Endoscope image processing method, apparatus, device, and storage medium

Abstract

The application discloses an endoscope image processing method, device and equipment and a storage medium, and belongs to the technical field of endoscopes. The method comprises the following steps: acquiring a visible light image and an infrared image collected by an endoscope, extracting and enhancing feature information in the visible light image and the infrared image to obtain an enhanced visible light image and an enhanced infrared image, performing multi-scale fusion and reconstruction on the enhanced visible light image and the enhanced infrared image to obtain a reconstructed image. The application overcomes the respective disadvantages of visible light imaging and infrared imaging, realizes the enhancement and multi-scale fusion of the visible light image and the infrared image, generates a fusion image with rich texture and prominent targets, can be applied to an endoscope with high requirements for imaging quality effect, improves the definition of endoscope imaging, and thus improves the operation efficiency and reduces the operation risk.

Description

Endoscopic image processing methods, devices, equipment, and storage media Technical Field

[0001] This invention relates to the field of endoscopy technology, and more particularly to an endoscopic image processing method, apparatus, device, and storage medium. Background Technology

[0002] Most current medical endoscopes primarily use visible light for imaging, such as 4K endoscopes, three-chip endoscopes, and 3D endoscopes. This is because visible light images closely match human visual perception, containing rich details, textures, edge gradients, and colors. However, the penetrating power of visible light in the 400-700 nanometer wavelength range is limited. Tiny particles generated during surgery can almost completely block visible light, affecting image quality, causing loss of image detail, impacting the surgeon's judgment, and increasing surgical risks.

[0003] Infrared light, with wavelengths ranging from 700 to 1000 nanometers, has strong penetrating power, allowing it to pass through fine particles generated during surgery and obtain more information than visible light. However, infrared imaging alone suffers from problems such as high noise and poor visual effects.

[0004] In recent years, with the development of sensor technology and image processing technology, although there are some imaging methods that use image fusion technology to fuse visible light images and infrared images, most of these methods are just simple image fusion, resulting in images with poor texture, lack of detail, and image distortion. Because medical endoscopes are related to life and health and have very high requirements for image quality, existing methods are still not applicable to the field of medical endoscopy. Summary of the Invention

[0005] The main objective of this invention is to provide an endoscopic image processing method, apparatus, device, and storage medium, which aims to achieve an endoscope that integrates visible light and infrared images and has high-quality imaging effects.

[0006] To achieve the above objectives, the present invention provides an endoscopic image processing method, comprising:

[0007] Acquire visible light and infrared images captured by the endoscope;

[0008] Feature information is extracted and enhanced from the visible light image and infrared image to obtain enhanced visible light image and infrared image;

[0009] The enhanced visible light image and infrared image are fused at multiple scales and reconstructed to obtain the reconstructed image.

[0010] Optionally, the step of extracting and enhancing the feature information in the visible light image and infrared image to obtain the enhanced visible light image and infrared image includes:

[0011] The visible light image and the infrared image are decomposed to obtain their respective base layer and detail layer, wherein the detail layer includes the image feature information extracted by the decomposition.

[0012] The respective base layer and detail layer are enhanced to obtain the enhanced base layer and detail layer;

[0013] The base layer and detail layer of the enhanced visible light image and infrared image are combined to obtain the enhanced visible light image and infrared image respectively.

[0014] Optionally, the step of decomposing the visible light image and the infrared image respectively to obtain their respective base layer and detail layer includes:

[0015] The visible light image and the infrared image are respectively subjected to guided filtering processing to generate the base layer of the visible light image and the infrared image;

[0016] By subtracting the base layers of the original images of the visible light image and the infrared image, detail layers of the visible light image and the infrared image are obtained respectively. The detail layers contain feature information extracted from the original images by subtraction.

[0017] Optionally, the step of enhancing the respective base layer and detail layer to obtain enhanced base layer and detail layer includes:

[0018] Adaptive enhancement is performed on the base layer of the visible light image and the infrared image to obtain the enhanced base layer of the visible light image and the infrared image;

[0019] Nonlinear enhancement is performed on the detail layers of the visible light image and the infrared image to obtain the enhanced detail layers of the visible light image and the infrared image.

[0020] Optionally, after the step of acquiring the visible light image and infrared image captured by the endoscope, the method further includes:

[0021] Detect whether the bit depths of the visible light image and the infrared image are consistent;

[0022] If they are inconsistent, the high bit depth image is converted to a low bit depth image to obtain the registered visible light image and infrared image;

[0023] The registered visible light image and infrared image are normalized respectively to obtain normalized visible light image and infrared image;

[0024] The step of extracting and enhancing feature information from the visible light image and infrared image to obtain enhanced visible light image and infrared image includes:

[0025] Feature information is extracted and enhanced from the normalized visible light image and infrared image to obtain enhanced visible light image and infrared image.

[0026] Optionally, the step of performing multi-scale fusion and reconstruction on the enhanced visible light image and infrared image to obtain the reconstructed image includes:

[0027] Gaussian pyramids are generated for the enhanced visible light image and infrared image, respectively, and the Gaussian pyramids include images at different scales;

[0028] Calculate the Laplacian pyramid for the image in the Gaussian pyramid, the Laplacian pyramid including image detail information at different scales;

[0029] The corresponding levels of the Laplacian pyramids of the enhanced visible light image and infrared image are weighted and fused to obtain the fused Laplacian pyramid.

[0030] The merged Laplace pyramid image is added to the corresponding level image of the Gaussian pyramid to obtain the reconstructed image.

[0031] Optionally, the step of generating corresponding Gaussian pyramids for the enhanced visible light image and infrared image, wherein the Gaussian pyramids include images at different scales, includes:

[0032] The enhanced visible light image and infrared image are downsampled and processed with a Gaussian filter to obtain the next level image;

[0033] Repeat the above steps for the next level image to generate a set of images corresponding to the visible light image and the infrared image;

[0034] The images in the set are arranged according to their scale to form a corresponding Gaussian pyramid, and the downsampling is used to reduce the scale of the images.

[0035] Optionally, the step of calculating the Laplacian pyramid for the image in the Gaussian pyramid, wherein the Laplacian pyramid includes image detail information at different scales, includes:

[0036] The images in the Gaussian pyramid corresponding to the visible light image and the infrared image are subtracted from their respective upper-level images to obtain a set of images corresponding to each;

[0037] The images in the image set are arranged according to scale to form a corresponding Laplacian pyramid. Each layer of the Laplacian pyramid contains detailed information of the corresponding layer of the Gaussian pyramid, which is obtained by subtraction.

[0038] Furthermore, the present invention also provides an endoscopic image processing apparatus, the apparatus comprising:

[0039] The endoscope image acquisition module is used to acquire visible light and infrared images captured by the endoscope.

[0040] The endoscopic image enhancement module extracts and enhances the feature information in the visible light image and infrared image to obtain enhanced visible light image and infrared image;

[0041] The endoscopic image fusion module is used to perform multi-scale fusion and reconstruction of the enhanced visible light image and infrared image to obtain the reconstructed image.

[0042] Furthermore, the present invention also provides an endoscopic image processing device, the device comprising: a memory, a processor, and an endoscopic image processing program stored in the memory and executable on the processor, the endoscopic image processing program being configured to implement the steps of the endoscopic image processing method as described above.

[0043] Furthermore, the present invention also provides a storage medium storing an endoscope image processing program, which, when executed by a processor, implements the steps of the endoscope image processing method described above.

[0044] This invention proposes an endoscopic image processing method, apparatus, device, and storage medium. By acquiring visible light and infrared images from an endoscope, the method extracts and enhances feature information from these images to obtain enhanced visible light and infrared images. Multi-scale fusion and reconstruction are then performed on these enhanced images to obtain a reconstructed image. This invention extracts and enhances feature information from visible light and infrared images, and performs multi-scale fusion and reconstruction on the enhanced images. This solves the technical problems in existing imaging technologies where individual visible light or infrared imaging is easily affected by environmental factors or high noise levels, and the resulting images from image fusion often lack rich texture, detail, and image distortion. It achieves the enhancement and multi-scale fusion of visible light and infrared images, generating fused images with rich texture and prominent targets. This can be applied to endoscopes with high imaging quality requirements, improving the clarity of endoscopic imaging, thereby increasing surgical efficiency and reducing surgical risks. Attached Figure Description

[0045] Figure 1 is a schematic diagram of the device structure of the hardware operating environment involved in the embodiment of the present invention;

[0046] Figure 2 is a flowchart of the first embodiment of the endoscopic image processing method of the present invention;

[0047] Figure 3 is a schematic diagram of the image enhancement and fusion process in an embodiment of the endoscopic image processing method of the present invention;

[0048] Figure 4 is a flowchart of the second embodiment of the endoscopic image processing method of the present invention;

[0049] Figure 5 is a detailed flowchart of step S201 in an embodiment of the endoscopic image processing method of the present invention;

[0050] Figure 6 is a detailed flowchart of step S202 in an embodiment of the endoscopic image processing method of the present invention;

[0051] Figure 7 is a flowchart illustrating the image enhancement method in an embodiment of the endoscopic image processing method of the present invention;

[0052] Figure 8 is a flowchart of the third embodiment of the endoscopic image processing method of the present invention;

[0053] Figure 9 is a detailed flowchart of step S200 in an embodiment of the endoscopic image processing method of the present invention;

[0054] Figure 10 is a flowchart of the fourth embodiment of the endoscopic image processing method of the present invention;

[0055] Figure 11 is a schematic diagram of the process of fusing the Laplace pyramid in an embodiment of the endoscopic image processing method of the present invention;

[0056] Figure 12 is a flowchart of the fifth embodiment of the endoscopic image processing method of the present invention;

[0057] Figure 13 is a flowchart of the sixth embodiment of the endoscopic image processing method of the present invention;

[0058] Figure 14 is a schematic diagram of the process of constructing a Laplacian pyramid from a Gaussian image pyramid in an embodiment of the endoscopic image processing method of the present invention;

[0059] Figure 15 is a schematic diagram of the functional modules of an embodiment of the endoscopic image processing device of the present invention.

[0060] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0061] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0062] Referring to Figure 1, Figure 1 is a schematic diagram of the endoscopic image processing device structure of the hardware operating environment involved in the embodiment of the present invention.

[0063] As shown in Figure 1, the endoscopic image processing device may include: a processor 1001, such as a central processing unit (CPU), a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. The communication bus 1002 is used to enable communication between these components. The user interface 1003 may include a display screen and an input unit such as a keyboard; optionally, the user interface 1003 may also include a standard wired interface or a wireless interface. The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a Wireless-fidelity (Wi-Fi) interface). The memory 1005 may be a high-speed random access memory (RAM) or a stable non-volatile memory (NVM), such as a disk drive. The memory 1005 may also optionally be a storage device independent of the aforementioned processor 1001.

[0064] Those skilled in the art will understand that the structure shown in Figure 1 does not constitute a limitation on the endoscopic image processing device, and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0065] As shown in Figure 1, the memory 1005, which serves as a storage medium, may include an operating system, a data storage module, a network communication module, a user interface module, and an endoscope image processing program.

[0066] In the endoscopic image processing device shown in Figure 1, the network interface 1004 is mainly used for data communication with other devices; the user interface 1003 is mainly used for data interaction with the user; the processor 1001 and the memory 1005 in the endoscopic image processing device of the present invention can be set in the endoscopic image processing device, and the endoscopic image processing device calls the endoscopic image processing program stored in the memory 1005 through the processor 1001 and executes the endoscopic image processing method provided in the embodiment of the present invention.

[0067] Referring to Figure 2, which is a flowchart illustrating a first embodiment of the endoscopic image processing method of the present invention, the first embodiment of the present invention provides an endoscopic image processing method, the endoscopic image processing method comprising:

[0068] Step S100: Acquire visible light and infrared images captured by the endoscope;

[0069] In this embodiment, the endoscopic image processing method can be implemented by an endoscopic image processing device, which includes light signal acquisition and processing devices, including visible light image sensors and infrared image sensors.

[0070] Specifically, a visible light image sensor can capture light signals within a certain area, and after the captured light signals are processed by the photoelectric conversion device integrated in the sensor, they are converted into corresponding electrical signals, thereby obtaining a visible light image of the target object.

[0071] In addition, an active light source can be integrated into the endoscope image processing device to provide illumination when the ambient light intensity of the endoscope is insufficient, thereby assisting the visible light sensor in imaging.

[0072] Infrared image sensors can capture infrared radiation emitted or reflected by a target object within a certain area. Infrared radiation can be the object's own thermal radiation or from an external infrared radiation source. This external source can be provided by the aforementioned active light source, or by infrared lasers, or by fluorescent contrast agents. When infrared radiation strikes the infrared image sensor, the infrared photosensitive material within the sensor converts the radiation into electrical charges. These charges are read out and converted into electrical signals, thus obtaining a visualized infrared image.

[0073] It should be noted that infrared light with wavelengths ranging from 700 to 1000 nanometers has strong penetrating power, allowing it to penetrate fine particles generated during surgery and obtain more information than visible light. Therefore, the advantage of infrared imaging lies in its ability to improve the clarity and recognizability of target objects in special surgical environments. These special surgical environments include situations where tissue bleeding obstructs the lens's field of view, lens fog obstructs the view, insufficient light makes details in dark areas unclear, and blood vessel patterns are not clearly visible.

[0074] In this embodiment, the target object can be internal biological tissue, an organ, or the environment in which the tissue or organ is located. It should be understood that the term "tissue" or "organ" here does not constitute a limitation on the target object.

[0075] The visible light image I = (R, G, B) acquired by the visible light image sensor and the infrared image FI acquired by the infrared sensor are both images of the same target object under the same environment and at the same time. Furthermore, these visible light and infrared images can be displayed on a monitor or other output devices.

[0076] Step S200: Extract and enhance the feature information in the visible light image and infrared image to obtain the enhanced visible light image and infrared image;

[0077] In this embodiment, in order to take advantage of the respective advantages of visible light images and infrared images, it is necessary to extract and enhance the feature information of each of these two types of images for use in subsequent image processing.

[0078] Specifically, in this embodiment, feature information is first extracted from visible light images and infrared images.

[0079] The feature information of visible light and infrared images includes: color features, shape features, texture features, edge features, and spatial relationship features. Texture features refer to the textural details in the image, such as surface smoothness and roughness; shape features refer to the shape and outline of objects in the image; edge features refer to the edges between objects in the image; and spatial relationship features refer to the layout and organization of objects in the image.

[0080] In this embodiment, the feature information extracted from visible light and infrared images can include any of the aforementioned feature information. This feature information can reflect certain characteristics of the target object. For example, the color features of a visible light image can reflect information about the blood environment of a biological tissue or organ, while shape, texture, and edges can reflect information such as the boundaries between different tissues or organs. Similarly, the shape features of an infrared image can reflect the specific shape and outline of the target object.

[0081] To obtain clearer feature information about these target objects, it is necessary to extract feature information from both visible light and infrared images, and then enhance the extracted features. This allows us to leverage the respective imaging advantages of these two types of images to achieve better imaging results.

[0082] Specifically, the methods for extracting feature information from visible light images and infrared images can include the following:

[0083] For visible light images, color features can be extracted using methods such as color space conversion and color histograms. Texture information of objects in visible light images can be captured using texture descriptors (such as gray-level co-occurrence matrix, local binary mode, etc.). Contour and boundary information of objects in visible light images can be extracted using edge detection algorithms (such as Canny edge detection).

[0084] For infrared images, temperature features can be extracted by calculating the grayscale or temperature values ​​of the image, thermal information of objects can be obtained by calculating the radiation intensity distribution of the infrared image, or texture information of objects in infrared images can be captured by texture descriptors (such as grayscale co-occurrence matrix, local binary mode, etc.).

[0085] After extracting the feature information, the visible light image and the infrared image are enhanced.

[0086] Specifically, the following general enhancement methods can be used for visible light images and infrared images from which feature information has been extracted:

[0087] Histogram equalization enhances the contrast of image features, making image details clearer.

[0088] Edge enhancement algorithms, such as Gaussian blur, Sobel operator, or Canny edge detection algorithm, can be used to highlight edge information in an image and enhance its contour and texture features.

[0089] For a specific target object, you can selectively enhance specific color channels in the image to highlight the target object or area.

[0090] Applying Gaussian filtering to an image can remove noise and improve image quality.

[0091] Adjusting the color balance in an image ensures even brightness across different color channels, thereby improving the image's color performance and visual effects.

[0092] In addition, the following enhancement methods can be used for infrared images:

[0093] To address the issue of the large dynamic range of infrared images, dynamic range compression can be employed to ensure that details in infrared images are clearly displayed across a wider brightness range.

[0094] Temperature correction is applied to infrared images, converting the original grayscale values ​​into temperature values ​​to enhance the temperature characteristics of the infrared images.

[0095] Infrared images are converted into pseudo-color images, and the contrast and detail of the images are enhanced by selecting an appropriate color mapping scheme.

[0096] Spatial filtering algorithms, such as mean filtering or median filtering, can be used to remove noise from infrared images and improve image quality and clarity.

[0097] Mapping the grayscale values ​​of infrared images to thermal color maps makes the temperature information in the images more intuitive and visual.

[0098] The feature information is extracted from the visible light image and the infrared image using the above feature information extraction method. After the feature information is extracted, the feature information is enhanced by the above feature information enhancement method to obtain the enhanced visible light image and the infrared image.

[0099] Step S300: Perform multi-scale fusion and reconstruction on the enhanced visible light image and infrared image to obtain the reconstructed image.

[0100] Specifically, in this embodiment, since visible light images and infrared images have different imaging advantages, the feature information that can be extracted is also different.

[0101] By performing multi-scale fusion of visible light and infrared images, the feature information of the two images can be integrated, thereby improving the visualization effect and information content of the fused image, and thus enhancing the clarity and recognizability of the target object.

[0102] Specifically, in this embodiment, the visible light image and the infrared image are first fused using a multi-scale fusion method based on an image pyramid.

[0103] Multi-scale fusion refers to the fusion of features at multiple scales in an image. Multi-scale fusion typically employs multiple branches, each processing a single scale feature. Fusing image feature information at multiple scales not only achieves better performance than single-scale feature fusion, but also, compared to fusion at a single scale feature, multi-scale data fusion better characterizes the essential properties of the target object.

[0104] In this embodiment, the multi-scale fusion method for visible light images and infrared images based on image pyramids may include the following steps:

[0105] First, we construct Gaussian and Laplacian pyramids for visible light images and infrared images to represent information from these two types of images at different scales.

[0106] Image pyramids are a multi-scale image representation method, typically divided into Gaussian pyramids and Laplacian pyramids. Gaussian pyramids contain different scale versions of an image, while Laplacian pyramids contain detailed information about the image at different scales.

[0107] Then, these pyramids are fused to obtain a multi-scale fused image. The pyramid fusion process includes the following steps:

[0108] First, the Laplacian pyramids from the visible light image and the infrared image are fused to obtain the fused Laplacian pyramid.

[0109] Then, the merged Laplace pyramid is fused with the Gaussian pyramid of the visible light image to obtain the merged pyramid.

[0110] Finally, the images of different scales in the merged pyramid are recombined into an image of the original size, thus obtaining a reconstructed image that can be displayed.

[0111] Referring to Figure 3, Figure 3 is a schematic diagram of the process of performing image enhancement on visible light images and infrared images respectively, and performing pyramid-based fusion on the enhanced images in this embodiment.

[0112] In this embodiment, visible light and infrared images acquired by an endoscope are obtained. Feature information in these images is extracted and enhanced to obtain enhanced visible light and infrared images. Multi-scale fusion and reconstruction are then performed on these enhanced images to obtain a reconstructed image. This solution, by extracting and enhancing feature information from visible light and infrared images, and then performing multi-scale fusion and reconstruction on the enhanced images, solves the problem that existing imaging technologies, where standalone visible light or infrared imaging is easily affected by environmental factors or high noise levels, cannot overcome the technical problems of image fusion resulting in images with insufficient texture, lack of detail, and image distortion. It achieves the enhancement and multi-scale fusion of visible light and infrared images, generating fused images with rich texture and prominent targets. This can then be applied to endoscopes with high requirements for imaging quality, improving the clarity of endoscopic imaging, thereby increasing surgical efficiency and reducing surgical risks.

[0113] Furthermore, referring to FIG4, the second embodiment of the present invention provides an endoscopic image processing method. Based on the embodiment shown in FIG2 above, step S200: extracting and enhancing feature information from the visible light image and infrared image to obtain enhanced visible light image and infrared image may include:

[0114] Step S201: Decompose the visible light image and the infrared image respectively to obtain their respective base layer and detail layer, wherein the detail layer includes the image feature information extracted by decomposition;

[0115] Specifically, in this embodiment, image filtering is first performed on the visible light image and the infrared image respectively to obtain smoother, more natural, and more realistic base layer images for the visible light image and the infrared image respectively.

[0116] Then, the visible light image and the infrared image are subtracted from their respective base layer images to obtain the detail layer images of the visible light image and the infrared image, respectively.

[0117] By subtracting the base layer images from the visible light and infrared images, various feature information can be extracted from both. This feature information is then contained in the product of the subtraction, which is the detail layer image.

[0118] Specifically, referring to Figure 5, which is a detailed flowchart of step S201, step S201 may include:

[0119] Step S2011: Perform guided filtering processing on the visible light image and the infrared image respectively to generate the base layer of the visible light image and the infrared image;

[0120] Step S2012: Subtract the base layers of the visible light image and the infrared image from their respective base layers to obtain detail layers of the visible light image and the infrared image, respectively. The detail layers contain feature information of the original images extracted by the subtraction.

[0121] More specifically, in this embodiment, the visible light image and the infrared image are first subjected to guided filtering. Guided filtering can make the image clearer and smoother. It can remove some noise while preserving the image edge information, making the image smoother, more natural, and more realistic.

[0122] In this embodiment, both visible light and infrared images can be used as the original images. Guided filtering, including a Gaussian filter, is applied to the original images to obtain the corresponding base layer image. Guided filtering filters only one color channel at a time; after applying guided filtering to each channel, the base layer image is obtained. In this embodiment, guided filtering is performed according to the following formula:

[0123] Base(r,g,b)=guide(I(r,g,b));

[0124] Where r, g, and b are the three color channels of the original image; I(r,g,b) is the original image and Base(r,g,b) is the base layer image.

[0125] Then, the base layer image obtained after guided filtering is subtracted from the original image to obtain the detail layer image. The subtraction is performed using the following formula:

[0126] Detail = I (0-1) (r,g,b)-Base(r,g,b);

[0127] The I here (0-1) (r,g,b), which is the original image, corresponds to the normalized visible light image and infrared image. The normalized image is obtained by normalizing the two images acquired by the visible light sensor and infrared sensor as described above.

[0128] Step S202: Enhance the respective base layer and detail layer to obtain the enhanced base layer and detail layer;

[0129] Specifically, referring to Figure 6, which is a detailed flowchart of step S202, step S202 may include:

[0130] Step S2021: Adaptively enhance the base layer of the visible light image and the infrared image to obtain the enhanced base layer of the visible light image and the infrared image;

[0131] More specifically, in this embodiment, adaptive enhancement is performed on the base layer images of the visible light and infrared images. This aims to automatically adjust the enhancement parameters based on the local features of the base layer images, thereby achieving adaptive enhancement for different regions. This allows for enhancement of features such as brightness and contrast in different areas while preserving details, resulting in better visual effects.

[0132] The adaptive enhancement methods in this embodiment include: Local Contrast Enhancement, Histogram Equalization, Adaptive Histogram Equalization, and Adaptive Enhancement Filters. These methods can enhance different features to achieve better visual effects in different regions. In practice, a suitable adaptive enhancement method can be selected based on specific requirements.

[0133] For example, there might be cases where the obtained base layer image is too dark and needs to be enhanced. However, the brightness level of the image is random. To avoid over-enhancing an image with appropriate brightness, a local adaptive image enhancement method is used. The formula for the adaptive enhancement method is as follows:

[0134] E_base=Base(x,y)^(α*W(x,y)+0.5);

[0135] Where α is the enhancement rate, W(x,y) is the average gray level of the neighborhood (local window) centered at Base(x,y), (x,y) is the horizontal and vertical coordinates of the image, and E_base is the base layer image after adaptive enhancement.

[0136] Step S2022: Perform nonlinear enhancement on the detail layers of the visible light image and the infrared image to obtain the enhanced detail layers of the visible light image and the infrared image.

[0137] Compared to linear enhancement methods, nonlinear enhancement methods can better handle complex textures, edges, and details in images, producing more natural and vivid image effects. To significantly improve image contrast, enhance image details, and improve image visual quality, nonlinear enhancement is chosen. In this embodiment, the nonlinear enhancement method includes:

[0138] Contrast stretching: This enhances image contrast by expanding the dynamic range of pixel values ​​through a non-linear mapping. It makes the darkest pixels darker and the brightest pixels brighter, thus improving the image's visual appeal.

[0139] Histogram equalization: By transforming the grayscale distribution of image pixels, it makes the image's histogram uniformly distributed across the entire grayscale range. This enhances image contrast and makes image details clearer.

[0140] Adaptive Histogram Equalization: An improved method of histogram equalization, this approach divides the image into multiple small regions and performs histogram equalization on each region separately, thereby achieving adaptive enhancement of local areas. This method can better handle local contrast variations in the image and avoids the over-enhancement problem that may be caused by global enhancement in histogram equalization.

[0141] In practice, appropriate nonlinear enhancement methods can be selected and adjusted according to requirements to achieve better image enhancement results.

[0142] For example, a nonlinear enhancement of the detail layer image can be performed by constructing a nonlinear matrix, with the enhancement referencing other pixels within the neighborhood window. The formula for enhancing the detail layer image by constructing a nonlinear matrix is ​​as follows:

[0143]

[0144] Here, u(i,j) represents a local window centered on the detail layer image detail(x,y). s is an adjustment parameter. The local window can be a 5×5 pixel value, a 16×16 pixel value, or any other arbitrary size. M(x,y) is the enhanced detail layer image; Y is an intermediate variable; and ο is the neighborhood window, which can be a filter window of any size, and the aforementioned local window is contained within the filter window.

[0145] Step S203: Combine the base layer and detail layer of the enhanced visible light image and infrared image respectively to obtain the enhanced visible light image and infrared image respectively.

[0146] Specifically, after enhancing the base layer and detail layer images of the visible light and infrared images respectively through the above steps, the base layer and detail layer of the visible light image are combined, and the base layer of the infrared image is combined. Through this combination, enhanced visible light and infrared images are obtained, following the formula:

[0147] I final =E_base + M * Detail;

[0148] Referring to Figure 7, which is a flowchart of the image enhancement method of this embodiment, the method includes decomposing an image into a base layer and a detail layer, and then combining the enhanced base layer and detail layer. The image includes a visible light image and an infrared image.

[0149] In this embodiment, the visible light image and the infrared image are decomposed separately, and the base layer and detail layer of each decomposed image are enhanced using appropriate enhancement methods. The enhanced base layer and detail layer are then combined to obtain the enhanced visible light image and infrared image. By extracting the feature information and detail information of these two images and enhancing this information, the clarity and recognizability of texture details and contour information in endoscopic imaging are improved.

[0150] Furthermore, referring to Figure 8, the third embodiment of the present invention provides an endoscopic image processing method. Based on the embodiment shown in Figure 2 above, after step S100: acquiring the visible light image and infrared image collected by the endoscope, it further includes:

[0151] Step S101: Detect whether the bit depth of the visible light image and the infrared image are consistent;

[0152] Specifically, in this embodiment, the images described below can be visible light images and infrared images. The bit depth of an image is a parameter used to represent the color depth or grayscale level of each pixel. Bit depth determines the number of colors or grayscale levels an image can display; it is also called color depth or grayscale level. Generally, the higher the bit depth, the richer the image details and color gradations, but the image file size will also increase.

[0153] Considering that subsequent feature extraction and enhancement operations on the visible light and infrared images require consistent image specifications, it is necessary to check whether the bit depths of the visible light and infrared images are consistent after acquiring them. If they are inconsistent, appropriate processing is required to ensure that the bit depths of the two images are consistent.

[0154] Step S102: If they are inconsistent, convert the high bit depth image to a low bit depth image to obtain the registered visible light image and infrared image;

[0155] Specifically, in this embodiment, if the bit depths of the visible light image and the infrared image are inconsistent, it is necessary to convert the high-bit-depth image to a low-bit-depth image to register the two images. The conversion method is to reduce the bit depth of the high-bit-depth image to be the same as the bit depth of the low-bit-depth image. Reducing the bit depth of the image can be achieved using the following methods:

[0156] Truncation: This is one of the simplest bit depth reduction methods. During truncation, the pixel values ​​of a high-bit-depth image are directly truncated (or rounded) to values ​​within the range of a low-bit-depth image. For example, the pixel values ​​(0-255) of a 24-bit image are truncated to the pixel values ​​(0-255) of an 8-bit image.

[0157] Quantization: Quantization maps pixel values ​​from a high-bit-depth image to a range of low-bit-depth pixels. This is done by dividing the image into intervals, mapping the pixel values ​​of the high-bit-depth image to corresponding low-bit-depth pixel values. For example, the pixel values ​​(0-65535) of a 16-bit image can be quantized to the pixel values ​​(0-255) of an 8-bit image, with each interval's pixel value mapped to a corresponding value.

[0158] Linear compression is a method for reducing bit depth while preserving the linear relationship between pixel values ​​in the original image. It maps pixel values ​​from a high bit depth image to a low bit depth range through linear scaling, thus preserving the overall contrast and brightness distribution of the image.

[0159] Gamma Transform: The Gamma transform is a non-linear compression method commonly used to convert high bit-depth images into low bit-depth images. By using the Gamma function to non-linearly map the pixel values ​​of a high bit-depth image, the contrast of low-brightness areas can be improved, but some details in high-brightness areas may be lost.

[0160] Color mapping: Color mapping maps colors from a high-bit-depth image to a lower-bit-depth range. By choosing a suitable mapping algorithm, it is possible to map high-bit-depth colors to a lower-bit-depth range while preserving color characteristics.

[0161] Step S103: Normalize the registered visible light image and infrared image respectively to obtain normalized visible light image and infrared image;

[0162] Specifically, in this embodiment, in order to standardize the pixel values ​​of the image to conform to specific statistical properties or numerical ranges, thereby facilitating subsequent image processing or analysis, it is necessary to normalize the visible light image and the infrared image. Image normalization can improve the stability, comparability, and processing efficiency of the image. In this embodiment, the image normalization process can be performed using the following method:

[0163] Linear normalization: Maps the pixel values ​​of an image to a specific range in a linear proportion, such as mapping pixel values ​​to the range of 0 to 1 or -1 to 1.

[0164] Mean-variance normalization: Standardizes the pixel values ​​of an image by applying the mean and variance, so that the mean of the image is 0 and the variance is 1.

[0165] Max-Min Normalization: Maps the pixel values ​​of an image according to the ratio of the maximum and minimum values, so that the pixel values ​​are within a specific range.

[0166] Histogram equalization: By adjusting the histogram distribution of an image, the dynamic range of pixel values ​​is increased, making the image more contrasting and brighter.

[0167] Specifically, in this embodiment, the normalization method can be implemented according to the following formula:

[0168]

[0169] Where I(x,y) is the pixel value at position (x,y) in the visible light image; min(I) and max(I) are the minimum and maximum values ​​of the pixels in the image; I 0-1 This is the normalized image.

[0170] Furthermore, referring to Figure 9, which is a detailed flowchart of step S200, step S200 may include:

[0171] Feature information is extracted and enhanced from the normalized visible light image and infrared image to obtain enhanced visible light image and infrared image.

[0172] In step S200 above, feature information is extracted and enhanced from the visible light image and the infrared image. Here, the visible light image and the infrared image refer to the two images after normalization processing, that is, the normalized visible light image and the infrared image.

[0173] In this embodiment, image registration is achieved by detecting whether the bit depths of the visible light image and the infrared image are consistent. If they are inconsistent, the higher bit depth image is converted to a lower bit depth image. The registered images are then normalized to eliminate the size differences between the images, making them more suitable for subsequent image processing tasks.

[0174] Furthermore, referring to Figure 10, the fourth embodiment of the present invention proposes an endoscopic image processing method. Based on the embodiment shown in Figure 2 above, step S300: performing multi-scale fusion and reconstruction on the enhanced visible light image and infrared image to obtain the reconstructed image may include:

[0175] Step S301: Generate corresponding Gaussian pyramids for the enhanced visible light image and infrared image respectively, wherein the Gaussian pyramids include images at different scales;

[0176] Step S302: Calculate the Laplacian pyramid for the image in the Gaussian pyramid, wherein the Laplacian pyramid includes image detail information at different scales;

[0177] Specifically, in this embodiment, the Gaussian pyramid and Laplacian pyramid mentioned in steps S301 and S302 both belong to the image pyramid. The image pyramid is a multi-scale image representation method used to process and analyze images at different scales.

[0178] The Gaussian pyramid contains versions of an image at different scales, while the Laplacian pyramid contains detailed information about the image at different scales. From high to low levels, each level is a scaled-down version of the image at the level above. The name "pyramid" comes from its structure; the images arranged by scale resemble a pyramid.

[0179] The advantage of applying image pyramids lies in their ability to perform multi-scale image analysis, extracting different information from different image levels. Furthermore, applying image pyramids to perform multi-scale image fusion results in a fused image that can include detailed and feature information from different image levels.

[0180] Step S303: Weighted fusion of the corresponding levels of the Laplacian pyramids of the enhanced visible light image and infrared image to obtain the fused Laplacian pyramid;

[0181] Specifically, referring to Figure 11, which is a schematic flowchart of the Laplacian pyramid fusion process in this embodiment, the weighted fusion of corresponding levels of the Laplacian pyramids of the enhanced visible light image and infrared image to obtain the fused Laplacian pyramid includes the following steps:

[0182] First, based on the Laplacian pyramids corresponding to the visible light image and the infrared image respectively, the images of the same level, such as their respective third-level images, are multiplied by a corresponding weight.

[0183] Then, by adding the results of the multiplication together, we obtain the image of the corresponding layer of the merged Laplace pyramid (the third layer as described above).

[0184] By fusing the two images at each corresponding level of the two pyramids using the above steps, a fused Laplace Pyramid can be obtained.

[0185] In this embodiment, to ensure that images at higher scales are more significantly affected, the weights use a Gaussian distribution function W. white and W infrerad The formula for calculating the fusion weights of the infrared and visible light images for each layer of the Laplacian pyramid is as follows:

[0186] W white =w white / (w white +w infrerad ), W white The weights are the corresponding visible light images.

[0187] W infrerad =w infrerad / (w white +w infrerad ), W infrerad The weights are the corresponding infrared images.

[0188] Where w white and w infrerad It can be obtained through the following formula:

[0189]

[0190]

[0191] Where σ is the fusion weight adjustment parameter, and ε is the fusion weight adjustment parameter, with a value range of 0-1. white For the white light images to be fused, I imfrared For the infrared images that need to be fused.

[0192] Step S304: Add the merged Laplace pyramid image to the corresponding level image of the Gaussian pyramid to obtain the reconstructed image.

[0193] Specifically, in this embodiment, the images at each level of the fused Laplacian pyramid contain detailed information at different scales of visible light and infrared images. However, an image pyramid is a collection of images, and the ultimate goal of image processing is to obtain a single image. Therefore, image reconstruction is also required on the fused Laplacian pyramid to obtain a reconstructed single image. The process of reconstructing the fused Laplacian pyramid includes the following steps:

[0194] First, the merged Laplacian pyramid is added to the corresponding levels of the Gaussian pyramid in the visible light image, i.e., layer 0 is added to layer 0, layer 1 is added to layer 1, layer n is added to layer n, and so on, until the top layer is reached, to obtain the merged pyramid.

[0195] Then, starting from the top image of the fused pyramid, the images of different levels in the fused pyramid are magnified to the size of the 0th level image, which is the size of the original image (the original image can be a visible light image or an infrared image), by using image upsampling.

[0196] Finally, the images of all levels of the merged pyramid are upsampled and enlarged to the original image size, then all images are superimposed, and the mean value of each pixel is calculated. The resulting mean image is the final reconstructed image.

[0197] In this embodiment, Gaussian pyramids and Laplacian pyramids are constructed for both the visible light image and the infrared image. The Laplacian pyramids of both images are then fused to obtain a fused Laplacian pyramid. This fused Laplacian pyramid is then fused with the Gaussian pyramid of the visible light image to obtain a fused pyramid. The images in the fused pyramid are then added and fused layer by layer to obtain the reconstructed image. The reconstructed image integrates the details and texture information of both the visible light and infrared images at different scales, resulting in clearer endoscopic imaging and more prominent texture details and contours.

[0198] Furthermore, referring to Figure 12, the fifth embodiment of the present invention proposes an endoscopic image processing method. Based on the embodiment shown in Figure 10 above, step S301: generating corresponding Gaussian pyramids for the enhanced visible light image and infrared image respectively, wherein the Gaussian pyramids include images at different scales, which may include:

[0199] Step S3011: The enhanced visible light image and infrared image are downsampled and processed by Gaussian filter to obtain the next level image;

[0200] Step S3012: Repeat the above steps for the next level image to generate a set of images corresponding to the visible light image and the infrared image;

[0201] Step S3013: Arrange the images in the set according to their scale to form a corresponding Gaussian pyramid, and the downsampling is used to reduce the scale of the images.

[0202] Specifically, in this embodiment, the process of constructing a Gaussian pyramid by repeatedly downsampling and blurring the original image is as follows:

[0203] First, starting with the original image, which is level 0 of the Gaussian pyramid, also known as the bottom layer, the original image is downsampled. This can be achieved by using filters or simply discarding some pixels, resulting in a downsampled image that is half the size of the original image. This downsampled image serves as the next level of the Gaussian pyramid.

[0204] Then, a Gaussian filter is applied to the downsampled image for linear blurring to remove high-frequency details and make the image smoother. The purpose of applying Gaussian filtering is to simulate the response of the visual system to remove noise and details, thereby reducing potential aliasing.

[0205] Repeat the downsampling and Gaussian filtering steps for the next level of images until the top layer of the pyramid (the smallest image) is reached, resulting in a series of images at different scales. Each downsampling and Gaussian filtering operation produces an image reduced by half, forming the next level of the pyramid. These images, arranged from smallest to largest and from top to bottom, form a pyramid-shaped image set, which is the Gaussian pyramid.

[0206] In this embodiment, the original image can be an enhanced visible light image or an enhanced infrared image. In particular, before constructing the Gaussian pyramid of the infrared image, the black and white infrared image needs to be copied and expanded into a three-channel format.

[0207] After constructing the Gaussian pyramid, each level of the pyramid represents a version of the original image at a different scale. The bottom layer of the pyramid is the original image, and as the levels increase, the image size decreases while the image details become increasingly blurred. The scale relationship between each level is achieved through a combination of downsampling and Gaussian filtering.

[0208] In this embodiment, a Gaussian pyramid was constructed by repeatedly downsampling and blurring the enhanced visible light image and the enhanced infrared image. This represented the feature information of these two images at different scales and provided a foundation for the subsequent construction of the corresponding Laplacian pyramid.

[0209] Furthermore, referring to Figure 13, the sixth embodiment of the present invention proposes an endoscopic image processing method. Based on the embodiment shown in Figure 10 above, step S302: calculating the Laplacian pyramid for the image in the Gaussian pyramid, wherein the Laplacian pyramid includes image detail information at different scales, including:

[0210] Step S3021: Subtract the images in the Gaussian pyramid corresponding to the visible light image and the infrared image from their respective previous level images to obtain a set of images corresponding to each;

[0211] Step S3022: Arrange the images in the image set according to their scale to form a corresponding Laplacian pyramid. Each layer of the Laplacian pyramid contains detailed information of the corresponding layer of the Gaussian pyramid, and the detailed information is obtained by subtraction.

[0212] Referring to Figure 14, which is a schematic diagram of the process of constructing the Placus pyramid using the Gaussian image pyramid in this embodiment.

[0213] Specifically, in this embodiment, the Laplacian Pyramid is constructed based on the Gaussian Pyramid. The construction process is as follows:

[0214] First, the images at each level of the Gaussian pyramid are upsampled. Upsampling can be achieved using interpolation methods, including nearest neighbor interpolation, bilinear interpolation, and bicubic interpolation. After upsampling, the image in the Gaussian pyramid has the same size as the image at the next higher level. For example, the image size after upsampling the first level image is the same as the size of the image at level 0 in the Gaussian pyramid.

[0215] Then, the upsampled image is processed by Gaussian filtering to smooth the image and remove any high-frequency noise that may be introduced.

[0216] Finally, the image of the corresponding level in the Gaussian pyramid is subtracted from the upsampled image of the next level, resulting in the image of the corresponding level in the Laplacian pyramid. For example, subtracting the upsampled image of the first level from the image of the 0th level in the Gaussian pyramid yields the image of the 0th level in the Laplacian pyramid. This operation removes low-frequency information from the Gaussian image while preserving the detail information of the original image at different scales.

[0217] Repeating the above steps multiple times will yield a collection of images. Arranging these images from smallest to largest scale, and from top to bottom, will create a pyramid-shaped image set, which is the Laplace pyramid.

[0218] In this embodiment, the Laplacian pyramid of the visible light image is obtained by processing the Gaussian pyramid of the visible light image through the above steps, and the Laplacian pyramid of the infrared image is obtained by processing the Gaussian pyramid of the infrared image through the above steps.

[0219] In this embodiment, a Laplacian pyramid is constructed by progressively extracting detail information from the original image (which may be an enhanced visible light image or an enhanced infrared image) at different scales. Each level of the Laplacian pyramid represents the detail information of the original image at that scale. This detail information can be reconstructed by superimposing the Laplacian image back into the corresponding Gaussian image level, thereby obtaining the reconstructed image.

[0220] Furthermore, referring to FIG15, this embodiment of the invention also provides an endoscopic image processing apparatus 150, the apparatus comprising:

[0221] The endoscope image acquisition module 151 is used to acquire visible light images and infrared images captured by the endoscope;

[0222] The endoscope image enhancement module 152 extracts and enhances the feature information in the visible light image and infrared image to obtain the enhanced visible light image and infrared image.

[0223] The endoscopic image fusion module 153 is used to fuse the enhanced visible light image and the enhanced infrared image to obtain a reconstructed image.

[0224] Furthermore, embodiments of the present invention also provide an endoscopic image processing device, the device comprising: a memory, a processor, and an endoscopic image processing program stored in the memory and executable on the processor, the endoscopic image processing program being configured to implement the steps of the various embodiments of the endoscopic image processing method described above.

[0225] Furthermore, the present invention also provides a computer-readable storage medium storing an endoscope image processing program, which, when executed by a processor, implements the steps of the various embodiments of the endoscope image processing method described above.

[0226] In the embodiments of the apparatus, terminal equipment, and computer-readable storage medium of the present invention, all the technical features of the above-described embodiments of the image feature encoding method are included. The extended and explanatory contents of the specification are basically the same as those of the above-described embodiments of the image feature encoding method, and will not be repeated here.

[0227] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.

[0228] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0229] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0230] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.

Claims

1. An endoscopic image processing method, characterized in that, The endoscopic image processing method includes the following steps: acquiring visible light and infrared images obtained by an endoscope; extracting and enhancing feature information from the visible light and infrared images to obtain enhanced visible light and infrared images; generating corresponding Gaussian pyramids for the enhanced visible light and infrared images, wherein the Gaussian pyramids include images at different scales; calculating a Laplacian pyramid for the images in the Gaussian pyramids, wherein the Laplacian pyramids include image detail information at different scales; weightedly fusing the corresponding levels of the Laplacian pyramids of the enhanced visible light and infrared images to obtain a fused Laplacian pyramid; and adding the fused Laplacian pyramid to the corresponding level of the Gaussian pyramid of the visible light image to obtain the reconstructed image.

2. The endoscopic image processing method as described in claim 1, characterized in that, The step of extracting and enhancing the feature information in the visible light image and infrared image to obtain enhanced visible light image and infrared image includes: decomposing the visible light image and infrared image respectively to obtain their respective base layer and detail layer, wherein the detail layer includes the image feature information extracted from the decomposition; enhancing the respective base layer and detail layer respectively to obtain enhanced base layer and detail layer; and combining the enhanced base layer and detail layer of the visible light image and infrared image respectively to obtain enhanced visible light image and infrared image.

3. The endoscopic image processing method as described in claim 2, characterized in that, The step of decomposing the visible light image and the infrared image to obtain their respective base layer and detail layer includes: performing guided filtering on the visible light image and the infrared image respectively to generate the base layer of the visible light image and the infrared image; subtracting the respective base layer from the original images of the visible light image and the infrared image to obtain the detail layer of the visible light image and the infrared image respectively, wherein the detail layer contains the feature information of the original image extracted by the subtraction.

4. The endoscopic image processing method as described in claim 2, characterized in that, The step of enhancing the respective base layer and detail layer to obtain the enhanced base layer and detail layer includes: adaptively enhancing the base layer of the visible light image and the infrared image to obtain the enhanced base layer of the visible light image and the infrared image; and nonlinearly enhancing the detail layer of the visible light image and the infrared image to obtain the enhanced detail layer of the visible light image and the infrared image.

5. The endoscopic image processing method as described in claim 1, characterized in that, The steps following the acquisition of visible light and infrared images from the endoscope further include: detecting whether the bit depths of the visible light and infrared images are consistent; if they are inconsistent, converting the high bit depth image to a low bit depth image to obtain registered visible light and infrared images; normalizing the registered visible light and infrared images respectively to obtain normalized visible light and infrared images; and extracting and enhancing the feature information in the visible light and infrared images to obtain enhanced visible light and infrared images, including: extracting and enhancing the feature information in the normalized visible light and infrared images to obtain enhanced visible light and infrared images.

6. The endoscopic image processing method as described in claim 1, characterized in that, The step of generating corresponding Gaussian pyramids for the enhanced visible light image and infrared image, wherein the Gaussian pyramid includes images at different scales, includes: performing downsampling and Gaussian filter processing on the enhanced visible light image and infrared image respectively to obtain the next level image; repeating the above steps on the next level image to generate a set of a series of images corresponding to the visible light image and infrared image; arranging the images in the set according to scale to form the corresponding Gaussian pyramid, wherein the downsampling is used to reduce the scale of the image.

7. The endoscopic image processing method as described in claim 1, characterized in that, The step of calculating the Laplacian pyramid for the images in the Gaussian pyramid, wherein the Laplacian pyramid includes image detail information at different scales, includes: subtracting the images in the Gaussian pyramid corresponding to the visible light image and the infrared image from their respective previous level images to obtain a set of images corresponding to each; arranging the images in the set of images according to scale to form the corresponding Laplacian pyramid, wherein each layer of the Laplacian pyramid contains detail information of the corresponding layer of the Gaussian pyramid, and the detail information is obtained by subtraction.

8. An endoscopic image processing device, characterized in that, The device includes: an endoscope image acquisition module for acquiring visible light and infrared images captured by an endoscope; an endoscope image enhancement module for extracting and enhancing feature information from the visible light and infrared images to obtain enhanced visible light and infrared images; an endoscope image fusion module for generating corresponding Gaussian pyramids for the enhanced visible light and infrared images, wherein the Gaussian pyramids include images at different scales; calculating a Laplacian pyramid for the images in the Gaussian pyramids, wherein the Laplacian pyramids include image detail information at different scales; weightedly fusing the corresponding levels of the Laplacian pyramids of the enhanced visible light and infrared images to obtain a fused Laplacian pyramid; and adding the fused Laplacian pyramid to the corresponding level of the Gaussian pyramid of the visible light image to obtain the reconstructed image.

9. An endoscopic image processing device, characterized in that, The device includes: a memory, a processor, and an endoscope image processing program stored in the memory and executable on the processor, the endoscope image processing program being configured to implement the endoscope image processing method as described in any one of claims 1 to 7.

10. A storage medium, characterized in that, The storage medium stores an endoscope image processing program, which, when executed by a processor, implements the endoscope image processing method as described in any one of claims 1 to 7.

Images