A cross-platform or cross-scenario medical image detection and feedback system and method thereof
Patent Information
- Authority / Receiving Office
- TW · TW
- Patent Type
- Patents
- Current Assignee / Owner
- NAT TAIWAN UNIV
- Filing Date
- 2023-08-10
- Publication Date
- 2026-08-01
Smart Images
Figure TWG2TB001903483_001 
Figure TWG2TB001903483_002 
Figure TWG2TB001903483_003
Abstract
Description
A Cross-Platform or Cross-Scenario Medical Image Detection and Feedback System and Its Method The present invention relates to a medical image detection technology, and particularly to a cross-platform or cross-scenario medical image detection and feedback system and its method. With the development of technology, medical images play an important role in diagnosis and treatment. However, traditional medical image systems usually have the following problems: 1. Traditional medical image systems can only operate under specific platforms or vendor systems, and cannot perform data exchange or integration across platforms or vendors. Moreover, the general public cannot obtain or directly analyze the original files of medical images (such as DICOM (Digital Imaging and Communications in Medicine)) from vendor systems. 2. Traditional medical image systems also require professional equipment and their operators, and it is difficult to adapt to different usage scenarios and requirements. 3. The medical images in traditional medical image systems need to be interpreted and explained by professionals, so the information contained in the medical image area displayed on the device (such as a computer screen or a wearable device display) cannot be fed back in real time, resulting in information gaps and inconveniences for general users. Therefore, how to detect the medical image area from a complex and changing environment to immediately feed back the information represented in the medical image area has become an urgent issue in the industry. To solve the aforementioned conventional technical problems or provide related effects, the present invention provides a cross-platform or cross-scenario medical image detection and feedback system, which includes: an image interface module that receives real environment images or virtual environment images from a user device, and both the real environment images and the virtual environment images include medical images; an image processing module that is communicatively connected to the image interface module to receive the real environment images or virtual environment images from the image interface module, and then detects the medical images from the real environment images or the virtual environment images; an image optimization module that is communicatively connected to the image processing module to receive the medical images from the image processing module, corrects the shape of the medical images, then removes the light and shadow in the medical images after shape correction, and enhances the pixels of the medical images after light and shadow removal, thereby obtaining optimized medical images; and an information feedback module that is communicatively connected to the image optimization module to receive the optimized medical images from the image optimization module, detects the medical image information in the optimized medical images, and then feeds back the medical image information to the user device. The present invention further provides a cross-platform or cross-scenario medical image detection and feedback method, which includes: receiving, by an image interface module, a real environment image or a virtual environment image from a user device, and both the real environment image and the virtual environment image include medical images; receiving, by an image processing module, the real environment image or the virtual environment image from the image interface module, and detecting the medical image from the real environment image or the virtual environment image; receiving, by an image optimization module, the medical image from the image processing module, correcting the shape of the medical image, removing the light and shadow in the shape-corrected medical image, and enhancing the pixels of the light-and-shadow-removed medical image, thereby obtaining an optimized medical image; and receiving, by an information feedback module, the optimized medical image from the image optimization module, detecting the medical image information in the optimized medical image, and further feeding back the medical image information to the user device. In the foregoing embodiment, the image optimization module includes an image correction model for correcting the shape of the medical image, a light and shadow removal model for removing the light and shadow in the shape-corrected medical image, and a pixel enhancement model for enhancing the pixels in the light-and-shadow-removed medical image. In the foregoing embodiment, the image optimization module uses a paired dataset to train the image correction model, the light and shadow removal model, and the pixel enhancement model. Among them, the image correction model, the light and shadow removal model, and the pixel enhancement model are all generative adversarial network models, and the paired dataset is composed of a plurality of original medical images and their corresponding plurality of training medical images, and the plurality of training medical images are obtained by reducing the resolution and distorting the plurality of original medical images. In the foregoing embodiment, the user device is an augmented reality device, and the user device displays the image type, the position, name, or prompt of the body part or organ structure on the medical image viewed by a user through the user device according to the medical image information. In the foregoing embodiment, the user device is a virtual reality device, and the user device further receives the optimized medical image from the information feedback module, displays the optimized medical image in virtual reality, and displays the image type, the position, name, or prompt of the body part or organ structure on the optimized medical image according to the medical image information. In the foregoing embodiment, the cross-platform or cross-scenario medical image detection and feedback system is established in a backend server. In the foregoing embodiment, the cross-platform or cross-scenario medical image detection and feedback system is established in the user device. As described above, the cross-platform or cross-scenario medical image detection and feedback system and method of the present invention capture medical images from real environment images or virtual environment images of a user device through an image processing module, and perform optimization processes such as shape correction, light and shadow removal, and pixel enhancement on the medical images by an image optimization module, thereby obtaining optimized medical images for subsequent identification of organ tissues in the images. In this regard, the image optimization module can improve the recognizability of medical images in complex and variable environments, and then immediately feedback the information represented by the medical images to the user, enabling the user to quickly and clearly know the information displayed by the medical images. 1: Cross-platform or cross-scenario medical image detection and feedback system 11: Image interface module 12: Image processing module 12a: Image capture model 13: Image optimization module 13a: Image correction model 13b: Light and shadow removal model 13c: Pixel enhancement model 14: Information feedback module 14a: Image detection model 2a: Real environment image 2b: Virtual environment image 7: Screen 8: User 9: User device A: Medical image S41 to S44: Steps S51A to S53A: Steps S51B to S53B: Steps S51C to S53C: Steps FIG. 1 is a schematic structural diagram of the cross-platform or cross-scenario medical image detection and feedback system of the present invention. FIG. 2A is a schematic diagram of the real environment image of the present invention. FIG. 2B is a schematic diagram of the virtual environment image of the present invention. FIG. 3A is a schematic diagram of the medical image in the real environment image of the present invention. FIG. 3B is a schematic diagram of the medical image in the virtual environment image of the present invention. FIG. 4 is a training flowchart of the image capture model of the present invention. FIG. 5A is a training flowchart of the image correction model of the present invention. FIG. 5B is a training flowchart of the light and shadow removal model of the present invention. FIG. 5C is a training flowchart of the pixel enhancement model of the present invention. The following describes the implementation modes of the present invention through specific embodiments. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. It should be noted that the structures, proportions, sizes, etc. shown in the drawings attached to this specification are only used to cooperate with the content disclosed in the specification for the understanding and reading of those skilled in the art, and are not used to limit the limiting conditions under which the present invention can be implemented. Therefore, they do not have substantial technical significance. Any modification of the structure, change of the proportional relationship, or adjustment of the size, without affecting the effects that the present invention can produce and the purposes that can be achieved, should still fall within the scope covered by the technical content disclosed in the present invention. At the same time, the terms such as "a", "first", "second", "upper", and "lower" cited in this specification are also only for the convenience of clear description, and are not used to limit the scope of implementation of the present invention. The change or adjustment of their relative relationships, without substantial change in the technical content, should be regarded as the scope within which the present invention can be implemented. FIG. 1 is a schematic structural diagram of a cross-platform or cross-scenario medical image detection and feedback system 1 of the present invention. The cross-platform or cross-scenario medical image detection and feedback system 1 includes: an image interface module 11, an image processing module 12, an image optimization module 13, and an information feedback module 14. Specifically, the cross-platform or cross-scenario medical image detection and feedback system 1 can be established in a server (such as a general-purpose server, a file server, a storage unit server, etc.) and a computer and other backend electronic devices with appropriate mechanisms, or in an augmented reality device (such as an AR (Augmented Reality) glasses) that supports edge computing, a virtual reality device (such as a VR (Virtual Reality) glasses), a smart phone, and a tablet computer 9 and other user devices with a camera and a screen. Among them, each module or model in the cross-platform or cross-scenario medical image detection and feedback system 1 (that is, the image interface module 11, the image processing module 12, the image optimization module 13, and the information feedback module 14) can be software, hardware, or firmware; if it is hardware, it can be a processing unit, a processor, a computer, or a server with data processing and computing capabilities; if it is software or firmware, it can include instructions executable by a processing unit, a processor, a computer, or a server, and can be installed on the same hardware device or distributed on different multiple hardware devices. The described image interface module 11 receives or obtains a real environment image 2a or a virtual environment image 2b. Among them, as shown in FIG. 2A, the real environment image 2a refers to an image containing medical images in the real environment, for example: all the screens containing medical images in the work field of medical staff or the living field of the general public; as shown in FIG. 2B, the virtual environment image 2b refers to an image containing medical images in a virtual environment (such as the screen display), for example: all the screens displayed in a computer platform, a manufacturer's instrument, or an application program. Furthermore, these images include complex physical, virtual, or augmented scenes. In an embodiment, the cross-platform or cross-scene medical image detection and feedback system 1 is established in the user device 9. The camera of the user device 9 captures the real environment image 2a or the virtual environment image 2b, so that the image interface module 11 obtains the real environment image 2a or the virtual environment image 2b. Among them, the user device 9 is a wearable device such as HoloLens 2 launched by Microsoft. In an embodiment, the cross-platform or cross-scene medical image detection and feedback system 1 is established in a backend device and is communicatively (or electrically) connected to a user device 9 (such as AR glasses), so that the image interface module 11 receives the real environment image 2a or the virtual environment image 2b from the user device 9. Among them, the backend device is a server (such as a general-purpose server, a file server, a storage unit server, etc.). The described image processing module 12 includes an image capture model 12a, and the image processing module 12 is communicatively (or electrically) connected to the image interface module 11 to obtain the real environment image 2a or the virtual environment image 2b from the image interface module 11. Then, the image capture model 12a detects a medical image A from the real environment image 2a or the virtual environment image 2b (as shown in FIGS. 3A and 3B). In this embodiment, the image capture model 12a can be a YOLO (You Only Look Once) object detection model in a Convolutional Neural Network (CNN), or a segmentation model of the Region-based Convolutional Neural Networks (R-CNN) series. The image capture model 12a performs deep learning based on a plurality of training images so that the trained image capture model 12a can extract the medical image A from the real environment image 2a or the virtual environment image 2b. Among them, the plurality of training images include a plurality of real environment images and a plurality of virtual environment images for training, and the positions of the medical images are labeled on the plurality of real environment images and the plurality of virtual environment images for training. Specifically, as shown in FIG. 4, the training flowchart of the image capture model 12a includes the following steps S41 to S44: In step S41, a plurality of real environment images and a plurality of virtual environment images are collected. In step S42, the regions of the medical images in the plurality of real environment images and the plurality of virtual environment images are labeled to obtain a plurality of real environment images and a plurality of virtual environment images for training. In step S43, a neural network model is trained based on the plurality of real environment images for training to obtain a pre-trained model. In step S44, the pre-trained model is trained based on the plurality of real environment images for training to train the image capture model 12a. In an embodiment, the image processing module 12 or an external electronic device (such as a computer, etc.) uses a method such as linear algebra to reduce the resolution and distort a plurality of original medical images (such as high-resolution pure medical image originals in DICOM or PNG format) to obtain a plurality of training medical images corresponding to the plurality of original medical images, so as to combine the plurality of original medical images and their corresponding plurality of training medical images into a paired data set for subsequent training of the deep learning model. The described image optimization module 13 is communicatively (or electrically) connected to the image processing module 12 to receive the medical image A from the image processing module 12. The image optimization module 13 performs shape correction on the medical image A to make the medical image A in a rectangular shape. Then, the image optimization module 13 removes the light and shadow from the shape-corrected medical image A to remove the light and shadow in the medical image A. Moreover, the image optimization module 13 performs super-resolution imaging (SR) (i.e., pixel enhancement) on the light-and-shadow-removed medical image A to improve the image quality of the medical image A, thereby obtaining the optimized medical image A. In an embodiment, the image optimization module 13 includes an image correction model 13a, and the medical image A is corrected to a rectangular shape by the image correction model 13a. Specifically, the image correction model 13a includes, but is not limited to, a model of a convolutional neural network (CNN) or a generative adversarial network (GAN). When the image correction model 13a is a generative adversarial network model, the image correction model 13a includes a first generator and a first discriminator. Among them, as shown in FIG. 5A, the training flowchart of the image correction model 13a includes the following steps S51A to S53A: In step S51A, the first generator in the image correction model 13a performs shape correction on the training medical images in the paired dataset to make the training medical images in a rectangular shape. In step S52A, the first discriminator in the image correction model 13a determines the authenticity of the shape-corrected training medical image based on the shape-corrected training medical image and its corresponding original medical image, that is, determines whether the probability distribution (such as uniform distribution, normal distribution, etc.) of the pixel group or the radiomic feature of the shape-corrected training medical image approaches the original medical image, thereby determining whether the shape-corrected training medical image is corrected to a rectangle. In step S53A, the image optimization module 13 uses the first loss function generated by determining the authenticity of the shape-corrected training medical image to optimize the image correction model 13a, where the first loss function includes adversarial training loss (GAN Loss), pixel recovery loss (Pixel Loss), and perceptual loss (Perceptual Loss). In addition, the image correction model 13a can also correct the medical image A by using a correction algorithm, and the correction algorithm can be a method based on geometric transformation, such as Keystone correction. Among them, Keystone correction is used to convert the trapezoidal perspective effect (caused by the angle of the screen or projector) into a rectangular shape, which can be achieved by calculating and applying the correction transformation to correct the distortion effect caused by the installation position or angle of the screen. In an embodiment, the image optimization module 13 includes a light and shadow removal model 13b, and the light and shadow removal model 13b removes the light and shadow in the shape-corrected medical image A. Specifically, the light and shadow removal model 13b includes, but is not limited to, a model of a convolutional neural network (CNN) or a generative adversarial network (GAN). When the light and shadow removal model 13b is a generative adversarial network model, the light and shadow removal model 13b includes a second generator and a second discriminator. Among them, as shown in FIG. 5B, the training flowchart of the light and shadow removal model 13b includes the following steps S51B to step S53B: In step S51B, the second generator in the light and shadow removal model 13b removes the light and shadow from the training medical images in the paired dataset. In step S52B, the second discriminator in the light and shadow removal model 13b determines the authenticity of the light and shadow-removed training medical image based on the light and shadow-removed training medical image and its corresponding original medical image, that is, determines the probability distribution (such as uniform distribution, normal distribution, etc.) of the pixel group of the light and shadow-removed training medical image or whether the radiomic feature approaches the original medical image, thereby determining whether the light and shadow of the light and shadow-removed training medical image have been removed. In step S53B, the image optimization module 13 uses the second loss function generated by determining the authenticity of the light and shadow-removed training medical image to optimize the light and shadow removal model 13b, where the second loss function includes adversarial training loss (GAN Loss), pixel recovery loss (Pixel Loss), and perceptual loss (Perceptual Loss). In one embodiment, the image optimization module 13 includes a pixel enhancement model 13c. The pixel enhancement model 13c performs super-resolution imaging (SR) on the medical image A from which light and shadow have been removed, so as to enhance the pixels of the medical image A and obtain an optimized medical image A. Specifically, the pixel enhancement model 13c includes, but is not limited to, a model of a convolutional neural network (CNN) or a generative adversarial network (GAN). When the pixel enhancement model 13c is a generative adversarial network model, the pixel enhancement model 13c includes a third generator and a third discriminator. Among them, as shown in FIG. 5C, the training flowchart of the pixel enhancement model 13c includes the following steps S51C to S53C: In step S51C, the third generator in the pixel enhancement model 13c enhances the pixels of the training medical images in the paired dataset. In step S52C, the third discriminator in the pixel enhancement model 13c determines the authenticity of the pixel-enhanced training medical image based on the pixel-enhanced training medical image and its corresponding original medical image, that is, determines the probability distribution (such as uniform distribution, normal distribution, etc.) of the pixel group of the pixel-enhanced training medical image or whether the radiomic feature approaches the original medical image, thereby determining whether the pixel-enhanced training medical image has completed super-resolution imaging (SR). In step S53C, the image optimization module 13 uses the third loss function generated by determining the authenticity of the pixel-enhanced training medical image to optimize the pixel enhancement model 13c. Among them, the third loss function includes adversarial training loss (GAN Loss), pixel recovery loss (Pixel Loss), and perceptual loss (Perceptual Loss). The information feedback module 14 includes an image detection model 14a, which is communicatively (or electrically) connected to the image optimization module 13 to receive the optimized medical image A from the image optimization module 13, and then uses the image detection model 14a to detect medical image information from the optimized medical image A and feedback the medical image information to the user device 9. Among them, the image detection model 14a can be an object detection model (YOLO) in a convolutional neural network (CNN) or a segmentation model based on the region-based convolutional neural network (R-CNN) series, and the image detection model 14a can detect medical image information from the optimized medical image A after being trained by deep learning. In one embodiment, when the user device 9 is an augmented reality device, the user device 9, based on the medical image information, instantaneously displays the image type, the position, name, or prompt of the body part or organ structure on the medical image A viewed by the user 8 through the user device 9. Specifically, at this time, the medical image A viewed by the user 8 is an image in the real environment, that is, through the user device 9, the organ tissue can be directly marked on the medical image A in the real environment. In one embodiment, when the user device 9 is a virtual reality device, the user device 9 further receives the optimized medical image from the information feedback module 14 to directly display the optimized medical image A in the virtual reality, and based on the medical image information, displays the image type, the position, name, or prompt of the body part or organ structure on the optimized medical image A. In one embodiment, the medical image information includes the image type (such as X-ray image, CT image, etc.), the body part (such as chest, arm, etc.), or the position, name, or prompt of the organ structure. For example, the user device 9, based on the medical image information, displays the image type and the name of the body part of the medical image A to the user 8; and based on the position (such as two-dimensional coordinates) of the organ structure or each part in the medical image information, marks an object frame on the medical image A (or the optimized medical image A) directly viewed by the user 8, and displays the name (such as heart, etc.) or prompt (such as heart hypertrophy, etc.) of the organ structure marked by the object frame. The following are specific embodiments of the present invention, and are described in conjunction with FIGS. 1 to 5C. The same parts as those in the above embodiments will not be repeated. In this embodiment, a user 8 wears an augmented reality device (i.e., the user device) 9 so that the user views the screen 7 through the augmented reality device 9. At this time, the lens on the augmented reality device 9 captures a real environment image 2a (as shown in FIG. 2A), and transmits the real environment image 2a to a cross-platform or cross-scenario medical image detection and feedback system 1 provided in the augmented reality device 9, or the cross-platform or cross-scenario medical image detection and feedback system 1 provided in the backend server. Next, the image interface module 11 in the cross-platform or cross-scenario medical image detection and feedback system 1 receives the real environment image 2a, and then the image processing module 12 uses the image capture model 12a to detect a medical image A (as shown in FIG. 3A). In this regard, the image optimization module 13 receives the medical image A from the image processing module 12. Among them, the image optimization module 13 uses the image correction model 13a to correct the medical image A into a rectangular shape, then uses the light and shadow removal model 13b to remove the light and shadow in the shape-corrected medical image A, and also uses the pixel enhancement model 13c to perform super-resolution imaging (SR) on the light and shadow-removed medical image A to enhance the pixels of the medical image A, thereby obtaining an optimized medical image A through the image optimization module 13. Finally, the information feedback module 14 receives the optimized medical image A from the image optimization module 13, and the information feedback module 14 uses the image detection model 14a to detect a medical image information from the optimized medical image A, and then feeds back the medical image information to the augmented reality device 9. The user device 9 uses the medical image information to immediately display the image type (such as X-ray image), body part (such as chest), or the position, name, or prompt of the organ structure (such as heart) on the medical image A viewed by the user 8 through the augmented reality device 9, thereby providing the user 8 with information such as the position and name of the organ tissue contained in the medical image A he / she views. In summary, for the cross-platform or cross-scenario medical image detection and feedback system and its method of the present invention, mainly the image processing module extracts medical images from the real environment images or virtual environment images of the user device, and then the image optimization module performs optimization processes such as shape correction, light and shadow removal, and pixel enhancement on the medical images, thereby obtaining optimized medical images for subsequent identification of medical image information in the images. Therefore, the present invention can detect medical images from a complex and variable environment, and then the image optimization module can improve its recognizability, and then immediately feedback the information represented by the medical images to the user, enabling him / her to quickly and clearly know the information displayed by the medical images. In addition, the cross-platform or cross-scenario medical image detection and feedback system and its method of the present invention have the following advantages or technical effects: First, the present invention can detect medical-related images therein by collecting the set of all pixels in the picture in a complex environment, and then identify the information represented by these images, without the need for professional equipment, operators, and professional interpretation, enabling general users without medical-related expertise to also understand the information represented in medical-related images. Second, the present invention completely eliminates the need to obtain original files (such as DICOM), enabling data exchange or integration across platforms or scenarios, thereby enhancing the versatility and compatibility of the system. Third, the present invention can instantaneously feedback medical-related image area information to the original computer or other wearable device displays, improving user convenience and efficiency while also reducing the information gap for general users. Fourth, compared with the prior art where medical images captured are recognized and then displayed along with their recognition results on another screen or display, causing inconvenience to general users. In contrast, the present invention enables general users to intuitively understand the information contained in the current medical images through an augmented reality device (AR glasses), significantly enhancing the convenience of use. The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the present invention. Any person skilled in the art can modify and change the above embodiments without departing from the spirit and scope of the present invention. Therefore, the scope of the patent protection of the present invention shall be as set forth in the patent application scope. 1: Medical Image Detection and Feedback System across Platforms or Scenarios 11: Image Interface Module 12: Image Processing Module 12a: Image Capture Model 13: Image Optimization Module 13a: Image Calibration Model 13b: Light and Shadow Removal Model 13c: Pixel Enhancement Model 14: Information Feedback Module 14a: Image Detection Model 7: Screen 8: User 9: User Device
Claims
1. A cross-platform or cross-scenario medical image detection and feedback system, comprising: an image interfacing module for receiving real-world or virtual-world images captured by a user device worn on a user's body, wherein both the real-world and virtual-world images contain medical images; an image processing module communicatively connected to the image interfacing module for receiving the real-world or virtual-world images from the image interfacing module and detecting the medical images from the real-world or virtual-world images; an image optimization module communicatively connected to the image processing module for receiving the medical images from the image processing module, correcting the shape of the medical images, removing shadows from the shape-corrected medical images, and enhancing the pixels in the shadow-removed medical images, thereby obtaining optimized medical images; and an information feedback module communicatively connected to the image optimization module for receiving the optimized medical images from the image optimization module, detecting medical image information in the optimized medical images, and then feeding the medical image information back to the user device, wherein... The image optimization module includes an image correction model for correcting the shape of the medical image, a shading model for removing light and shadow from the shape-corrected medical image, and a pixel enhancement model for enhancing pixels in the shading-removed medical image. The image optimization module trains the image correction model, the shading model, and the pixel enhancement model using a pairwise dataset. The image correction model, the shading model, and the pixel enhancement model are all generative adversarial network models. The pairwise dataset consists of a plurality of original medical images and their corresponding plurality of training medical images. The plurality of training medical images are obtained by reducing the resolution and distorting the plurality of original medical images.
2. The cross-platform or cross-scenario medical image detection and feedback system as described in Request 1, wherein, The user device is an augmented reality device, and based on the medical image information, the user device displays the image type, location, name or prompts of body parts or organ structures on the medical images viewed by the user through the user device.
3. The cross-platform or cross-scenario medical image detection and feedback system as described in Request 1, wherein, The user device is a virtual reality device, and the user device receives the optimized medical image from the information feedback module, displays the optimized medical image in virtual reality, and displays the image type, body part or organ structure location, name or prompt on the optimized medical image according to the medical image information.
4. The cross-platform or cross-scenario medical image detection and feedback system as described in Request 1, wherein, The system is built on a backend server.
5. A cross-platform or cross-scenario medical image detection and feedback system as described in Request 1, wherein, The system is built into the user's device.
6. A cross-platform or cross-scenario medical image detection and feedback method, comprising: receiving, via an image interfacing module, real-world or virtual-world images captured by a user device worn on a user's body, wherein both the real-world or virtual-world images contain medical images; receiving, via an image processing module, the real-world or virtual-world images from the image interfacing module, and detecting the medical images from the real-world or virtual-world images; receiving, via an image optimization module, the medical images from the image processing module, correcting the shape of the medical images, removing shadows from the shape-corrected medical images, and enhancing the pixels of the shadow-removed medical images, thereby obtaining optimized medical images, wherein... The image optimization module performs shape correction on the medical image using an image correction model, removes light and shadow from the shape-corrected medical image using a shading model in the same module, and enhances the pixel count of the shading-removed medical image using a pixel enhancement model. The module also trains the image correction model, shading model, and pixel enhancement model using a pairwise dataset. All three models are generative adversarial networks (GANs), and the pairwise dataset consists of multiple original medical images and their corresponding multiple training medical images. The training medical images are obtained by reducing the resolution and distorting the original medical images. Finally, an information feedback module receives the optimized medical image from the image optimization module, detects medical image information within the optimized image, and feeds this information back to the user device.
7. The cross-platform or cross-scenario medical image detection and feedback method as described in claim 6 further includes the user device, which is an augmented reality device, displaying the image type, body part or organ structure location, name or prompt on the medical image viewed by the user through the user device based on the medical image information.
8. The cross-platform or cross-scenario medical image detection and feedback method as described in claim 6 further includes the user device, which is a virtual reality device, receiving the optimized medical image from the information feedback module to display the optimized medical image in virtual reality, and displaying the image type, body part or organ structure location, name or prompt on the optimized medical image based on the medical image information.