Eyeball A ultrasonic biological measurement method and system under assistance of B ultrasonic image

Abstract

The invention discloses an eyeball A-ultrasonic biological measurement method and system under the assistance of a B-ultrasonic image, and relates to the field of ophthalmology department image processing.The method comprises the steps that standard scanning is conducted on an eyeball through the B-ultrasonic scanning technology, and a standard eyeball section image is obtained; constructing an automatic judgment model to mark the standard eyeball section image to obtain three surface distances; performing distance parameter measurement on the standard eyeball section image based on the three surface distances to obtain a first measurement result; performing standard scanning on the eyeball by adopting an A-ultrasonic scanning technology to obtain a standard A-ultrasonic scanning waveform image; marking the standard A ultrasonic scanning waveform image based on the first measurement result to obtain four surface waves; based on the four surface waves, a second measurement result is obtained, and A-ultrasonic waveform biological parameter measurement is completed. According to the method, the existing A-ultrasonic technology is easy to inaccurately identify eyeball waveforms in part of cases, and the accuracy and efficiency of A-ultrasonic waveform identification can be effectively improved.

Description

Technical Field

[0001] The present invention relates to the field of ophthalmic image processing, and in particular to an eyeball A-ultrasound biological measurement method and system assisted by B-ultrasound images. Background Art

[0002] Eyeball biometry refers to the use of various inspection methods to measure various parameters of the eyeball structure, providing a reference for the diagnosis and treatment of eye diseases. A-ultrasound, also known as A-type ultrasound, detects the echo of sound waves based on the relationship between the time and amplitude of sound waves. Ophthalmic ultrasound imaging uses A-ultrasound to measure ophthalmic biological parameters. The principle is that when the probe is placed in front of the eyeball, the sound wave propagates forward. According to the propagation characteristics of the sound wave, each time it encounters an interface with density difference, it will reflect once, forming a wave crest. In this way, the reflected echo will be arranged in sequence according to the time of return. The height of the wave crest indicates the intensity of the echo. The higher the wave crest, the stronger the echo. A-ultrasound detects the position of the wave crest and calculates the corresponding eyeball biological parameters (such as corneal thickness, anterior chamber depth, lens thickness, vitreous length and axial length of the eyeball) through sound velocity matching. The accurate measurement of these parameters is crucial for calculating the degree of the artificial lens to be implanted in the eye (an error of 1.0mm in axial measurement will produce an average refractive error of up to 3.0D).

[0003] Although the existing A-ultrasound biological measurement technology can determine the interfaces of different structures in the eyeball for most people, thereby accurately measuring various biological parameters, the A-ultrasound waveform judgment of patients with special types of cataracts, retinal lesions and vitreous opacities often has many mixed waves, which are difficult to accurately identify. The operator is required to have rich experience to adjust the waveform judgment, which affects clinical work efficiency and is also prone to errors in the measurement of eyeball biological parameters.

[0004] Existing B-ultrasound imaging technology can easily determine the interfaces of different structures in the eye (the posterior surface of the cornea, the anterior surface of the lens, the posterior surface of the lens and the retinal surface, etc.). However, due to the insufficient resolution of B-ultrasound imaging, it is difficult to accurately define and measure the distances between the interfaces of various structures in the eye, which affects the precise measurement of ocular biological parameters such as anterior chamber depth, lens thickness, vitreous length and axial length. Summary of the invention

[0005] To solve the above technical problems, the present invention provides an accurate biological measurement method and system for automatic recognition of eye A-ultrasound waveforms combined with B-ultrasound image measurement technology, which is conducive to efficient and accurate measurement of eye A-ultrasound biological parameters.

[0006] To achieve the above object, the present invention provides an eyeball A-ultrasound biological measurement method assisted by B-ultrasound images, the steps comprising:

[0007] Use B-ultrasound scanning technology to perform standard scans on the eyeball and obtain standard cross-sectional images of the eyeball;

[0008] An automatic judgment model is constructed to mark the standard eyeball section image to obtain: the anterior surface of the cornea, recorded as S1; the anterior surface of the lens, recorded as S2; the posterior surface of the lens, recorded as S3 and the retinal surface, recorded as S4, and three surface distances between these four surfaces. The automatic judgment model adopts an image segmentation and positioning algorithm based on the YOLO model to realize automatic position judgment, and the steps include:

[0009] 1. Image Annotation

[0010] Use Labelme annotation software to perform instance segmentation and annotation on the image, annotate the border positions of four categories: the posterior surface of the cornea, the anterior surface of the lens, the posterior surface of the lens, and the retinal surface, and then generate a label file containing the object category and the bounding box;

[0011] 2. Model training

[0012] Select the YOLOv11 model for training, adjust the input resolution of the image to 640*640, set the batch size to 16, use the Adam optimizer, set the initial learning rate to 0.001, set the learning rate decay factor to 0.01, set the random horizontal flip to 0.5, set the random scale to 0.5, and enable the Mosaic data enhancement strategy;

[0013] 3. Model Reasoning

[0014] The confidence threshold of the reasoning process is set to 0.25, and the model outputs frame coordinates of four categories: the anterior surface of the cornea, the anterior surface of the lens, the posterior surface of the lens, and the retinal surface. The center point coordinates of each frame are calculated according to these frame coordinates, and then the three surface distances are calculated by the value of the center point coordinate X-axis; based on the three surface distances, the distance parameters of the standard eye section image are measured to obtain a first measurement result; the distance parameters include: the distance from the anterior surface of the cornea to the anterior surface of the lens, the distance from the anterior surface of the cornea to the posterior surface of the lens, and the distance from the anterior surface of the cornea to the retinal surface;

[0015] A-ultrasound scanning technology is used to perform a standard scan of the eyeball to obtain a standard A-ultrasound scanning waveform image;

[0016] Based on the first measurement result, the standard A-scan waveform image is marked to obtain four surface waves, including: anterior corneal surface wave, anterior lens surface wave, posterior lens surface wave and anterior retinal surface wave;

[0017] Based on the four surface waves, a second measurement result is obtained, and the A-ultrasound waveform biological parameter measurement is completed.

[0018] Preferably, the distance from the anterior surface of the cornea to the anterior surface of the lens is recorded as D1; ​​the distance from the anterior surface of the cornea to the posterior surface of the lens is recorded as D2; and the distance from the anterior surface of the cornea to the surface of the retina is recorded as D3.

[0019] Preferably, in the standard A-scan waveform image, from left to right, the first single high wave is marked as the surface wave of S1, and D1 is used as a reference to select the surface wave with the smallest adjacent distance to D1 and marked as S2; D2 is used as a reference to select the surface wave with the smallest adjacent distance to D2 and marked as S3; D3 is used as a reference to select the surface wave with the smallest adjacent distance to D3 and marked as S4.

[0020] Preferably, based on the surface waves of the four surfaces S1, S2, S3 and S4, the distance from S1 to S2 is measured as the anterior chamber depth, the distance from S2 to S3 is measured as the lens thickness, the distance between S3 and S4 is measured as the vitreous cavity depth, and the distance between S1 and S4 is measured as the axial length.

[0021] The present invention also provides an eyeball A-ultrasound biological measurement system assisted by B-ultrasound images, the system is used to implement the above method, including: a first acquisition module, a first marking module, a second acquisition module, a second marking module and a measurement module;

[0022] The first acquisition module is used to perform a standard scan of the eyeball using B-ultrasound scanning technology to obtain a standard eyeball section image;

[0023] The first marking module is used to construct an automatic judgment model to mark the standard eyeball section image to obtain: the front surface of the cornea, recorded as S1; the front surface of the lens, recorded as S2; the back surface of the lens, recorded as S3 and the retinal surface, recorded as S4, and the three surface distances between these four surfaces. The automatic judgment model adopts an image segmentation and positioning algorithm based on the YOLO model to realize automatic position judgment, and the steps include:

[0024] 1. Image Annotation

[0025] Use Labelme annotation software to perform instance segmentation and annotation on the image, annotate the border positions of four categories: the anterior surface of the cornea, the anterior surface of the lens, the posterior surface of the lens, and the retinal surface, and then generate a label file containing the object category and the bounding box;

[0026] 2. Model training

[0027] Select the YOLOv11 model for training, adjust the input resolution of the image to 640*640, set the batch size to 16, use the Adam optimizer, set the initial learning rate to 0.001, set the learning rate decay factor to 0.01, set the random horizontal flip to 0.5, set the random scale to 0.5, and enable the Mosaic data enhancement strategy;

[0028] 3. Model Reasoning

[0029] The confidence threshold of the reasoning process is set to 0.25, and the model outputs frame coordinates of four categories: the anterior surface of the cornea, the anterior surface of the lens, the posterior surface of the lens, and the retinal surface. The center point coordinates of each frame are calculated according to these frame coordinates, and then the three surface distances are calculated by the value of the center point coordinate X-axis; based on the three surface distances, the distance parameters of the standard eye section image are measured to obtain a first measurement result; the distance parameters include: the distance from the anterior surface of the cornea to the anterior surface of the lens, the distance from the anterior surface of the cornea to the posterior surface of the lens, and the distance from the anterior surface of the cornea to the retinal surface;

[0030] The second acquisition module is used to perform a standard scan of the eyeball using A-scan technology to obtain a standard A-scan waveform image;

[0031] The second marking module is used to mark the standard A-scan waveform based on the first measurement result to obtain four surface waves, including: anterior corneal surface wave, anterior lens surface wave, posterior lens surface wave and anterior retinal surface wave;

[0032] The measurement module is used to complete the measurement of A-ultrasound waveform biological parameters based on the four surface waves.

[0033] Compared with the prior art, the present invention has the following beneficial effects:

[0034] In view of the fact that existing clinical A-ultrasound technology is prone to inaccurate eye waveform recognition in some cases, the present invention adds B-ultrasound image measurement information to assist in the recognition of A-ultrasound waveforms, which can effectively improve the accuracy and efficiency of A-ultrasound waveform recognition, thereby improving the accuracy of eye biological parameter measurement. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the technical solution of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0036] Figure 1 A flowchart for obtaining a first measurement result according to an embodiment of the present invention;

[0037] Figure 2 A flow chart of obtaining a second measurement result according to an embodiment of the present invention;

[0038] Figure 3 A standard A-scan waveform image of an embodiment of the present invention;

[0039] Figure 4 A schematic diagram of marking a scanned section image according to an embodiment of the present invention;

[0040] Figure 5 This is an A-ultrasound scanning waveform image based on B-ultrasound assistance according to an embodiment of the present invention. DETAILED DESCRIPTION

[0041] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0042] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0043] Embodiment 1

[0044] This embodiment provides a method for eyeball A-ultrasound biological measurement assisted by B-ultrasound images, the steps comprising:

[0045] S1. Use B-ultrasound scanning technology to perform a standard scan of the eyeball to obtain a standard eyeball cross-sectional image.

[0046] S2. Build an automatic judgment model to mark standard eye section images.

[0047] By marking the standard eyeball section image, we can obtain the following: the anterior surface of the cornea, denoted as S1; the anterior surface of the lens, denoted as S2; the posterior surface of the lens, denoted as S3 and the retinal surface, denoted as S4. The three surface distances between these four surfaces are obtained. The automatic judgment model adopts an image segmentation and positioning algorithm based on the YOLO model to realize automatic position judgment. The steps include:

[0048] 1. Image Annotation

[0049] The Labelme annotation software was used to perform instance segmentation and annotation on the image, and the border positions of the four categories of the anterior surface of the cornea, the anterior surface of the lens, the posterior surface of the lens, and the retinal surface were annotated respectively, and then a label file containing the object category and the bounding box was generated. Since the B-ultrasound resolution cannot accurately measure the cornea, this embodiment uses the average corneal thickness of 0.5mm as a constant, and the posterior surface of the cornea is shifted 0.5mm to the left to mark it as the anterior surface of the cornea. A total of 1,000 B-ultrasound images were annotated.

[0050] 2. Model training

[0051] The YOLOv11 (COCO pre-trained) model was selected for training. The input resolution of the image was adjusted to 640*640, the batch size was set to 16, the Adam optimizer, the initial learning rate was set to 0.001, the learning rate decay factor was set to 0.01, the random horizontal flip (fliplr) was set to 0.5, the random scale (scale) was set to 0.5, and the Mosaic data enhancement strategy was enabled. The operating system was Ubuntu 20.04, the GPU was 1 NVIDIA GeForce RTX 3090 (video memory 24G), the programming framework was PyTorch, and a total of 100 rounds of training were performed to achieve model training for the segmentation and positioning of the anterior surface of the cornea, the anterior surface of the lens, the posterior surface of the lens, and the retinal surface on B-ultrasound images.

[0052] 3. Model Reasoning

[0053] The confidence threshold of the inference process is set to 0.25. The model outputs the coordinates of four categories of borders: the anterior surface of the cornea, the anterior surface of the lens, the posterior surface of the lens, and the retinal surface. The center point coordinates of each border are calculated based on these border coordinates, and then the three surface distances between the four surfaces are calculated through the value of the X-axis of the center point coordinates. Finally, based on the three surface distances, the distance parameters of the standard eye section image are measured to obtain the first measurement result, including: the distance from the anterior surface of the cornea to the anterior surface of the lens, denoted as D1, the distance from the anterior surface of the cornea to the posterior surface of the lens, denoted as D2, and the distance from the anterior surface of the cornea to the retinal surface, denoted as D3. Among them, D1 is the anterior chamber depth, D3 is the axial length of the eye, D2-D1 is the lens thickness, and D3-D2 is the vitreous cavity depth. The overall process is as follows: Figure 1 shown.

[0054] S3. Use A-ultrasound scanning technology to perform a standard scan on the eyeball and obtain a standard A-ultrasound scanning waveform image.

[0055] S4. Based on the first measurement result, mark the standard A-scan waveform image to obtain four surface waves.

[0056] The four surface waves include: the anterior corneal surface wave, the anterior lens surface wave, the posterior lens surface wave and the anterior retinal surface wave. The method for identifying the A-ultrasound peaks of different intraocular structural interfaces includes: in the standard A-ultrasound scanning waveform image, from left to right, the first single high wave is marked as the surface wave S1, and D1 is used as a reference to select the surface wave with the smallest adjacent distance to D1 and marked as S2; using D2 as a reference, select the surface wave with the smallest adjacent distance to D2 and marked as S3; using D3 as a reference, select the surface wave with the smallest adjacent distance to D3 and marked as S4.

[0057] S5. Based on the four surface waves, a second measurement result is obtained, and the A-ultrasound waveform biological parameter measurement is completed.

[0058] Based on the surface waves of the four surfaces S1, S2, S3 and S4, the distance from S1 to S2 is measured as the anterior chamber depth, the distance from S2 to S3 is measured as the lens thickness, the distance between S3 and S4 is measured as the vitreous cavity depth, and the distance between S1 and S4 is measured as the axial length. The second measurement result is obtained to complete the measurement of A-ultrasound waveform biological parameters. The process is as follows: Figure 2 shown.

[0059] Embodiment 2

[0060] This embodiment is described by taking retinal detachment as an example.

[0061] Use existing technology to perform A-ultrasound scanning on the eyeball to be measured, and obtain a standard A-ultrasound scanning waveform (such as Figure 3 As shown in the figure, the single high waves from left to right are the anterior surface of the cornea S1, the anterior surface of the lens S2, the posterior surface of the lens S3, and the retinal surface S4. The measured axial length (the distance from the S1 interface to the S4 interface) is 23.45 mm.

[0062] Perform B-ultrasound scanning on the eyeball to be measured to obtain the required standard scanning section. The scanned section image is judged and marked for the four interfaces of the anterior surface of the cornea, the anterior surface of the lens, the posterior surface of the lens, and the retinal surface, and the anterior chamber depth, lens thickness, vitreous cavity depth, and axial length are measured respectively (such as Figure 4 As shown), the axial length of the eye is 23.97mm.

[0063] Re-label the A-ultrasound waveform from left to right, and label the first single high wave as the corneal anterior surface wave (S1 B ), taking the anterior chamber depth measured by B-ultrasound as a reference, the single high wave with the minimum distance adjacent to the anterior chamber depth was selected as the anterior surface wave of the lens (S2 B ); Taking the sum of the anterior chamber depth and the lens thickness measured by B-ultrasound as a reference, the single high wave with the minimum distance adjacent to the measured value is selected as the posterior surface wave of the lens (S3B ); Taking the axial length measured by B-ultrasound as a reference, the single high wave with the smallest adjacent distance is selected as the anterior retinal surface wave (S4 B ),See Figure 5 .

[0064] Measure the above S1 B To S2 B The distance is taken as the anterior chamber depth, S2 B To S3 B The distance is taken as the lens thickness, S3 B To S4 B The distance between the vitreous cavity and the vitreous cavity depth, S1 B To S4 B The distance between the two is taken as the axial length of the eye. Figure 5 As shown, the corneal anterior surface wave, lens anterior surface wave, and lens posterior surface wave identified by the present technical solution are different from those of the above-mentioned prior art ( Figure 3 ) is consistent with the retinal surface wave S4 B This is inconsistent with the S4 identified by the prior art. From the B-ultrasound scan section of the eyeball, it can be judged that the current embodiment has retinal detachment, while the prior art identifies the raised retina as the surface of the eyeball wall, which cannot accurately reflect the true axial length of the eye. In this embodiment, the axial length measured by the current technical solution is 23.95mm, compared with the axial length of 23.45 measured by the prior art. The difference of 0.50mm will result in a refractive error of about 1.5D, which will have a significant impact on the patient's intraocular lens power prediction calculation.

[0065] The embodiments described above are only descriptions of the preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the protection scope determined by the claims of the present invention.

Claims

1. A method for measuring eyeball biologically with A-ultrasound assisted by B-ultrasound images, characterized in that the steps include: Use B-ultrasound scanning technology to perform standard scans on the eyeball and obtain standard cross-sectional images of the eyeball; An automatic judgment model is constructed to mark the standard eyeball section image to obtain: the anterior surface of the cornea, denoted as S1; The front surface of the lens is denoted as S2; the rear surface of the lens is denoted as S3; and the retinal surface is denoted as S4. The three surface distances between these four surfaces, the automatic judgment model adopts an image segmentation and positioning algorithm based on the YOLO model to realize automatic position judgment, and the steps include: Use Labelme annotation software to perform instance segmentation and annotation on the image, annotate the border positions of four categories: the anterior surface of the cornea, the anterior surface of the lens, the posterior surface of the lens, and the retinal surface, and then generate a label file containing the object category and the bounding box; Select the YOLOv11 model for training, adjust the input resolution of the image to 640*640, set the batch size to 16, use the Adam optimizer, set the initial learning rate to 0.001, set the learning rate decay factor to 0.01, set the random horizontal flip to 0.5, set the random scale to 0.5, and enable the Mosaic data enhancement strategy; The confidence threshold of the reasoning process is set to 0.25, and the model outputs frame coordinates of four categories: the anterior surface of the cornea, the anterior surface of the lens, the posterior surface of the lens, and the retinal surface. The center point coordinates of each frame are calculated according to these frame coordinates, and then the three surface distances are calculated by the value of the center point coordinate X-axis; based on the three surface distances, the distance parameters of the standard eye section image are measured to obtain a first measurement result; the distance parameters include: the distance from the anterior surface of the cornea to the anterior surface of the lens, the distance from the anterior surface of the cornea to the posterior surface of the lens, and the distance from the anterior surface of the cornea to the retinal surface; A-ultrasound scanning technology is used to perform a standard scan of the eyeball to obtain a standard A-ultrasound scanning waveform image; Based on the first measurement result, the standard A-scan waveform image is marked to obtain four surface waves, including: anterior corneal surface wave, anterior lens surface wave, posterior lens surface wave and anterior retinal surface wave; Based on the four surface waves, a second measurement result is obtained, and the A-ultrasound waveform biological parameter measurement is completed.

2. The eyeball A-ultrasound biological measurement method assisted by B-ultrasound images according to claim 1, characterized in that: The distance from the anterior surface of the cornea to the anterior surface of the lens is recorded as D1; ​​the distance from the anterior surface of the cornea to the posterior surface of the lens is recorded as D2; and the distance from the anterior surface of the cornea to the retinal surface is recorded as D3.

3. The eyeball A-ultrasound biological measurement method assisted by B-ultrasound images according to claim 2, characterized in that: In the standard A-scan waveform image, from left to right, the first single high wave is marked as the surface wave of S1, and D1 is used as a reference to select the surface wave with the smallest adjacent distance to D1 and marked as S2; D2 is used as a reference to select the surface wave with the smallest adjacent distance to D2 and marked as S3; D3 is used as a reference to select the surface wave with the smallest adjacent distance to D3 and marked as S4.

4. The eyeball A-ultrasound biological measurement method assisted by B-ultrasound images according to claim 3, characterized in that: Based on the surface waves of the four surfaces S1, S2, S3 and S4, the distance from S1 to S2 was measured as the anterior chamber depth, the distance from S2 to S3 was measured as the lens thickness, the distance from S3 to S4 was measured as the vitreous cavity depth, and the distance from S1 to S4 was measured as the axial length.

5. A B-ultrasound image-assisted eyeball A-ultrasound biological measurement system, the system is used to implement the method according to any one of claims 1 to 4, characterized in that: include: A first acquisition module, a first marking module, a second acquisition module, a second marking module and a measurement module; The first acquisition module is used to perform a standard scan of the eyeball using B-ultrasound scanning technology to obtain a standard eyeball section image; The first marking module is used to construct an automatic judgment model to mark the standard eyeball section image to obtain: the anterior surface of the cornea, denoted as S1; The front surface of the lens is denoted as S2; the rear surface of the lens is denoted as S3; and the retinal surface is denoted as S4. The three surface distances between these four surfaces, the automatic judgment model adopts an image segmentation and positioning algorithm based on the YOLO model to realize automatic position judgment, and the steps include: Use Labelme annotation software to perform instance segmentation and annotation on the image, annotate the border positions of four categories: the anterior surface of the cornea, the anterior surface of the lens, the posterior surface of the lens, and the retinal surface, and then generate a label file containing the object category and the bounding box; Select the YOLOv11 model for training, adjust the input resolution of the image to 640*640, set the batch size to 16, use the Adam optimizer, set the initial learning rate to 0.001, set the learning rate decay factor to 0.01, set the random horizontal flip to 0.5, set the random scale to 0.5, and enable the Mosaic data enhancement strategy; The confidence threshold of the reasoning process is set to 0.25, and the model outputs frame coordinates of four categories: the anterior surface of the cornea, the anterior surface of the lens, the posterior surface of the lens, and the retinal surface. The center point coordinates of each frame are calculated according to these frame coordinates, and then the three surface distances are calculated by the value of the center point coordinate X-axis; based on the three surface distances, the distance parameters of the standard eye section image are measured to obtain a first measurement result; the distance parameters include: the distance from the anterior surface of the cornea to the anterior surface of the lens, the distance from the anterior surface of the cornea to the posterior surface of the lens, and the distance from the anterior surface of the cornea to the retinal surface; The second acquisition module is used to perform a standard scan of the eyeball using A-scan technology to obtain a standard A-scan waveform image; The second marking module is used to mark the standard A-scan waveform based on the first measurement result to obtain four surface waves, including: anterior corneal surface wave, anterior lens surface wave, posterior lens surface wave and anterior retinal surface wave; The measurement module is used to complete the measurement of A-ultrasound waveform biological parameters based on the four surface waves.

Images