Wavefront sensing technology-based cataract vision real-time enhancement system and method

Abstract

The invention discloses a cataract vision real-time enhancement system and method based on a wavefront sensing technology, and the system comprises a camera module, a beam expander, a polarizing film, a beam splitter, a light modulation part, a scattering medium and a focusing lens, and the beam expander, the polarizing film, the beam splitter, the scattering medium and the focusing lens are sequentially connected. The light modulation part carries out light wave front phase modulation on the light beams separated by the beam splitter, and the camera module collects a clear image which is not disturbed by the scattering medium and a blurred image which is modulated by the light modulation part, penetrates through the scattering medium and is focused by the focusing lens. The target area image quality is optimized based on the clear image and the blurred image to optimize the phase of the light modulation component. According to the invention, the limitation of the existing cataract treatment method can be overcome, the visual sense of the cataract patient is enhanced in real time, and the life quality of the patient is improved.

Description

Technical Field

[0001] The invention relates to a real-time cataract vision enhancement system and method based on wavefront sensing technology, belonging to the technical field of light wavefront modulation and vision enhancement. Background Art

[0002] In modern medicine, cataract is one of the main causes of visual impairment and blindness. Traditional cataract treatments mainly include surgery and drug therapy, but these methods have certain risks and limitations. With the development of optical technology, visual enhancement technology based on optical principles has gradually become a research hotspot. However, how to achieve real-time enhancement of cataract patients' vision under non-invasive conditions is still a problem that needs to be solved.

[0003] Optical wavefront modulation technology has made significant progress in the field of vision enhancement in recent years. Cataract scattering crystals cause the wavefront of light entering the eye to be distorted. Adjusting the wavefront through a spatial light modulator can effectively correct vision defects and improve visual quality. For cataract patients, traditional vision correction methods have limited effects due to the opacity of the intraocular lens. Vision enhancement systems based on wavefront sensing technology can achieve real-time enhancement of cataract patients' vision under non-invasive conditions through precise wavefront modulation.

[0004] As one of the core technologies of artificial intelligence, deep neural network (DNN) has shown strong performance in image processing and optimization. Convolutional neural network (CNN) is particularly good at processing image data, and can effectively extract the spatial features of images. It is widely used in tasks such as image classification, image segmentation and image enhancement. In recent years, deep learning models such as U-Net have been increasingly used in medical image processing, providing efficient image restoration and enhancement solutions. Summary of the invention

[0005] The purpose of the present invention is to address the deficiencies of the above-mentioned prior art and to provide a real-time cataract vision enhancement system based on wavefront sensing technology, which can overcome the limitations of existing cataract treatment methods, achieve real-time enhancement of the vision of cataract patients, and improve the quality of life of patients. Specifically, through the combination of optical paths, software and hardware modules and optimization methods, wavefront sensing technology and deep neural network algorithms are used to achieve non-invasive visual enhancement.

[0006] Another object of the present invention is to provide a real-time enhancement method for cataract vision.

[0007] The purpose of the present invention can be achieved by adopting the following technical solutions:

[0008] A real-time cataract vision enhancement system based on wavefront sensing technology comprises a camera module, a beam expander, a polarizer, a beam splitter, a light modulation component, a scattering medium and a focusing lens, wherein the beam expander, the polarizer, the beam splitter, the scattering medium and the focusing lens are connected in sequence, the light modulation component performs phase modulation of the light wavefront on the light beam separated by the beam splitter, the camera module collects a clear image that is not disturbed by the scattering medium and a blurred image that is focused by the focusing lens after the light beam modulated by the light modulation component passes through the scattering medium, and the phase of the light modulation component is optimized based on the image quality of the target area of ​​the clear image and the blurred image.

[0009] Furthermore, the camera module includes a first camera and a second camera, the first camera captures a clear image without being disturbed by a scattering medium, and the second camera captures a blurred image focused by a focusing lens after a light beam modulated by a light modulation component passes through a scattering medium, for simulating the retina of a human eye.

[0010] Another object of the present invention can be achieved by adopting the following technical solutions:

[0011] A method for real-time enhancement of cataract vision is implemented based on the above-mentioned real-time enhancement system for cataract vision, and the method comprises:

[0012] The first camera of the camera module captures a clear image that is not disturbed by a scattering medium;

[0013] The light beam generated by the laser is expanded by a beam expander, the polarizer controls the polarization state of the expanded light beam, the beam splitter separates the light beam, the light modulation component modulates the phase of the light wavefront of the separated light beam, and the modulated light beam returns to the beam splitter. The modulated light beam of the beam splitter is focused by a focusing lens through a scattering medium;

[0014] The second camera of the camera module captures a blurred image of the modulated light beam passing through the scattering medium and focused by the focusing lens;

[0015] Optimize the phase of the optical modulation component based on the image quality of the target area of ​​the clear image and the blurred image to obtain the global optimal modulation phase data;

[0016] The deep learning neural network model is trained based on the clear image, blurred image and global optimal modulation phase data;

[0017] Based on the trained deep learning neural network model, the real-time acquired image data is input, and the phase modulation parameters are predicted and optimized to adjust the light modulation components in real time, so that the real-time enhancement effect of cataract patients on the target area of ​​the visual image is optimized.

[0018] Furthermore, the optimization of the phase of the light modulation component based on the image quality of the target area of ​​the clear image and the blurred image specifically includes:

[0019] Based on the image quality of the target area of ​​the clear image and the blurred image, an iterative optimization algorithm is used to optimize the phase of the light modulation component.

[0020] Furthermore, the image quality of the target area based on the clear image and the blurred image is optimized by using an iterative optimization algorithm, which specifically includes:

[0021] The image quality of the target area of ​​the clear image and the blurred image is compared as a basis for judging whether the iterative optimization algorithm has converged, so that the phase of the optical modulation component after iteration gradually converges to the optimal modulation phase.

[0022] Furthermore, the iterative optimization algorithm adopts a continuous sequence algorithm to search for a global optimal solution by gradually adjusting the phase to obtain an optimal modulation phase.

[0023] Furthermore, the iterative optimization algorithm uses a genetic algorithm to search for a global optimal solution by simulating natural selection and genetic variation to obtain an optimal modulation phase.

[0024] Furthermore, the deep learning neural network model is trained according to the clear image, the blurred image and the global optimal modulation phase data, specifically including:

[0025] The blurred image and the corresponding global optimal modulation phase are input, and the target output is a clear image. The mean square error (MSE) is used as the loss function, and the Adam optimizer is used to minimize the loss function. The corresponding parameters of the deep learning neural network model are optimized. The parameters of the deep learning neural network model are adjusted during the training process so that the deep learning neural network model can accurately predict the optimal phase modulation scheme, thereby enhancing the image quality.

[0026] Furthermore, the deep learning neural network model is a U-Net convolutional neural network model.

[0027] Furthermore, the U-Net convolutional neural network model includes an encoder and a decoder;

[0028] The encoder gradually extracts the features of the input image through a series of convolutional layers and pooling layers, reducing the spatial dimension layer by layer. The bottleneck layer connects the encoder and the decoder and extracts high-level features through convolution operations. The decoder gradually restores the spatial resolution of the image through deconvolution layers and directly passes the feature map of the corresponding layer of the encoder to the decoder through skip connections, thereby retaining more image details.

[0029] The present invention has the following beneficial effects compared with the prior art:

[0030] 1. The system of the present invention includes a camera module, a beam expander, a polarizer, a beam splitter, a spatial light modulator, a scattering medium, and a focusing lens. The camera module is responsible for capturing the clear image that a normal human eye should receive and the blurred image after the modulated light beam passes through the scattering medium. The beam expander expands the light beam to an appropriate size. The polarizer controls the polarization state of the light to optimize the system performance. The beam splitter realizes the separation and combination of optical signals. The spatial light modulator modulates the light wavefront. The scattering medium simulates the eye turbidity of cataract patients for system testing. Through the optimized combination of these modules, efficient and accurate visual enhancement effects can be achieved in the optical path construction.

[0031] 2. The present invention introduces an iterative optimization algorithm, which can effectively improve the image quality of the target area by iterating the feedback of the quality of the target area of ​​the image. Compared with traditional static image processing technology, this method has higher dynamic optimization ability and adaptability, and can adjust the optimization strategy in real time to obtain the best visual enhancement effect.

[0032] 3. The present invention demonstrates excellent performance in image processing and enhancement by using a deep learning neural network model. The deep learning neural network model preferably adopts a U-Net convolutional neural network model. The encoder-decoder structure of the U-Net convolutional neural network model enables it to efficiently extract and restore the spatial features of the image, and is particularly suitable for processing the visual images of cataract patients. Compared with traditional image processing methods, the deep learning model has higher accuracy and efficiency, and can predict and generate the optimal phase modulation scheme in real time, thereby significantly enhancing the visual quality.

[0033] 4. The present invention achieves non-invasive visual enhancement through precise wavefront modulation. Compared with traditional surgery and drug treatments, this method not only avoids the risks and discomfort during the treatment process, but also provides immediate visual improvement effects. Patients can obtain significant visual improvement without undergoing complicated medical procedures, greatly improving their quality of life.

[0034] 5. The present invention not only shows remarkable effects in enhancing cataract vision, but also has broad application prospects. Through the combination of precise wavefront modulation and deep learning models, it achieves non-invasive visual enhancement, which can effectively improve the quality of life of patients with low vision. It can also be extended to the frontier fields of optics and medicine. It can be used for patients with other ophthalmic diseases (such as corneal scars, etc.) to enhance vision and reduce surgical risks. It can even be used in intelligent visual aids to help the visually impaired get a clearer visual experience. With the increase in market demand, this technology has huge potential for integration and commercialization. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. 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 the structures shown in these drawings without paying creative work.

[0036] Figure 1 Schematic diagram of the structure of a real-time cataract vision enhancement system based on wavefront sensing technology according to an embodiment of the present invention.

[0037] Figure 2 Schematic diagram of the optical path implementation principle of the real-time cataract vision enhancement system based on wavefront sensing technology in an embodiment of the present invention.

[0038] Among them, 1-first camera, 2-beam expander, 3-polarizer, 4-beam splitter, 5-spatial light modulator, 6-scattering medium, 7-focusing lens, 8-second camera, 9-iterative optimization algorithm, 10-phase, 11-deep neural network model, 12-enhanced image. DETAILED DESCRIPTION

[0039] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0040] It should be understood that when used in this specification and the appended claims, the terms "include" and "comprises" indicate the presence of described features, integers, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or combinations thereof.

[0041] The description of the embodiments should be understood in conjunction with the attached drawings. The embodiments are only used to illustrate specific embodiments that can be used to implement the present invention, and do not limit the present invention. The directional terms mentioned in the present invention, such as "upper", "lower", "front", "back", "left", "right", "inner", "outer", "side", etc., are only referenced to the directions of the attached drawings. Therefore, the directional terms used are used to illustrate and understand the present invention, but not to limit the present invention.

[0042] Embodiment 1:

[0043] like Figure 1As shown, this embodiment provides a real-time cataract vision enhancement system based on wavefront sensing technology, which includes a camera module, a beam expander 2, a polarizer 3, a beam splitter 4, a spatial light modulator 5, a scattering medium 6 and a focusing lens 7. The beam expander 2, the polarizer 3, the beam splitter 4, the scattering medium 6 and the focusing lens 7 are connected in sequence, and the spatial light modulator performs phase modulation on the light wavefront of the light beam separated by the beam splitter 4. The camera module collects a clear image that is not disturbed by the scattering medium 6 and a blurred image that is focused by the focusing lens 7 after the light beam modulated by the spatial light modulator 5 passes through the scattering medium 6, and the phase of the spatial light modulator 5 is optimized based on the image quality of the target area of ​​the clear image and the blurred image.

[0044] Furthermore, the camera module includes two cameras, namely a first camera 1 and a second camera 8. The first camera 1 captures a clear image that is not disturbed by the scattering medium 6, that is, it captures the image that a normal human eye should receive, and uses this as the target of iterative optimization. The second camera 8 captures a blurred image that is a light beam modulated by the spatial light modulator 5 and passed through the scattering medium 6 and focused by the focusing lens 7, so as to simulate the retina of the human eye.

[0045] In this embodiment, the beam expander 2 adopts a continuously variable magnification beam expander, which can expand the clear image light beam collected by the camera. Specifically, it provides a hand wheel for variable magnification and focusing to modulate the beam expansion ratio and the divergence angle of the output beam. The beam expander 2 expands the light beam to an appropriate size, and by precisely adjusting the size and divergence angle of the light beam, it ensures that the parallel light beam generated by the beam expansion can evenly cover the effective area of ​​the spatial light modulator 5, thereby improving the accuracy and efficiency of the experiment.

[0046] In this embodiment, the polarizer 3 adopts a thin film polarizer, which can control the polarization state of the expanded light beam. The polarizer 3 is embedded in the mechanical clamping ring with a rotating lens mount, which can continuously rotate 360° to control the polarization state of the light beam to match the arrangement direction of the liquid crystal molecules for phase modulation by the spatial light modulator; the beam splitter 4 can realize the separation and combination of the light beam, ensuring that it is vertically incident on the spatial light modulator 5 and the outgoing light is perpendicular to the incident light, and the subsequent optical elements will not affect the light beam to be modulated from being incident on the light modulator 5; the spatial light modulator 5 performs phase modulation on the light wavefront of the separated light beam, and its phase is generated by an iterative optimization algorithm 9; the scattering medium 6 adopts a turbid scattering medium to simulate cataracts. The patient's eye turbidity is used to simulate the real patient's condition; the modulated light beam of the beam splitter 4 is focused by the focusing lens through the scattering medium and captured by the second camera 8 to obtain a restored image (blurred image) after being disturbed by the scattering medium; the images collected by the first camera 1 and the second camera 8 are compared for the image quality of the target area, which is used as a basis for judging whether the iterative optimization algorithm 9 converges, so that the iterative phase 10 gradually converges to the optimal modulation phase, and the restored image of the modulated light beam by the spatial light modulator 5 after being disturbed by the scattering medium is realized. The optimal phase is applied to the spatial light modulator 5 to generate a phase modulation wavefront in advance, which can compensate for the phase disturbance caused by the scattering medium, so that the image quality of the target area is restored.

[0047] Furthermore, in the construction of the system of this embodiment, the first camera 1 captures a clear image that is not disturbed by the scattering medium, and the second camera 1 captures a blurred image that is focused by the focusing lens 7 after the light beam modulated by the spatial light modulator 5 passes through the scattering medium 6. It is necessary to ensure that the images captured by the first camera 1 and the second camera 2 are consistent in size and resolution and calibrated for consistency; in addition, the position and angle of each component in the system need to be precisely adjusted to ensure that the light beam can be accurately transmitted and modulated. To achieve this, high-precision mechanical adjustment devices and stable brackets can be used to fix the optical components.

[0048] Furthermore, the iterative optimization algorithm 9 is used to obtain the optimal modulation phase to modulate the disturbance wavefront to enhance the blurred image; the iterative optimization algorithm 9 can adopt a continuous sequential algorithm (Continuous Sequential Algorithm, CSA), which searches for the optimal solution by gradually adjusting the phase to obtain the optimal modulation phase; the iterative optimization algorithm 9 can also adopt a genetic algorithm, which searches for the global optimal solution by simulating natural selection and genetic variation to obtain the optimal modulation phase; the combination of these two algorithms can effectively improve the optimization efficiency and the accuracy of the results; when the iterative optimization algorithm converges to the optimal phase, the spatial light modulator 5 will generate a phase modulated wavefront, which can compensate for the phase disturbance when the light beam passes through the scattering medium, thereby restoring the image quality of the target area; in the iterative optimization algorithm 9, enhancement (Enhancement, ENH) and correlation coefficients (Correlation Coefficients, CC) are used to evaluate the enhancement effect of the target area of ​​the image after iteration.

[0049] In this embodiment, MATLAB is used to drive the spatial light modulator to perform phase iteration. The image data collected by the camera is fed back and processed into a grayscale matrix. The convergence is determined by evaluating the parameters to determine whether to perform the next iteration cycle. Finally, a clear image, a blurred image and an image data set with a global optimal modulation phase are obtained.

[0050] like Figure 1 and Figure 2 As shown, this embodiment also provides a real-time cataract vision enhancement method, which includes the following steps:

[0051] S1. The first camera 1 of the camera module captures a clear image without being disturbed by a scattering medium.

[0052] S2. The light beam generated by the laser is expanded by the beam expander 2, the polarizer 3 controls the polarization state of the expanded light beam, the beam splitter 4 separates the light beam, the spatial light modulator 5 modulates the phase of the light wavefront of the separated light beam, and the modulated light beam returns to the beam splitter 4. The modulated light beam of the beam splitter 4 is focused by the focusing lens 7 through the scattering medium 6.

[0053] S3 . The second camera 8 of the camera module captures a blurred image of the modulated light beam that passes through the scattering medium 6 and is focused by the focusing lens 7 .

[0054] S4. Optimize the spatial light modulator phase based on the image quality of the target area of ​​the clear image and the blurred image to obtain the global optimal modulation phase data.

[0055] In this embodiment, an iterative optimization algorithm is used to optimize the phase of the spatial light modulator based on the image quality of the target area of ​​the clear image and the blurred image. Specifically, the image quality of the target area of ​​the clear image and the blurred image is compared as a basis for judging whether the iterative optimization algorithm has converged, so that the spatial light modulator phase after the iteration gradually converges to the optimal modulation phase, and finally the global optimal modulation phase data is obtained.

[0056] S5. Train the deep learning neural network model based on the clear image, blurred image and global optimal modulation phase data.

[0057] Furthermore, the deep learning neural network model is a U-Net convolutional neural network model, which includes an encoder and a decoder; the encoder gradually extracts the features of the input image through a series of convolutional layers and pooling layers, reducing the spatial dimension layer by layer but increasing the number of features, the bottleneck layer connects the encoder and the decoder, and extracts high-level features through more convolution operations; the decoder gradually restores the spatial resolution of the image through the deconvolution layer, and directly passes the feature map of the corresponding layer of the encoder to the decoder through skip connections, thereby retaining more image details.

[0058] A large number of clear images, blurred images and global optimal modulation phase data collected in this embodiment constitute an image data set. When training the deep learning neural network model, the blurred image and the corresponding global optimal modulation phase are input, and the target output is a clear image. The mean square error MSE is used as the loss function, and the Adam optimizer is used to minimize the loss function to optimize the corresponding parameters of the deep learning neural network model. The parameters of the deep learning neural network model are adjusted during the training process so that the deep learning neural network model can accurately predict the optimal phase modulation scheme, thereby enhancing the image quality.

[0059] S6. Based on the trained deep learning neural network model, the real-time acquired image data is input, and the phase modulation parameters are predicted and optimized to adjust the spatial light modulator in real time, so that the real-time enhancement effect of cataract patients in the target area of ​​the visual image is optimized, and efficient enhanced image 12 is achieved.

[0060] Those skilled in the art will appreciate that the spatial light modulator in the above embodiment may also be replaced by a digital micromirror array to achieve the phase modulation function.

[0061] In summary, the system of the present invention includes a camera module, a beam expander, a polarizer, a beam splitter, a spatial light modulator, a scattering medium and a focusing lens. The camera module collects a clear image without being disturbed by the scattering medium and a blurred image after the modulated light beam is disturbed, and feeds back the quality of the target area of ​​the image through an iterative optimization algorithm, and optimizes the phase of the spatial light modulator to improve the image quality of the target area; the system collects clear images, blurred images and a global optimal modulation phase data set during the optical path hardware and software iteration process, adopts a deep learning neural network model for training, and uses the deep learning neural network model to predict and generate the optimal phase modulation scheme, so as to achieve real-time enhancement of the vision of cataract patients, realize non-invasive and efficient visual enhancement, and has significant clinical and commercial application potential.

[0062] The above is only a preferred embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical solution and inventive concept of the present invention within the scope disclosed by the present invention, which shall fall within the protection scope of the present invention.

Claims

1. A real-time cataract vision enhancement system based on wavefront sensing technology, characterized in that: The invention comprises a camera module, a beam expander, a polarizer, a beam splitter, a light modulation component, a scattering medium and a focusing lens, wherein the beam expander, the polarizer, the beam splitter, the scattering medium and the focusing lens are connected in sequence, the light modulation component performs phase modulation on the light wavefront of the light beam separated by the beam splitter, the camera module collects a clear image which is not disturbed by the scattering medium and a blurred image which is focused by the focusing lens after the light beam modulated by the light modulation component passes through the scattering medium, and the phase of the light modulation component is optimized based on the image quality of the target area of ​​the clear image and the blurred image.

2. The real-time cataract vision enhancement system according to claim 1, characterized in that: The camera module includes a first camera and a second camera, wherein the first camera collects a clear image without being disturbed by a scattering medium, and the second camera collects a blurred image focused by a focusing lens after a light beam modulated by a light modulation component passes through a scattering medium, for simulating the retina of a human eye.

3. A method for real-time enhancement of cataract vision, implemented based on the real-time enhancement system for cataract vision according to any one of claims 1-2, characterized in that: The method comprises: The first camera of the camera module captures a clear image that is not disturbed by a scattering medium; The light beam generated by the laser is expanded by a beam expander, the polarizer controls the polarization state of the expanded light beam, the beam splitter separates the light beam, the light modulation component modulates the phase of the light wavefront of the separated light beam, and the modulated light beam returns to the beam splitter. The modulated light beam of the beam splitter is focused by a focusing lens through a scattering medium; The second camera of the camera module captures a blurred image of the modulated light beam passing through the scattering medium and focused by the focusing lens; Optimize the phase of the optical modulation component based on the image quality of the target area of ​​the clear image and the blurred image to obtain the global optimal modulation phase data; The deep learning neural network model is trained based on the clear image, blurred image and global optimal modulation phase data; Based on the trained deep learning neural network model, the real-time acquired image data is input, and the phase modulation parameters are predicted and optimized to adjust the light modulation components in real time, so that the real-time enhancement effect of cataract patients on the target area of ​​the visual image is optimized.

4. The method for real-time enhancement of cataract vision according to claim 3, characterized in that: The method of optimizing the phase of the light modulation component based on the image quality of the target area of ​​the clear image and the blurred image specifically includes: Based on the image quality of the target area of ​​the clear image and the blurred image, an iterative optimization algorithm is used to optimize the phase of the light modulation component.

5. The method for real-time enhancement of cataract vision according to claim 4, characterized in that: The method of optimizing the phase of the light modulation component based on the image quality of the target area of ​​the clear image and the blurred image by using an iterative optimization algorithm specifically includes: The image quality of the target area of ​​the clear image and the blurred image is compared as a basis for judging whether the iterative optimization algorithm has converged, so that the phase of the optical modulation component after iteration gradually converges to the optimal modulation phase.

6. The method for real-time enhancement of cataract vision according to any one of claims 4-5, characterized in that: The iterative optimization algorithm adopts a continuous sequence algorithm to search for a global optimal solution by gradually adjusting the phase to obtain an optimal modulation phase.

7. The method for real-time enhancement of cataract vision according to any one of claims 4 to 5, characterized in that: The iterative optimization algorithm adopts a genetic algorithm to search for a global optimal solution by simulating natural selection and genetic variation to obtain an optimal modulation phase.

8. The method for real-time enhancement of cataract vision according to claim 3, characterized in that: The deep learning neural network model is trained according to the clear image, the blurred image and the global optimal modulation phase data, specifically including: The blurred image and the corresponding global optimal modulation phase are input, and the target output is a clear image. The mean square error (MSE) is used as the loss function, and the Adam optimizer is used to minimize the loss function. The corresponding parameters of the deep learning neural network model are optimized. The parameters of the deep learning neural network model are adjusted during the training process so that the deep learning neural network model can accurately predict the optimal phase modulation scheme, thereby enhancing the image quality.

9. The method for real-time enhancement of cataract vision according to claim 3, characterized in that: The deep learning neural network model is a U-Net convolutional neural network model.

10. The method for real-time enhancement of cataract vision according to claim 9, characterized in that: The U-Net convolutional neural network model includes an encoder and a decoder; The encoder gradually extracts the features of the input image through a series of convolutional layers and pooling layers, reducing the spatial dimension layer by layer. The bottleneck layer connects the encoder and the decoder and extracts high-level features through convolution operations. The decoder gradually restores the spatial resolution of the image through deconvolution layers and directly passes the feature map of the corresponding layer of the encoder to the decoder through skip connections, thereby retaining more image details.

Images