对所述多张OCT图像进行特征提取,得到所述多张OCT图像对应的多个图像特征,其中,所述多张OCT图像中的每张OCT图像对应的图像特征包括所述每张OCT图像中的多尺度特征;将所述多张OCT图像对应的多个图像特征进行空间信息融合,得到所述多张OCT图像对应的特征;根据所述多张OCT图像对应的特征,确定抗VEGF疗效预测结果。
本申请获取多张OCT图像后,不需要进行分割,减少了分割网络中的标注成本,同时避免了由于分割不精确等造成的抗VEGF疗效预测的精确度误差。另外,提取到的每张OCT图像对应的图像特征包括多尺度特征,增加了提取的特征的丰富性和全面性,从而提高了抗VEGF疗效预测的精确度。并且,将多个图像特征进行空间信息融合,有效利用了多张OCT图像的空间特性,丰富了特征的空间信息,也提高了抗VEGF疗效预测的精确度。
为了更清楚地说明本申请实施例中的技术方案,下面将对实施例中所需使用的附图作简单地介绍,显而易见地,下面描述中的附图仅仅是本申请的一些实施例,对于本领域普通技术人员来讲,在不付出创造性劳动的前提下,还可以根据这些附图获得其他的附图。
图1为本申请实施例提供的一种抗血管内皮生长因子VEGF疗效预测的网结构示意图。
图2为本申请实施例提供的一种抗血管内皮生长因子VEGF疗效预测方法的流程示意图。
图3为本申请实施例提供的一种特征提取网络的示意图。
图4为本申请实施例提供的另一种抗血管内皮生长因子VEGF疗效预测方法的流程示意图。
图5为本申请实施例提供的一种抗血管内皮生长因子VEGF疗效预测装置的示意图。
图6为本申请实施例涉及的硬件运行环境的电子设备结构示意图。
本申请实施例提供的抗血管内皮生长因子VEGF疗效预测装置及方法,有利于提高抗VEGF疗效预测的精确度。
为了使本技术领域的人员更好地理解本申请方案,下面将结合本申请实施例中的附图,对本申请实施例中的技术方案进行清楚、完整地描述,显然,所描述的实施例仅仅是本申请一部分的实施例,而不是全部的实施例。基于本申请中的实施例,本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例,都应当属于本申请保护的范围。
本申请实施例中涉及的“至少一个”是指一个或者多个,“多个”是指两个或两个以上。“和/或”,描述关联对象的关联关系,表示可以存在三种关系,例如,A和/或B,可以表示:单独存在A,同时存在A和B,单独存在B的情况,其中A,B可以是单数或者复数。字符“/”一般表示前后关联对象是一种“或”的关系。“以下至少一项(个)”或其类似表达,是指的这些项中的任意组合,包括单项(个)或复数项(个)的任意组合。
本申请的说明书和权利要求书及上述附图中的术语“第一”、“第二”、“第三”、“第四”等是用于区别不同对象,而不是用于描述特定顺序、时序、优先级或者重要程度。例如,第一信息和第二信息,只是为了区分不同的信息,而并不是表示这两种信息的内容、优先级、发送顺序或者重要程度的不同。此外,术语“包括”和“具有”以及它们任何变形,意图在于覆盖不排他的包含。例如包含了一系列步骤或单元的过程、方法、***、产品或设备没有限定于已列出的步骤或单元,而是可选地还包括没有列出的步骤或单元,或可选地还包括对于这些过程、方法、产品或设备固有的其它步骤或单元。
本申请的技术方案可应用于人工智能、智慧城市、数字医疗和/或区块链技术领域,以实现智慧医疗。可选的,本申请涉及的数据如图像、特征和/或预测结果等可存储于数据库中,或者可以存储于区块链中,比如通过区块链分布式存储,本申请不做限定。
为了便于理解本申请,首先对本申请涉及的概念进行解释。
湿性老年黄斑变性(age-related macular degeneration,AMD):AMD是一种主要的致盲性眼病,为黄斑区结构的衰老性改变。主要表现为视网膜色素上皮细胞对视细胞外节盘膜吞噬消化能力下降,结果使未被完全消化的盘膜残余小体潴留于基底部细胞原浆中,并向细胞外排出,沉积于Bruch膜,形成玻璃膜疣。
血管内皮生长因子(vascular endothelial growth factor,VEGF):VEGF又称血管通透因子(vascular permeability factor,VPF),是一种高度特异性的促血管内皮细胞生长因子,具有促进血管通透性增加、细胞外基质变性、血管内皮细胞迁移、增殖和血管形成等作用。抗血管内皮生长因子VEGF眼内注射是对湿性AMD的一种有效疗法,但是抗VEGF注射疗法的费用高昂,并且有严格的适应征,对于不同的患者疗效不一。
光学相干断层扫描仪(optical coherence tomography,OCT)是目前常用的用于诊断眼科疾病的设备,利用类似超声波成像的光反射技术,对眼科疾病的检测、治疗等提供图像方面的参考。
如上介绍了本申请的背景技术,下面介绍本申请实施例的技术特征。
首先参见图1,图1为本申请实施例提供的一种抗血管内皮生长因子VEGF疗效预测的网结构示意图。如图1所示,本申请实施例中的网络结构中,包括对OCT图像进行特征提取的特征提取网络,以及针对多张OCT图像的空间信息进行融合的长短期记忆人工神经网络,最终实现抗VEGF疗效的预测。
首先,获取针对黄斑区采集的多张光学相干断层扫描仪OCT图像。在一种可能的实施方式中,该多张OCT图像可以是十二线OCT图像,也即12张OCT图像。
其次,将该多张OCT图像输入特征提取网络,得到该多张OCT图像对应的多个图像特征,其中,该特征提取网络为残差网络结构,该特征提取网络中的卷积层用于提取每张OCT图像中的多尺度特征。
然后,将该多张OCT图像对应的多个图像特征的时间序列输入长短期记忆人工神经网络,得到该多张OCT图像对应的特征,其中,长短期记忆人工神经网络用于针对多张OCT图像的空间信息进行融合。
最后,根据该多张OCT图像对应的特征,对抗VEGF疗效进行预测,得到抗VEGF疗效预测结果,其中,抗VEGF疗效预测结果包括视力提升或视力恶化。具体的,抗VEGF疗效预测结果可以为医生的治疗方案提供有效参考。在进行抗VEGF疗效预测时,根据多张OCT图像对应的特征,确定抗VEGF疗效预测概率。当抗VEGF疗效预测概率小于预设概率阈值时,确定抗VEGF疗效预测结果为视力恶化,表示不建议进行抗VEGF注射。当抗VEGF疗效预测概率大于预设概率阈值时,确定抗VEGF疗效预测结果为视力提升,表示建议进行抗VEGF注射。当抗VEGF疗效预测概率等于预设概率阈值时,确定抗VEGF疗效预测结果为视力恶化或者视力提升,具体为视力恶化还是视力提升可以根据需求确定。
参见图2,图2为本申请实施例提供的一种抗血管内皮生长因子VEGF疗效预测方法的流程示意图。其中,如图2所示,本申请实施例提供的一种抗VEGF疗效预测方法可以包括以下步骤。
201、获取针对黄斑区采集的多张光学相干断层扫描仪OCT图像。
可选的,获取通过光学相干断层扫描仪OCT对患者的眼部进行扫描得到的多张初始OCT图像,提取该多张初始OCT图像中针对黄斑区采集的区域,以得到针对黄斑区采集的多张OCT图像。
具体的,在对患者进行抗VEGF疗效预测时,首先使用OCT设备对该患者的眼部进行扫描,可以得到多张初始OCT图像,其中,该多张初始OCT图像中的每张OCT图像包括患者的视网膜组织信息。得到该多张初始OCT图像后,提取该多张初始OCT图像中针对黄斑区扫描到的图像部分,其中,黄斑区位于视网膜中央,黄斑区是人眼视网膜中央视觉细胞最集中的部位。通过提取针对黄斑区扫描到的区域,可以去除***的不含有效信息的区域,从而在之后的计算过程中节省计算时间。
可选的,获取了针对黄斑区采集到的多张OCT图像后,还可以对该多张OCT图像进行预处理,具体包括:对每张OCT图像进行图像校正,校正后进行对比度增强处理。其中,图像校正包括图像倾斜校正和/或图像亮度校正。
具体的,在对患者眼部进行扫描时,受外界因素影响例如光线影响、对患者眼部进行扫描时患者眼部的角度变化等影响,得到的多张OCT图像中可能会有部分图像出现过亮或者过暗的情况,或者出现图像倾斜的情况,这些情况不利于后续对OCT图像进行进一步处理。因此,在对该多张OCT图像进行特征提取之前,首先要对该多张OCT图像进行预处理,包括进行图像校正处理和对比度增强处理,从而对过亮或过暗的图像进行校正,以及对倾斜的图像进行校正,同时,提高图像的对比度,从而改善图像的视觉效果。
在一种可能的实施方式中,该多张OCT图像可以是十二线OCT图像,也即12张OCT图像。
202、对所述多张OCT图像进行特征提取,得到所述多张OCT图像对应的多个图像特征,其中,所述多张OCT图像中的每张OCT图像对应的图像特征包括所述每张OCT图像中的多尺度特征。
可选的,将所述多张OCT图像输入特征提取网络,得到所述多张OCT图像对应的多个图像特征。参见图3,图3为本申请实施例提供的一种特征提取网络的示意图。如图3所示,该特征提取网络为残差网络结构,在残差网络结构中,通过捷径连接的方式,直接将输入传到输出作为初始结果。其中,该特征提取网络中的卷积层用于提取每张OCT图像中的多尺度特征。
具体的,在确定抗VEGF疗效预测结果之前,首先通过特征提取网络提取该多张OCT图像对应的多个图像特征。在特征提取过程中,首先将该多张OCT图像输入特征提取网络,该特征提取网络为残差网络结构,其中,残差网络为卷积神经网络中的一种,容易优化,并且能够通过增加深度来提高准确率,残差网络内部的残差块使用跳跃连接,缓解了在深度神经网络中增加深度带来的梯度消失问题。该特征提取网络中的卷积层用于提取每张OCT图像中的多尺度特征,其中,不同尺度的特征反映的图像特征不同,较浅尺度的特征反映了较浅层次的图像特征,例如边缘等,较深尺度的特征反映了较深层次的图像特征例如物体轮廓等。
通过结合残差网络结构和多尺度特征提取功能的特征提取网络对多张OCT图像进行特征提取,一方面,通过残差网络结构的特征提取网络对多张OCT图像进行特征提取,可以减少计算成本,并且还可以缓解深度神经网络中增加深度带来的梯度消失问题。另一方面,通过具有多尺度特征提取功能的特征提取网络对多张OCT图像进行特征提取,可以有效获得该OCT图像中的多尺度特征,增加了提取的特征的丰富性和全面性,从而使得预测结果更精确。
203、将所述多张OCT图像对应的多个图像特征进行空间信息融合,得到所述多张OCT图像对应的特征。
可选的,获取所述多张OCT图像中的每张OCT图像对应的时间;根据所述每张OCT图像对应的时间,确定所述多张OCT图像对应的多个图像特征的时间序列;将所述多张OCT图像对应的多个图像特征的时间序列输入长短期记忆人工神经网络,得到所述多张 OCT图像对应的特征。
具体的,通过特征提取网络得到多张OCT图像对应的多个图像特征之后,对该多个图像特征进行空间信息融合,从而丰富特征的空间信息。在进行空间信息融合时,首先获取每张OCT图像对应的时间,根据每张OCT图像对应的时间,按照时间先后顺序对每张OCT图像对应的图像特征进行排序,从而得到该多张OCT图像对应的多个图像特征的时间序列。最后,将该多张OCT图像对应的多个图像特征的时间序列输入长短期记忆人工神经网络(long short-term memory,LSTM),得到该多张OCT图像对应的特征。其中,长短期记忆人工神经网络是一种时间循环神经网络,包括遗忘门、输入门和输出门等。将空间序列中提取到的图像特征以时间序列的形式输入长短期记忆人工神经网络中,从而将同一患者的多张OCT图像对应的多个图像特征进行了空间信息融合,丰富了特征的空间信息,使得网络的预测方式更接近医生阅片的过程,从而使得预测结果更精确。
204、根据所述多张OCT图像对应的特征,确定抗VEGF疗效预测结果。
可选的,根据所述多张OCT图像对应的特征,对抗VEGF疗效进行二分类,得到所述抗VEGF疗效预测结果,其中,所述抗VEGF疗效预测结果包括视力提升或视力恶化。
具体的,抗VEGF疗效预测结果可以为医生的治疗方案提供有效参考。在进行抗VEGF疗效预测时,根据多张OCT图像对应的特征,确定抗VEGF疗效预测概率。当抗VEGF疗效预测概率小于预设概率阈值时,确定抗VEGF疗效预测结果为视力恶化,表示不建议进行抗VEGF注射。当抗VEGF疗效预测概率大于预设概率阈值时,确定抗VEGF疗效预测结果为视力提升,表示建议进行抗VEGF注射。当抗VEGF疗效预测概率等于预设概率阈值时,确定抗VEGF疗效预测结果为视力恶化或者视力提升,具体为视力恶化还是视力提升可以根据需求确定。
可以看出,通过本申请实施例提供的抗血管内皮生长因子VEGF疗效预测方法,获取多张OCT图像后,不需要进行分割,减少了分割网络中的标注成本,同时避免了由于分割不精确等造成的抗VEGF疗效预测的精确度误差。另外,提取到的每张OCT图像对应的图像特征包括多尺度特征,增加了提取的特征的丰富性和全面性,从而提高了抗VEGF疗效预测的精确度。并且,将多个图像特征进行空间信息融合,有效利用了多张OCT图像的空间特性,丰富了特征的空间信息,也提高了抗VEGF疗效预测的精确度。
在本申请的一个实施方式中,本申请的方案还可以应用到智慧医疗领域。比如,接收医生输入的多张OCT图像,通过本申请提供的抗VEGF疗效预测方法,确定抗VEGF疗效预测结果。由于通过本申请提供的抗VEGF疗效预测方法,可以较为精准的确定抗VEGF疗效预测结果,这样可以为医生的治疗方案提供较为精准的判断依据,从而提高医生的治疗效率和精准度。
参见图4,图4为本申请实施例提供的另一种抗血管内皮生长因子VEGF疗效预测方法的流程示意图。其中,如图4所示,本申请实施例提供的另一种抗血管内皮生长因子VEGF疗效预测方法可以包括以下步骤。
401、获取针对黄斑区采集的多张光学相干断层扫描仪OCT图像。
可选的,获取通过光学相干断层扫描仪OCT对患者的眼部进行扫描得到的多张初始OCT图像,提取该多张初始OCT图像中针对黄斑区采集的区域,以得到针对黄斑区采集的多张OCT图像。
具体的,在对患者进行抗VEGF疗效预测时,首先使用OCT设备对该患者的眼部进行扫描,可以得到多张初始OCT图像,其中,该多张初始OCT图像中的每张OCT图像包括患者的视网膜组织信息。得到该多张初始OCT图像后,提取该多张初始OCT图像中针对黄斑区扫描到的图像部分,其中,黄斑区位于视网膜中央,黄斑区是人眼视网膜中央视觉细胞最集中的部位。通过提取针对黄斑区扫描到的区域,可以去除***的不含有效信息的 区域,从而在之后的计算过程中节省计算时间。
在一种可能的实施方式中,该多张OCT图像可以是十二线OCT图像,也即12张OCT图像。
402、对该多张OCT图像进行预处理。
对该多张OCT图像进行预处理包括:对每张OCT图像进行图像校正,校正后进行对比度增强处理。其中,图像校正包括图像倾斜校正和/或图像亮度校正。
具体的,在对患者眼部进行扫描时,受外界因素影响例如光线影响、对患者眼部进行扫描时患者眼部的角度变化等影响,得到的多张OCT图像中可能会有部分图像出现过亮或者过暗的情况,或者出现图像倾斜的情况,这些情况不利于后续对OCT图像进行进一步处理。因此,在对该多张OCT图像进行特征提取之前,首先要对该多张OCT图像进行预处理,包括进行图像校正处理和对比度增强处理,从而对过亮或过暗的图像进行校正,以及对倾斜的图像进行校正,同时,提高图像的对比度,从而改善图像的视觉效果。
在一种可能的实施方式中,对该多张OCT图像进行图像亮度校正的方法可以是,对该多张OCT图像进行伽马Gamma变换,其中,对每张OCT图像进行Gamma变换的公式为V
out=V
in
γ,其中,V分别代表每张OCT图像的R、G、B三通道。
在一种可能的实施方式中,对该多张OCT图像进行对比度增强处理的公式为
其中,V分别代表每张OCT图像的R、G、B三通道。也就是说,对于每张OCT图像的R、G、B通道数据,分别对其进行减去该通道所有数据中的最小值,然后除以该通道所有数据中的最大值减最小值,最后再乘以255将其恢复到[0,255]的取值范围的操作。这样可以使得每张OCT图像的R、G、B三通道数据在0-255上分布更均匀,提高了图像的对比度,达到改善图像主观视觉效果以及增强图像细节的目的。
403、将该多张OCT图像输入特征提取网络,得到该多张OCT图像对应的多个图像特征。
具体的,该特征提取网络为残差网络结构,该特征提取网络中的卷积层用于提取每张OCT图像中的多尺度特征。
在确定抗VEGF疗效预测结果之前,首先通过特征提取网络提取该多张OCT图像对应的多个图像特征。在特征提取过程中,首先将该多张OCT图像输入特征提取网络,该特征提取网络为残差网络结构,其中,残差网络为卷积神经网络中的一种,容易优化,并且能够通过增加深度来提高准确率,残差网络内部的残差块使用跳跃连接,缓解了在深度神经网络中增加深度带来的梯度消失问题。该特征提取网络中的卷积层用于提取每张OCT图像中的多尺度特征,其中,不同尺度的特征反映的图像特征不同,较浅尺度的特征反映了较浅层次的图像特征,例如边缘等,较深尺度的特征反映了较深层次的图像特征例如物体轮廓等。
在一种可能的实施方式中,该多张OCT图像为十二线OCT图像(也即12张OCT图像)。将12张OCT图像输入特征提取网络,得到该12张OCT图像对应的12个图像特征。
通过结合残差网络结构和多尺度特征提取功能的特征提取网络对多张OCT图像进行特征提取,一方面,通过残差网络结构的特征提取网络对多张OCT图像进行特征提取,可以减少计算成本,并且还可以缓解深度神经网络中增加深度带来的梯度消失问题。另一方面,通过具有多尺度特征提取功能的特征提取网络对多张OCT图像进行特征提取,可以有效获得该OCT图像中的多尺度特征,增加了提取的特征的丰富性和全面性,从而使得预测结果更精确。
404、确定该多张OCT图像对应的多个图像特征的时间序列。
具体的,在进行空间信息融合时,首先获取每张OCT图像对应的时间,根据每张OCT图像对应的时间,按照时间先后顺序对每张OCT图像对应的图像特征进行排序,从而得到该多张OCT图像对应的多个图像特征的时间序列。
405、将该多张OCT图像对应的多个图像特征的时间序列输入长短期记忆人工神经网络,得到该多张OCT图像对应的特征。
具体的,通过特征提取网络得到多张OCT图像对应的多个图像特征之后,对该多个图像特征进行空间信息融合,从而丰富特征的空间信息。得到该多张OCT图像对应的多个图像特征的时间序列后,将该多张OCT图像对应的多个图像特征的时间序列输入长短期记忆人工神经网络(long short-term memory,LSTM),得到该多张OCT图像对应的特征。其中,长短期记忆人工神经网络是一种时间循环神经网络,包括遗忘门、输入门和输出门等。将空间序列中提取到的图像特征以时间序列的形式输入长短期记忆人工神经网络中,从而将同一患者的多张OCT图像对应的多个图像特征进行了空间信息融合,丰富了特征的空间信息,使得网络的预测方式更接近医生阅片的过程,从而使得预测结果更精确。
406、根据该多张OCT图像对应的特征,确定抗VEGF疗效预测概率。
具体的,抗VEGF疗效预测概率用于确定抗VEGF疗效预测结果。其中,抗VEGF疗效预测结果包括视力提升或视力恶化,抗VEGF疗效预测结果可以为医生的治疗方案提供有效参考。
407、判断抗VEGF疗效预测概率是否小于预设概率阈值。
408、当抗VEGF疗效预测概率小于预设概率阈值时,确定抗VEGF疗效预测结果为视力恶化。
抗VEGF疗效预测结果为视力恶化时,表示不建议进行抗VEGF注射。
409、当抗VEGF疗效预测概率不小于预设概率阈值时,确定抗VEGF疗效预测结果为视力提升。
抗VEGF疗效预测结果为视力提升时,表示建议进行抗VEGF注射。
可以看出,通过本申请实施例提供的抗VEGF疗效预测方法,通过结合残差网络结构和多尺度特征提取功能的特征提取网络对多张OCT图像进行特征提取,一方面,通过残差网络结构的特征提取网络对多张OCT图像进行特征提取,可以减少计算成本,并且还可以缓解深度神经网络中增加深度带来的梯度消失问题。另一方面,通过具有多尺度特征提取功能的特征提取网络对多张OCT图像进行特征提取,可以有效获得该OCT图像中的多尺度特征,增加了提取的特征的丰富性和全面性,从而使得预测结果更精确。将空间序列中提取到的图像特征以时间序列的形式输入长短期记忆人工神经网络中,从而将同一患者的多张OCT图像对应的多个图像特征进行了空间信息融合,丰富了特征的空间信息,使得网络的预测方式更接近医生阅片的过程,从而使得预测结果更精确。
参见图5,图5为本申请实施例提供的一种抗血管内皮生长因子VEGF疗效预测装置的示意图。其中,如图5所示,本申请实施例提供的一种抗血管内皮生长因子VEGF疗效预测装置可以包括以下模块。
获取模块501,用于获取针对黄斑区采集的多张光学相干断层扫描仪OCT图像。
特征提取模块502,用于对所述多张OCT图像进行特征提取,得到所述多张OCT图像对应的多个图像特征,其中,所述多张OCT图像中的每张OCT图像对应的图像特征包括所述每张OCT图像中的多尺度特征。
空间信息融合模块503,用于将所述多张OCT图像对应的多个图像特征进行空间信息融合,得到所述多张OCT图像对应的特征。
确定模块504,用于根据所述多张OCT图像对应的特征,确定抗VEGF疗效预测结果。
在一种可能的实施方式中,所述获取模块501具体用于:获取通过光学相干断层扫描仪OCT对眼部进行扫描得到的多张初始OCT图像;提取所述多张初始OCT图像中针对黄斑区采集的区域,以得到针对黄斑区采集的所述多张OCT图像。
在一种可能的实施方式中,所述装置还包括处理模块,所述处理模块用于:对所述多张OCT图像进行图像校正处理,得到校正后的所述多张OCT图像,其中,所述图像校正处理包括图像倾斜校正和/或图像亮度校正;对校正后的所述多张OCT图像进行对比度增强处理,得到对比度增强后的所述多张OCT图像。
在一种可能的实施方式中,所述特征提取模块502具体用于:将所述多张OCT图像输入特征提取网络,得到所述多张OCT图像对应的多个图像特征,其中,所述特征提取网络为残差网络结构,所述特征提取网络中的卷积层用于提取所述每张OCT图像中的多尺度特征。
在一种可能的实施方式中,所述空间信息融合模块503具体用于:获取所述多张OCT图像中的每张OCT图像对应的时间;根据所述每张OCT图像对应的时间,确定所述多张OCT图像对应的多个图像特征的时间序列;将所述多张OCT图像对应的多个图像特征的时间序列输入长短期记忆人工神经网络,得到所述多张OCT图像对应的特征。
在一种可能的实施方式中,所述确定模块504具体用于:根据所述多张OCT图像对应的特征,对抗VEGF疗效进行二分类,得到所述抗VEGF疗效预测结果,其中,所述抗VEGF疗效预测结果包括视力提升或视力恶化。
在一种可能的实施方式中,所述确定模块504具体用于:根据所述多张OCT图像对应的特征,确定抗VEGF疗效预测概率;当所述抗VEGF疗效预测概率小于预设概率阈值时,确定所述抗VEGF疗效预测结果为视力恶化;当所述抗VEGF疗效预测概率大于所述预设概率阈值时,确定所述抗VEGF疗效预测结果为视力提升;当所述抗VEGF疗效预测概率等于所述预设概率阈值时,确定所述抗VEGF疗效预测结果为视力恶化或者视力提升。
本申请实施例中抗血管内皮生长因子VEGF疗效预测装置的具体实施可参见上述抗血管内皮生长因子VEGF疗效预测方法的各实施例,在此不做赘述。
参见图6,图6为本申请的实施例涉及的硬件运行环境的电子设备结构示意图。其中,如图6所示,本申请的实施例涉及的硬件运行环境的电子设备可以包括:处理器601,例如CPU。存储器602,可选的,存储器可以为高速RAM存储器,也可以是稳定的存储器,例如磁盘存储器。通信接口603,用于实现处理器601和存储器602之间的连接通信。
本领域技术人员可以理解,图6中示出的电子设备的结构并不构成对电子设备的限定,可以包括比图示更多或更少的部件,或者组合某些部件,或者不同的部件布置。
如图6所示,存储器602中可以包括操作***、网络通信模块以及抗血管内皮生长因子VEGF疗效预测程序。操作***是管理和控制电子设备硬件和软件资源的程序,支持抗血管内皮生长因子VEGF疗效预测程序以及其他软件或程序的运行。网络通信模块用于实现存储器602内部各组件之间的通信,以及与电子设备中其他硬件和软件之间通信。
在图6所示的电子设备中,处理器601用于执行存储器602中存储的抗血管内皮生长因子VEGF疗效预测程序,实现以下步骤:获取针对黄斑区采集的多张光学相干断层扫描仪OCT图像;对所述多张OCT图像进行特征提取,得到所述多张OCT图像对应的多个图像特征,其中,所述多张OCT图像中的每张OCT图像对应的图像特征包括所述每张OCT图像中的多尺度特征;将所述多张OCT图像对应的多个图像特征进行空间信息融合,得到所述多张OCT图像对应的特征;根据所述多张OCT图像对应的特征,确定抗VEGF疗效预测结果。
本申请实施例中电子设备的具体实施可参见上述抗血管内皮生长因子VEGF疗效预测方法的各实施例,在此不做赘述。
本申请的另一个实施例提供了一种计算机可读存储介质,计算机可读存储介质存储有计算机程序,计算机程序被处理器执行以实现以下步骤:获取针对黄斑区采集的多张光学相干断层扫描仪OCT图像;对所述多张OCT图像进行特征提取,得到所述多张OCT图像对应的多个图像特征,其中,所述多张OCT图像中的每张OCT图像对应的图像特征包括所述每张OCT图像中的多尺度特征;将所述多张OCT图像对应的多个图像特征进行空间信息融合,得到所述多张OCT图像对应的特征;根据所述多张OCT图像对应的特征,确定抗VEGF疗效预测结果。
本申请实施例中计算机可读存储介质的具体实施可参见上述抗血管内皮生长因子VEGF疗效预测方法的各实施例,在此不做赘述。
可选的,本申请涉及的存储介质如计算机可读存储介质可以是非易失性的,也可以是易失性的。
还需要说明的是,对于前述的各方法实施例,为了简单描述,故将其都表述为一系列的动作组合,但是本领域技术人员应该知悉,本申请并不受所描述的动作顺序的限制,因为依据本申请,某些步骤可以采用其他顺序或者同时进行。其次,本领域技术人员也应该知悉,说明书中所描述的实施例均属于优选实施例,所涉及的动作和模块并不一定是本申请所必须的。在上述实施例中,对各个实施例的描述都各有侧重,某个实施例中没有详述的部分,可以参见其他实施例的相关描述。
以上所述,以上实施例仅用以说明本申请的技术方案,而非对其限制;尽管参照前述实施例对本申请进行了详细的说明,本领域的普通技术人员应当理解:其依然可以对前述各实施例所记载的技术方案进行修改,或者对其中部分技术特征进行等同替换;而这些修改或者替换,并不使相应技术方案的本质脱离本申请各实施例技术方案的范围。
Claims (20)
- 一种抗血管内皮生长因子VEGF疗效预测装置,包括:获取模块,用于获取针对黄斑区采集的多张光学相干断层扫描仪OCT图像;特征提取模块,用于对所述多张OCT图像进行特征提取,得到所述多张OCT图像对应的多个图像特征,其中,所述多张OCT图像中的每张OCT图像对应的图像特征包括所述每张OCT图像中的多尺度特征;空间信息融合模块,用于将所述多张OCT图像对应的多个图像特征进行空间信息融合,得到所述多张OCT图像对应的特征;确定模块,用于根据所述多张OCT图像对应的特征,确定抗VEGF疗效预测结果。
- 根据权利要求1所述的装置,其中,所述获取模块具体用于:获取通过光学相干断层扫描仪OCT对眼部进行扫描得到的多张初始OCT图像;提取所述多张初始OCT图像中针对黄斑区采集的区域,以得到针对黄斑区采集的所述多张OCT图像。
- 根据权利要求1所述的装置,其中,所述装置还包括处理模块,所述处理模块用于:对所述多张OCT图像进行图像校正处理,得到校正后的所述多张OCT图像,其中,所述图像校正处理包括图像倾斜校正和/或图像亮度校正;对校正后的所述多张OCT图像进行对比度增强处理,得到对比度增强后的所述多张OCT图像。
- 根据权利要求1至3任一项所述的装置,其中,所述特征提取模块具体用于:将所述多张OCT图像输入特征提取网络,得到所述多张OCT图像对应的多个图像特征,其中,所述特征提取网络为残差网络结构,所述特征提取网络中的卷积层用于提取所述每张OCT图像中的多尺度特征。
- 根据权利要求1至3任一项所述的装置,其中,所述空间信息融合模块具体用于:获取所述多张OCT图像中的每张OCT图像对应的时间;根据所述每张OCT图像对应的时间,确定所述多张OCT图像对应的多个图像特征的时间序列;将所述多张OCT图像对应的多个图像特征的时间序列输入长短期记忆人工神经网络,得到所述多张OCT图像对应的特征。
- 根据权利要求1至3任一项所述的装置,其中,所述确定模块具体用于:根据所述多张OCT图像对应的特征,对抗VEGF疗效进行二分类,得到所述抗VEGF疗效预测结果,其中,所述抗VEGF疗效预测结果包括视力提升或视力恶化。
- 根据权利要求6所述的装置,其中,所述确定模块具体用于:根据所述多张OCT图像对应的特征,确定抗VEGF疗效预测概率;当所述抗VEGF疗效预测概率小于预设概率阈值时,确定所述抗VEGF疗效预测结果为视力恶化;当所述抗VEGF疗效预测概率大于所述预设概率阈值时,确定所述抗VEGF疗效预测结果为视力提升;当所述抗VEGF疗效预测概率等于所述预设概率阈值时,确定所述抗VEGF疗效预测结果为视力恶化或者视力提升。
- 一种抗血管内皮生长因子VEGF疗效预测方法,包括:获取针对黄斑区采集的多张光学相干断层扫描仪OCT图像;对所述多张OCT图像进行特征提取,得到所述多张OCT图像对应的多个图像特征,其中,所述多张OCT图像中的每张OCT图像对应的图像特征包括所述每张OCT图像中的多尺度特征;将所述多张OCT图像对应的多个图像特征进行空间信息融合,得到所述多张OCT图像对应的特征;根据所述多张OCT图像对应的特征,确定抗VEGF疗效预测结果。
- 一种电子设备,所述电子设备包括处理器、存储器、通信接口以及一个或多个程序,其中,所述一个或多个程序被存储在所述存储器中,并且被配置由所述处理器执行,以实现以下方法:获取针对黄斑区采集的多张光学相干断层扫描仪OCT图像;对所述多张OCT图像进行特征提取,得到所述多张OCT图像对应的多个图像特征,其中,所述多张OCT图像中的每张OCT图像对应的图像特征包括所述每张OCT图像中的多尺度特征;将所述多张OCT图像对应的多个图像特征进行空间信息融合,得到所述多张OCT图像对应的特征;根据所述多张OCT图像对应的特征,确定抗VEGF疗效预测结果。
- 根据权利要求9所述的电子设备,其中,所述获取针对黄斑区采集的多张光学相干断层扫描仪OCT图像时,具体实现:获取通过光学相干断层扫描仪OCT对眼部进行扫描得到的多张初始OCT图像;提取所述多张初始OCT图像中针对黄斑区采集的区域,以得到针对黄斑区采集的所述多张OCT图像。
- 根据权利要求9所述的电子设备,其中,所述处理器还用于执行:对所述多张OCT图像进行图像校正处理,得到校正后的所述多张OCT图像,其中,所述图像校正处理包括图像倾斜校正和/或图像亮度校正;对校正后的所述多张OCT图像进行对比度增强处理,得到对比度增强后的所述多张OCT图像。
- 根据权利要求9至11任一项所述的电子设备,其中,所述对所述多张OCT图像进行特征提取,得到所述多张OCT图像对应的多个图像特征时,具体实现:将所述多张OCT图像输入特征提取网络,得到所述多张OCT图像对应的多个图像特征,其中,所述特征提取网络为残差网络结构,所述特征提取网络中的卷积层用于提取所述每张OCT图像中的多尺度特征。
- 根据权利要求9至11任一项所述的电子设备,其中,所述将所述多张OCT图像对应的多个图像特征进行空间信息融合,得到所述多张OCT图像对应的特征时,具体实现:获取所述多张OCT图像中的每张OCT图像对应的时间;根据所述每张OCT图像对应的时间,确定所述多张OCT图像对应的多个图像特征的时间序列;将所述多张OCT图像对应的多个图像特征的时间序列输入长短期记忆人工神经网络,得到所述多张OCT图像对应的特征。
- 根据权利要求9至11任一项所述的电子设备,其中,所述根据所述多张OCT图像对应的特征,确定抗VEGF疗效预测结果时,具体实现:根据所述多张OCT图像对应的特征,对抗VEGF疗效进行二分类,得到所述抗VEGF疗效预测结果,其中,所述抗VEGF疗效预测结果包括视力提升或视力恶化。
- 一种计算机可读存储介质,所述计算机可读存储介质存储有计算机程序,所述计算机程序被处理器执行以实现以下方法:获取针对黄斑区采集的多张光学相干断层扫描仪OCT图像;对所述多张OCT图像进行特征提取,得到所述多张OCT图像对应的多个图像特征,其中,所述多张OCT图像中的每张OCT图像对应的图像特征包括所述每张OCT图像中的 多尺度特征;将所述多张OCT图像对应的多个图像特征进行空间信息融合,得到所述多张OCT图像对应的特征;根据所述多张OCT图像对应的特征,确定抗VEGF疗效预测结果。
- 根据权利要求15所述的计算机可读存储介质,其中,所述获取针对黄斑区采集的多张光学相干断层扫描仪OCT图像时,具体实现:获取通过光学相干断层扫描仪OCT对眼部进行扫描得到的多张初始OCT图像;提取所述多张初始OCT图像中针对黄斑区采集的区域,以得到针对黄斑区采集的所述多张OCT图像。
- 根据权利要求15所述的计算机可读存储介质,其中,所述计算机程序被处理器执行时还用于实现:对所述多张OCT图像进行图像校正处理,得到校正后的所述多张OCT图像,其中,所述图像校正处理包括图像倾斜校正和/或图像亮度校正;对校正后的所述多张OCT图像进行对比度增强处理,得到对比度增强后的所述多张OCT图像。
- 根据权利要求15至17任一项所述的计算机可读存储介质,其中,所述对所述多张OCT图像进行特征提取,得到所述多张OCT图像对应的多个图像特征时,具体实现:将所述多张OCT图像输入特征提取网络,得到所述多张OCT图像对应的多个图像特征,其中,所述特征提取网络为残差网络结构,所述特征提取网络中的卷积层用于提取所述每张OCT图像中的多尺度特征。
- 根据权利要求15至17任一项所述的计算机可读存储介质,其中,所述将所述多张OCT图像对应的多个图像特征进行空间信息融合,得到所述多张OCT图像对应的特征时,具体实现:获取所述多张OCT图像中的每张OCT图像对应的时间;根据所述每张OCT图像对应的时间,确定所述多张OCT图像对应的多个图像特征的时间序列;将所述多张OCT图像对应的多个图像特征的时间序列输入长短期记忆人工神经网络,得到所述多张OCT图像对应的特征。
- 根据权利要求15至17任一项所述的计算机可读存储介质,其中,所述根据所述多张OCT图像对应的特征,确定抗VEGF疗效预测结果时,具体实现:根据所述多张OCT图像对应的特征,对抗VEGF疗效进行二分类,得到所述抗VEGF疗效预测结果,其中,所述抗VEGF疗效预测结果包括视力提升或视力恶化。
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011061099.8A CN111920375B (zh) | 2020-09-30 | 2020-09-30 | 一种抗血管内皮生长因子vegf疗效预测装置及方法 |
CN202011061099.8 | 2020-09-30 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2021139446A1 true WO2021139446A1 (zh) | 2021-07-15 |
Family
ID=73334769
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/CN2020/132473 WO2021139446A1 (zh) | 2020-09-30 | 2020-11-27 | 一种抗血管内皮生长因子vegf疗效预测装置及方法 |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN111920375B (zh) |
WO (1) | WO2021139446A1 (zh) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111920375B (zh) * | 2020-09-30 | 2021-01-15 | 平安科技(深圳)有限公司 | 一种抗血管内皮生长因子vegf疗效预测装置及方法 |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106455972A (zh) * | 2014-04-28 | 2017-02-22 | 西北大学 | 功能性光学相干断层成像的装置、方法和*** |
CN110309862A (zh) * | 2019-06-11 | 2019-10-08 | 广东省人民医院(广东省医学科学院) | 基于集成机器学习的dme预后信息预测***及其应用方法 |
CN110472530A (zh) * | 2019-07-29 | 2019-11-19 | 中山大学 | 基于小波变换和迁移学习的视网膜oct图像分类方法 |
US20200027209A1 (en) * | 2018-07-18 | 2020-01-23 | Case Western Reserve University | Predicting response to anti-vascular endothelial growth factor therapy with computer-extracted morphology and spatial arrangement features of leakage patterns on baseline fluorescein angiography in diabetic macular edema |
CN111611877A (zh) * | 2020-04-30 | 2020-09-01 | 杭州电子科技大学 | 基于多时空信息融合的抗年龄干扰的人脸识别方法 |
CN111920375A (zh) * | 2020-09-30 | 2020-11-13 | 平安科技(深圳)有限公司 | 一种抗血管内皮生长因子vegf疗效预测装置及方法 |
Family Cites Families (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP2442715A4 (en) * | 2009-06-18 | 2018-01-03 | Richard M. Awdeh | System and method for molecular in vivo imaging and theranostics |
CN108229580B (zh) * | 2018-01-26 | 2020-12-11 | 浙江大学 | 一种基于注意力机制及特征融合的眼底图中糖网特征分级装置 |
CN109376767B (zh) * | 2018-09-20 | 2021-07-13 | 中国科学技术大学 | 基于深度学习的视网膜oct图像分类方法 |
CN109727259A (zh) * | 2019-01-07 | 2019-05-07 | 哈尔滨理工大学 | 一种基于残差u-net网络的视网膜图像分割算法 |
CN109858429B (zh) * | 2019-01-28 | 2021-01-19 | 北京航空航天大学 | 一种基于卷积神经网络的眼底图像病变程度识别与可视化*** |
CN110147715A (zh) * | 2019-04-01 | 2019-08-20 | 江西比格威医疗科技有限公司 | 一种视网膜OCT图像Bruch膜开角自动检测方法 |
CN111184508B (zh) * | 2020-01-19 | 2021-05-18 | 武汉大学 | 一种基于联合神经网络的心电信号检测装置及分析方法 |
CN111329445B (zh) * | 2020-02-20 | 2023-09-15 | 广东工业大学 | 基于组卷积残差网络和长短期记忆网络的房颤识别方法 |
-
2020
- 2020-09-30 CN CN202011061099.8A patent/CN111920375B/zh active Active
- 2020-11-27 WO PCT/CN2020/132473 patent/WO2021139446A1/zh active Application Filing
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106455972A (zh) * | 2014-04-28 | 2017-02-22 | 西北大学 | 功能性光学相干断层成像的装置、方法和*** |
US20200027209A1 (en) * | 2018-07-18 | 2020-01-23 | Case Western Reserve University | Predicting response to anti-vascular endothelial growth factor therapy with computer-extracted morphology and spatial arrangement features of leakage patterns on baseline fluorescein angiography in diabetic macular edema |
CN110309862A (zh) * | 2019-06-11 | 2019-10-08 | 广东省人民医院(广东省医学科学院) | 基于集成机器学习的dme预后信息预测***及其应用方法 |
CN110472530A (zh) * | 2019-07-29 | 2019-11-19 | 中山大学 | 基于小波变换和迁移学习的视网膜oct图像分类方法 |
CN111611877A (zh) * | 2020-04-30 | 2020-09-01 | 杭州电子科技大学 | 基于多时空信息融合的抗年龄干扰的人脸识别方法 |
CN111920375A (zh) * | 2020-09-30 | 2020-11-13 | 平安科技(深圳)有限公司 | 一种抗血管内皮生长因子vegf疗效预测装置及方法 |
Non-Patent Citations (1)
Title |
---|
SU, QIANG ET AL.: "Clinical Efficacy of Anti-VEGF on Exudative Senile Macular Degeneration based on ANGIO-OCT", CHINESE REMEDIES & CLINICS, vol. 20, no. 16, 31 August 2020 (2020-08-31), pages 2683 - 2686, XP055827602, ISSN: 1671-2560 * |
Also Published As
Publication number | Publication date |
---|---|
CN111920375B (zh) | 2021-01-15 |
CN111920375A (zh) | 2020-11-13 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Bhaskaranand et al. | The value of automated diabetic retinopathy screening with the EyeArt system: a study of more than 100,000 consecutive encounters from people with diabetes | |
Varadarajan et al. | Deep learning for predicting refractive error from retinal fundus images | |
US20220076420A1 (en) | Retinopathy recognition system | |
WO2020199593A1 (zh) | 图像分割模型训练方法、图像分割方法、装置、设备及介质 | |
US10722180B2 (en) | Deep learning-based diagnosis and referral of ophthalmic diseases and disorders | |
CN110400289B (zh) | 眼底图像识别方法、装置、设备和存储介质 | |
US20220165418A1 (en) | Image-based detection of ophthalmic and systemic diseases | |
Hassan et al. | Joint segmentation and quantification of chorioretinal biomarkers in optical coherence tomography scans: A deep learning approach | |
US20210391056A1 (en) | Health big data service method and system based on remote fundus screening | |
Wainwright et al. | Effect of image quality, color, and format on the measurement of retinal vascular fractal dimension | |
CN110415245A (zh) | 眼部数据确定方法、模型训练方法及设备 | |
Tsikata et al. | Automated brightness and contrast adjustment of color fundus photographs for the grading of age-related macular degeneration | |
Dong et al. | Deep learning-based estimation of axial length and subfoveal choroidal thickness from color fundus photographs | |
WO2021139446A1 (zh) | 一种抗血管内皮生长因子vegf疗效预测装置及方法 | |
Giancardo | Automated fundus images analysis techniques to screen retinal diseases in diabetic patients | |
CN116763250A (zh) | 眼前节图像的眼压信息提取方法、终端和存储介质 | |
WO2023103609A1 (zh) | 用于眼前节octa的眼动追踪方法、装置、设备和存储介质 | |
Thanh et al. | A real-time classification of glaucoma from retinal fundus images using AI technology | |
Burke et al. | An Open-Source Deep Learning Algorithm for Efficient and Fully Automatic Analysis of the Choroid in Optical Coherence Tomography | |
Yin et al. | A cloud-based system for automatic glaucoma screening | |
Aswathy | Diabetic Retinopathy Detection Using Semantic Segmentation And Optic Disc Localization | |
Shon et al. | Development of a β-variational autoencoder for disentangled latent space representation of anterior segment optical coherence tomography images | |
Almazroa et al. | The appropriateness of digital diabetic retinopathy screening images for a computer-aided glaucoma screening system | |
Sharma et al. | Analysis of eye disease classification by fundus images using different machine/deep/transfer learning techniques | |
Hakeem et al. | Inception V3 and CNN Approach to Classify Diabetic Retinopathy Disease |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 20911442 Country of ref document: EP Kind code of ref document: A1 |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
32PN | Ep: public notification in the ep bulletin as address of the adressee cannot be established |
Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 25/05/2023) |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 20911442 Country of ref document: EP Kind code of ref document: A1 |