Image processing method, image processing device and electronic equipment

Abstract

The invention discloses an image processing method, an image processing device and electronic equipment, and the method comprises the steps: obtaining a target image, the target image is obtained through rendering a plurality of Gaussian bodies in a target model, and the target image comprises a first number of image pixels; based on the color contribution degree of the Gaussian body to each image pixel, a target pixel set is determined, target pixels in the target pixel set correspond to the same target Gaussian body, and the color contribution degree of the target Gaussian body to each target pixel meets a first preset condition; and determining whether the target image is blurred based on the ratio of the second number of the target pixels to the first number.

Description

Technical Field

[0001] The present application relates to the technical field of image processing, and in particular to an image processing method, an image processing device, and an electronic device. Background Art

[0002] The 3D Gaussian model can render two-dimensional images corresponding to multiple perspectives of the scene. However, the rendered two-dimensional images are prone to blurring. In related technologies, manual detection is used to determine whether the rendered image is blurred, which is inefficient. Summary of the Invention

[0003] Embodiments of the present application provide an image processing method, an image processing device, and an electronic device.

[0004] In one aspect, an embodiment of the present application provides an image processing method, comprising:

[0005] Acquire a target image, where the target image is rendered by a plurality of Gaussian volumes in a target model, and the target image includes a first number of image pixels;

[0006] determining a target pixel set based on a color contribution of the Gaussian body to each of the image pixels, wherein target pixels in the target pixel set correspond to the same target Gaussian body, and a color contribution of the target Gaussian body to each of the target pixels satisfies a first preset condition;

[0007] Whether the target image is blurred is determined based on a ratio of a second number of the target pixels to the first number.

[0008] Optionally, the method further includes:

[0009] For a first pixel in the image pixels, determining a first Gaussian body and a second Gaussian body corresponding to the first pixel, wherein the second Gaussian body is closer to a viewing angle position of the target image than the first Gaussian body;

[0010] A color contribution of the first Gaussian body to the first pixel is determined based on the opacity of the first Gaussian body relative to the first pixel and the opacity of the second Gaussian body relative to the first pixel.

[0011] Optionally, the method further includes:

[0012] For a first pixel in the image pixels, determining color contributions of a plurality of candidate Gaussian bodies corresponding to the first pixel relative to the first pixel;

[0013] The Gaussian with the highest color contribution among the multiple candidate Gaussian bodies is determined as the target Gaussian body for the first pixel.

[0014] Optionally, determining whether the target image is blurred based on a ratio of the second number of target pixels to the first number includes:

[0015] In a case where a ratio of the second number of target pixels to the first number is greater than a target threshold, it is determined that the target image is blurred.

[0016] Optionally, the method further includes:

[0017] determining a target image area in the target image according to distribution information of the target pixels in the target image;

[0018] The target image area is determined to be a blurred area in the target image.

[0019] Optionally, the target image is a first image, and the method further includes:

[0020] Determining a second image, where the second image is an image obtained by rendering the target model, and the second image is a blurred image;

[0021] In the case that the first image is blurred, a target model region in the target model is determined according to the first image and the second image, where the target model region represents a region in the target model that does not meet training requirements.

[0022] Optionally, determining a target model area in the target model according to the first image and the second image includes:

[0023] determining a target image area in the first image according to distribution information of the target pixels in the first image;

[0024] A target model area in the target model is determined according to the target image area and the second image.

[0025] Optionally, the method further includes:

[0026] receiving rendering task information, wherein the rendering task information represents rendering an image corresponding to a target perspective through the target model;

[0027] In response to the area corresponding to the target perspective including at least a portion of the target model area, an image corresponding to the target perspective is obtained by a method other than a target method, where the target method is to render the image corresponding to the target perspective by using a target model.

[0028] The present application also provides an image processing device, including:

[0029] an acquisition module, configured to acquire a target image, wherein the target image is rendered by a plurality of Gaussian volumes in a target model, and the target image includes a first number of image pixels;

[0030] a processing module, configured to determine a target pixel set based on a color contribution of the Gaussian body to each of the image pixels, wherein the target pixels in the target pixel set correspond to the same target Gaussian body, and the color contribution of the target Gaussian body to each of the target pixels satisfies a first preset condition;

[0031] A determination module is configured to determine whether the target image is blurred based on a ratio of a second number of target pixels to the first number of target pixels.

[0032] On the other hand, an embodiment of the present application further provides an electronic device, including a memory and a processor, wherein the memory stores an executable program, and the processor executes the executable program to implement the steps of the above method. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] Figure 1 This is a flowchart of an image processing method according to an embodiment of the present application;

[0034] Figure 2 Another flowchart of the image processing method according to an embodiment of the present application;

[0035] Figure 3 is another flow chart of the image processing method according to an embodiment of the present application;

[0036] Figure 4 is another flow chart of the image processing method according to an embodiment of the present application;

[0037] Figure 5 is another flow chart of the image processing method according to an embodiment of the present application;

[0038] Figure 6 A flowchart of an embodiment of an image processing method according to an embodiment of the present application;

[0039] Figure 7 is another flow chart of the image processing method according to an embodiment of the present application;

[0040] Figure 8 This is a structural block diagram of an image processing device according to an embodiment of the present application;

[0041] Figure 9 This is a structural block diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0042] Various aspects and features of the present application are described herein with reference to the accompanying drawings.

[0043] It should be understood that various modifications may be made to the embodiments of the present application. Therefore, the above description should not be considered as limiting, but merely as an example of an embodiment. Other modifications within the scope and spirit of the present application will occur to those skilled in the art.

[0044] The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate embodiments of the present application and, together with the general description of the present application given above and the detailed description of the embodiments given below, serve to explain the principles of the present application.

[0045] These and other characteristics of the present application will become apparent from the following description of a preferred form of embodiment given as a non-limiting example with reference to the accompanying drawings.

[0046] It should also be understood that although the present application has been described with reference to certain specific examples, those skilled in the art will readily be able to implement many other equivalent forms of the present application.

[0047] The above and other aspects, features and advantages of the present application will become more apparent from the following detailed description when taken in conjunction with the accompanying drawings.

[0048] Specific embodiments of the present application will be described hereinafter with reference to the accompanying drawings; however, it should be understood that the embodiments described are merely examples of the present application and may be implemented in a variety of ways. Familiar and / or repetitive functions and structures are not described in detail to avoid obscuring the present application with unnecessary or redundant details. Therefore, the specific structural and functional details described herein are not intended to be limiting, but rather serve merely as a basis and representative basis for the claims to teach those skilled in the art to variously utilize the present application with substantially any suitable detailed structure.

[0049] This specification may use the phrases "in one embodiment," "in another embodiment," "in yet another embodiment," or "in other embodiments," which may all refer to one or more of the same or different embodiments according to the present application.

[0050] Figure 1 1 shows a flow chart of an image processing method according to an embodiment of the present application. Figure 1 Shown, including:

[0051] S100, obtaining a target image, where the target image is rendered by a plurality of Gaussian volumes in a target model, and the target image includes a first number of image pixels;

[0052] The image processing method of this embodiment is applied to three-dimensional Gaussian rendering, and the target model may be a three-dimensional Gaussian model. For example, a 2D image of a scene may be captured by a camera, and the initial model may be trained using the 2D image to obtain a three-dimensional Gaussian model, which includes multiple Gaussian bodies. Given a viewing angle position, multiple Gaussian bodies in the target model are rendered to obtain a target image. The target image includes a first number of image pixels, and the image pixels may be pixel blocks that constitute the target image. Specifically, the colors corresponding to the multiple Gaussian bodies may be weighted summed up at a selected viewing angle position to obtain the target image. The weights corresponding to different Gaussian bodies represent different colors. The colors of the image pixels contained in the target image are composed of the weights corresponding to the multiple Gaussian bodies. The multiple Gaussian bodies in the target model have different color contributions to the image pixels contained in the target image.

[0053] S200, determining a target pixel set based on a color contribution of the Gaussian body to each of the image pixels, wherein target pixels in the target pixel set correspond to the same target Gaussian body, and a color contribution of the target Gaussian body to each of the target pixels satisfies a first preset condition;

[0054] In this embodiment, after rendering multiple Gaussian bodies in a target model to obtain a target image, the color contribution of each Gaussian body to each image pixel included in the target image is determined. Based on the color contribution of each Gaussian body to each image pixel, a target pixel set is determined. The target pixel set includes a second number of target pixels. The target pixel set corresponds to the image region formed by the target Gaussian body. The target pixels may be a pixel block included in the target pixel set. The target pixels in the target pixel set correspond to the same target Gaussian body, and the color contribution of the target Gaussian body to each target pixel satisfies a first preset condition. The first preset condition indicates that the color contribution of the target Gaussian body to each target pixel is greater than the color contributions of other Gaussian bodies to the target pixel, or that the color contribution of the target Gaussian body to each target pixel is the largest among the multiple Gaussian bodies. The number of target Gaussian bodies may be one or more. For example, if there are two target Gaussian bodies, the color contribution of both target Gaussian bodies to the target pixel is greater than that of the other Gaussian bodies. If there is only one target Gaussian body, the color contribution of the one target Gaussian body to the target pixel is the largest among the multiple Gaussian bodies.

[0055] S300: Determine whether the target image is blurred based on a ratio of a second number of target pixels to the first number of target pixels.

[0056] In this embodiment, a target pixel set is determined based on the color contribution of the Gaussian to each of the image pixels, and after the target Gaussian is determined, a ratio of the second number of target pixels to the first number of image pixels can be determined based on the second number of target pixels and the first number of image pixels; and based on the ratio of the second number of target pixels to the first number of image pixels, it is determined whether the target image is blurred. If the value of the ratio of the second number of target pixels to the first number of image pixels is small, it indicates that the weight of the target Gaussian is small, that is, the target model does not contain many Gaussians with high color contributions, and the training method for characterizing the Gaussian does not cause obvious blurred areas in the target image; if the value of the ratio of the second number of target pixels to the first number of image pixels is large, it indicates that the weight of the target Gaussian is large, that is, the target model contains many Gaussians with high color contributions, and the training method for characterizing the Gaussian causes obvious blurred areas in the target image.

[0057] The present application determines a target pixel set based on the color contribution of the Gaussian body to each image pixel of the target image in the above manner, and determines whether the target image is blurred based on the ratio of the second number of target pixels in the target pixel set to the first number of image pixels of the target image. The present application determines whether the target model contains a Gaussian body with a high color contribution by the ratio of the second number of target pixels to the first number of image pixels, and then determines whether the target image is blurred. This does not require data collection to train a classification model, simplifies the method of determining blurred areas in three-dimensional Gaussian scenes, and improves the user experience.

[0058] In one embodiment of the present application, Figure 2 As shown, based on Example 1, the method further includes:

[0059] S210: For a first pixel in the image pixels, determine a first Gaussian body and a second Gaussian body corresponding to the first pixel, wherein the second Gaussian body is closer to a viewing angle position of the target image than the first Gaussian body;

[0060] In this embodiment, the process of determining the color contribution of a Gaussian volume to an image pixel can determine the color contribution of each Gaussian volume to the first pixel in the image pixels, where the first pixel is a pixel point included in the image pixel. Specifically, first, given a viewing angle of a target image, a viewing cone is generated, and the Gaussian volumes within the viewing cone are filtered. The Gaussian volumes within the viewing cone are sorted from near to far according to the viewing angle of the target image. All Gaussian volumes are projected onto a two-dimensional image plane, and alpha blending is performed in the sorted order from near to far to obtain the rendered target image. For the first pixel included in the image pixel of the target image, any Gaussian volume corresponding to the first pixel can be determined as the first Gaussian volume, and the Gaussian volume at a viewing angle closer to the target image relative to the first Gaussian volume can be determined as the second Gaussian volume. Then, the opacity of the first Gaussian volume relative to the first pixel is determined, as well as the opacity of the second Gaussian volume relative to the first pixel.

[0061] S220 : Determine a color contribution of the first Gaussian body to the first pixel based on the opacity of the first Gaussian body relative to the first pixel and the opacity of the second Gaussian body relative to the first pixel.

[0062] In this embodiment, after determining a first Gaussian volume corresponding to a first pixel and a second Gaussian volume located closer to the target image's viewing angle relative to the first Gaussian volume, the color contribution of the first Gaussian volume to the first pixel is determined based on the opacity of the first Gaussian volume relative to the first pixel and the opacity of the second Gaussian volume relative to the first pixel. Specifically, the color of the first pixel is determined based on the color and opacity of the first Gaussian volume and the opacity of the second Gaussian volume, and then the color contribution of the first Gaussian volume to the first pixel is determined.

[0063] For example, the first pixel is represented as the i-th pixel in the target image, and the color of the i-th pixel is calculated by the following formula:

[0064] C i =∑ n≤N c n α n ·T n ,T n =∏ m＜n (1-α m )

[0065] Among them, c n is the color of the first Gaussian body, which can be expressed as the color of the nth Gaussian body in the above formula; α n is the opacity of the first Gaussian body, which can be expressed as the opacity of the nth Gaussian body in the above formula; T nis the opacity of the second Gaussian body, which can be expressed as the cumulative opacity of the n-1 Gaussian bodies before the nth Gaussian body in the above formula; N is the number of Gaussian bodies; α m is the opacity of each second Gaussian body, and m represents the number of second Gaussians.

[0066] α n ·T n In the above formula, it can be expressed as the color contribution of the nth Gaussian to the i-th pixel on the rendering target image, which can be specifically expressed as the following formula:

[0067]

[0068] Through the above method, the color contribution of each Gaussian body to the first pixel can be obtained.

[0069] In some embodiments, the color contribution of each Gaussian body to the first pixel may be determined according to the correlation between the color of the first pixel and the color parameters in each Gaussian body.

[0070] In one embodiment of the present application, Figure 3 As shown, based on Example 1, the method further includes:

[0071] S230: For a first pixel in the image pixels, determine color contributions of a plurality of candidate Gaussian bodies corresponding to the first pixel relative to the first pixel;

[0072] S240 : Determine a Gaussian body with the highest color contribution among the multiple candidate Gaussian bodies as a target Gaussian body for the first pixel.

[0073] In this embodiment, after determining the color contribution of each Gaussian body relative to a first pixel in the image, multiple candidate Gaussian bodies corresponding to the first pixel can be determined for each Gaussian body, and the color contributions of each of the multiple candidate Gaussian bodies corresponding to the first pixel relative to the first pixel can be determined. The candidate Gaussian body can be a Gaussian body with a greater color contribution to the first pixel, that is, a Gaussian body representing a more pronounced color effect for the first pixel in the image. After determining the multiple candidate Gaussian bodies corresponding to each first pixel, a selection can be made from the multiple candidate Gaussian bodies. After determining the color contributions of each of the multiple candidate Gaussian bodies corresponding to the first pixel relative to the first pixel, the Gaussian body with the highest color contribution among the multiple candidate Gaussian bodies can be determined as a target Gaussian body. The target Gaussian body represents the most pronounced color effect for the first pixel in the image. For example, for each first pixel in the image, four candidate Gaussian bodies with greater color contributions to the first pixel can be determined. One of the four candidate Gaussian bodies has the highest color contribution, and this candidate Gaussian body can be determined as the target Gaussian body. In some embodiments, the top N Gaussian bodies ranked by color contribution among the multiple candidate Gaussian bodies may be used as target Gaussian bodies, and the target pixels in the target pixel set correspond to the same N target Gaussian bodies, where N is an integer greater than 1.

[0074] In one embodiment of the present application, determining whether the target image is blurred based on the ratio of the second number of target pixels to the first number includes:

[0075] In a case where a ratio of the second number of target pixels to the first number is greater than a target threshold, it is determined that the target image is blurred.

[0076] In this embodiment, the ratio of the second number of target pixels to the first number is compared with a target threshold to determine whether the target image is blurred. The target threshold may be a specified ratio threshold of the number of pixels in the target image, and may be set based on the number of image pixels included in the target image. If the ratio of the second number of target pixels to the first number is not greater than the target threshold, it is determined that the target image is not blurred, there is no problem with the training method for representing the Gaussian volume, and no blurred area has appeared in the target image. If the ratio of the second number of target pixels to the first number is greater than the target threshold, it is determined that the target image is blurred, there is a problem with the training method for representing the Gaussian volume, and a blurred area has appeared in the target image. For example, if the number of image pixels included in the target image is 10,000, the target threshold is set to 0.1 based on the number of image pixels in the target image, the second number of target pixels is 1,200, and the ratio of the second number of target pixels to the first number is 0.12. After comparing the ratio of the second number of target pixels to the first number with the target threshold, it is determined that the ratio of the second number of target pixels to the first number is greater than the target threshold, the target image is blurred, and there is a problem with the training method for representing the Gaussian volume, which has caused a blurred area in the target image.

[0077] In one embodiment of the present application, Figure 4 As shown, the method further includes:

[0078] S400, determining a target image area in the target image according to distribution information of the target pixels in the target image;

[0079] S500: Determine that the target image area is a blurred area in the target image.

[0080] In this embodiment, after determining the target Gaussian body, the distribution information of the target pixels corresponding to the target Gaussian body in the target image can be determined. The distribution information of the target pixels in the target image can be the specific position of the target pixels in the target image; after determining the distribution information of each target pixel in the target image, the target image area in the target image can be determined based on the distribution information of each target pixel in the target image, and the target image area is composed of multiple target pixels. After determining the target image area in the target image, the target image area can be determined as a blurred area in the target image. If the proportion of target pixels corresponding to the same target Gaussian body in a region is larger, that is, the denser the distribution, the blurred image representing the region is. Therefore, the blurred area can be further determined from the target image based on the distribution information of the target pixels in the target image. Furthermore, image processing can be performed on the blurred area, thereby improving the clarity of the target image.

[0081] In one embodiment of the present application, Figure 5As shown, based on Example 5, the method further includes: the target image is a first image, and the method further includes:

[0082] S410, determining a second image, where the second image is an image obtained by rendering the target model, and the second image is a blurred image;

[0083] S420: When the first image is blurred, determine a target model region in the target model according to the first image and the second image, where the target model region represents a region in the target model that does not meet training requirements.

[0084] In this embodiment, after rendering multiple Gaussian volumes using a target model to obtain a target image, i.e., a first image, a different viewing angle can be given and multiple Gaussian volumes in the target model can be rendered to obtain a second image different from the first image. The multiple Gaussian volumes used to render the second image include the target Gaussian volume. Due to the high color contribution of the target Gaussian volume, the second image rendered by the target model is blurred. After rendering the blurred second image using the target model, if the target image, i.e., the first image, is determined to be blurred, a target model region in the target model can be determined based on the first and second images. Specifically, the viewing frustum corresponding to the first image and the viewing frustum corresponding to the second image can be determined. The viewing frustum corresponding to the first image and the viewing frustum corresponding to the second image can be superimposed to obtain the target model region in the target model, i.e., the three-dimensional blurred region in the three-dimensional Gaussian model. The target model region represents the region of the target model that does not meet the training requirements. That is, the image rendered using the target model will appear blurred. The blurred portion of the image corresponds to the target model region. The viewing frustum corresponding to the first image and the viewing frustum corresponding to the second image have an overlapping area. The viewing frustum corresponding to the first image is the viewing frustum corresponding to the perspective used to render the first image. The viewing frustum corresponding to the second image is the viewing frustum corresponding to the perspective used to render the second image. The first perspective used to render the first image is different from the second perspective used to render the second image. Furthermore, when the first image, the second image, and the plurality of third images are all blurred images, the target model region in the target model can be determined based on the first image, the second image, and the plurality of third images, and the first image, the second image, and the third image are all images rendered as the target model.

[0085] In one embodiment of the present application, Figure 6 As shown, based on Example 5, determining the target model area in the target model according to the first image and the second image includes:

[0086] S4201, determining a target image area in the first image according to distribution information of the target pixels in the first image;

[0087] S4202: Determine a target model area in the target model according to the target image area and the second image.

[0088] In this embodiment, when the first image is blurred, a target Gaussian volume corresponding to a target pixel in the first image can be determined. Then, distribution information of the target pixels corresponding to the target Gaussian volume in the first image can be determined. The distribution information of the target pixels in the first image can be the specific location of the target pixels in the first image. After determining the distribution information of each target pixel in the first image, a target image region in the first image can be determined based on the distribution information of each target pixel in the first image. The target image region is composed of each target pixel. The target image region is the blurred region in the first image and is the region where the blur effect is concentrated in the first image. After determining the target image region in the first image, a target model region in the target model can be determined based on the target image region and the second image. Specifically, a viewing frustum corresponding to the target image region in the first image and a viewing frustum corresponding to the second image can be determined. The viewing frustum corresponding to the target image region in the first image and the viewing frustum corresponding to the second image can be superimposed to obtain the target model region in the target model, i.e., the three-dimensional blurred region in the three-dimensional Gaussian model. Since the target image region in the first image and the second image are both regions with more concentrated blur, the three-dimensional blurred region in the three-dimensional Gaussian model can be obtained by superimposing the viewing frustum corresponding to the target image region in the first image and the viewing frustum corresponding to the second image.

[0089] In one embodiment of the present application, Figure 7 As shown, the method further includes:

[0090] S600, receiving rendering task information, where the rendering task information represents rendering an image corresponding to a target perspective through the target model;

[0091] S700, in response to the area corresponding to the target perspective including at least a portion of the target model area, obtaining an image corresponding to the target perspective by a method other than a target method, where the target method is to render the image corresponding to the target perspective by using a target model.

[0092] In this embodiment, when receiving rendering task information, the image corresponding to the target perspective can be rendered using the target model based on the rendering task information. Before rendering the image corresponding to the target perspective using the target model, it can be determined whether the area corresponding to the target perspective includes all or part of the target model area in the target model.

[0093] If the area corresponding to the target perspective includes all or part of the target model area in the target model, it indicates that the target model area is involved in the process of the target model rendering the image corresponding to the target perspective, that is, the area in the target model that does not meet the training requirements. In this case, if the image corresponding to the target perspective is rendered by the target model according to the rendering task information, the final rendered image will have obvious blurred areas. The image corresponding to the target perspective is obtained by a method other than the target method, and the target method is to render the image corresponding to the target perspective by the target model. That is, when it is confirmed that rendering using the target model will result in an image with obvious blurred areas, the target model is not used for rendering to avoid obvious blurred areas in the final rendered image.

[0094] Based on the same inventive concept, the second aspect of this application also provides an image processing device corresponding to an image processing method. Since the principle of solving the problem by the image processing device in this application is similar to that of the above-mentioned image processing method of this application, the implementation of the image processing device can refer to the implementation of the method, and the repeated parts will not be repeated.

[0095] Figure 8 The structure diagram of the image processing device provided in the embodiment of the present application is shown, which specifically includes:

[0096] an acquisition module, configured to acquire a target image, wherein the target image is rendered by a plurality of Gaussian volumes in a target model, and the target image includes a first number of image pixels;

[0097] a processing module, configured to determine a target pixel set based on a color contribution of the Gaussian body to each of the image pixels, wherein the target pixels in the target pixel set correspond to the same target Gaussian body, and the color contribution of the target Gaussian body to each of the target pixels satisfies a first preset condition;

[0098] A determination module is configured to determine whether the target image is blurred based on a ratio of a second number of target pixels to the first number of target pixels.

[0099] In one embodiment of the present application, the processing module is further configured to:

[0100] For a first pixel in the image pixels, determining a first Gaussian body and a second Gaussian body corresponding to the first pixel, wherein the second Gaussian body is closer to a viewing angle position of the target image than the first Gaussian body;

[0101] A color contribution of the first Gaussian body to the first pixel is determined based on the opacity of the first Gaussian body relative to the first pixel and the opacity of the second Gaussian body relative to the first pixel.

[0102] In one embodiment of the present application, the processing module is further configured to:

[0103] For a first pixel in the image pixels, determining color contributions of a plurality of candidate Gaussian bodies corresponding to the first pixel relative to the first pixel;

[0104] The Gaussian with the highest color contribution among the multiple candidate Gaussian bodies is determined as the target Gaussian body for the first pixel.

[0105] In one embodiment of the present application, the determination module is further configured to:

[0106] In a case where a ratio of the second number of target pixels to the first number is greater than a target threshold, it is determined that the target image is blurred.

[0107] In one embodiment of the present application, the processing module is further configured to:

[0108] determining a target image area in the target image according to distribution information of the target pixels in the target image;

[0109] The target image area is determined to be a blurred area in the target image.

[0110] In one embodiment of the present application, the processing module is further configured to:

[0111] Determining a second image, where the second image is an image obtained by rendering the target model, and the second image is a blurred image;

[0112] In the case that the first image is blurred, a target model region in the target model is determined according to the first image and the second image, where the target model region represents a region in the target model that does not meet training requirements.

[0113] In one embodiment of the present application, the determination module is further configured to:

[0114] determining a target image area in the first image according to distribution information of the target pixels in the first image;

[0115] A target model area in the target model is determined according to the target image area and the second image.

[0116] In one embodiment of the present application, the determination module is further configured to:

[0117] receiving rendering task information, wherein the rendering task information represents rendering an image corresponding to a target perspective through the target model;

[0118] In response to the area corresponding to the target perspective including at least a portion of the target model area, an image corresponding to the target perspective is obtained by a method other than a target method, where the target method is to render the image corresponding to the target perspective by using a target model.

[0119] Based on the same inventive concept, Figure 9 As shown, this embodiment also includes an electronic device, including:

[0120] a memory for storing executable programs;

[0121] The processor is configured to execute the executable program to implement the above method.

[0122] The above embodiments are merely exemplary embodiments of the present application and are not intended to limit the scope of the present application. The scope of protection of the present application is defined by the claims. Those skilled in the art may make various modifications or equivalent substitutions to the present application within the essence and scope of protection of the present application, and such modifications or equivalent substitutions shall also be deemed to fall within the scope of protection of the present application.

Claims

1. An image processing method, comprising: Acquire a target image, where the target image is rendered by a plurality of Gaussian volumes in a target model, and the target image includes a first number of image pixels; determining a target pixel set based on a color contribution of the Gaussian body to each of the image pixels, wherein target pixels in the target pixel set correspond to the same target Gaussian body, and a color contribution of the target Gaussian body to each of the target pixels satisfies a first preset condition; Whether the target image is blurred is determined based on a ratio of a second number of the target pixels to the first number.

2. The method according to claim 1, further comprising: For a first pixel in the image pixels, determining a first Gaussian body and a second Gaussian body corresponding to the first pixel, wherein the second Gaussian body is closer to a viewing angle position of the target image than the first Gaussian body; A color contribution of the first Gaussian body to the first pixel is determined based on the opacity of the first Gaussian body relative to the first pixel and the opacity of the second Gaussian body relative to the first pixel.

3. The method according to claim 1, further comprising: For a first pixel in the image pixels, determining color contributions of a plurality of candidate Gaussian bodies corresponding to the first pixel relative to the first pixel; The Gaussian with the highest color contribution among the multiple candidate Gaussian bodies is determined as the target Gaussian body for the first pixel.

4. The method according to claim 1 , wherein determining whether the target image is blurred based on a ratio of the second number of the target pixels to the first number of the target pixels comprises: In a case where a ratio of the second number of target pixels to the first number is greater than a target threshold, it is determined that the target image is blurred.

5. The method according to claim 1, further comprising: determining a target image area in the target image according to distribution information of the target pixels in the target image; The target image area is determined to be a blurred area in the target image.

6. The method according to claim 1, wherein the target image is a first image, the method further comprising: Determining a second image, where the second image is an image obtained by rendering the target model, and the second image is a blurred image; In the case that the first image is blurred, a target model region in the target model is determined according to the first image and the second image, where the target model region represents a region in the target model that does not meet training requirements.

7. The method according to claim 6, wherein determining the target model area in the target model according to the first image and the second image comprises: determining a target image area in the first image according to distribution information of the target pixels in the first image; A target model area in the target model is determined according to the target image area and the second image.

8. The method according to claim 6, further comprising: receiving rendering task information, wherein the rendering task information represents rendering an image corresponding to a target perspective through the target model; In response to the area corresponding to the target perspective including at least a portion of the target model area, an image corresponding to the target perspective is obtained by a method other than a target method, where the target method is to render the image corresponding to the target perspective by using a target model.

9. An image processing device comprising: an acquisition module, configured to acquire a target image, wherein the target image is rendered by a plurality of Gaussian volumes in a target model, and the target image includes a first number of image pixels; a processing module, configured to determine a target pixel set based on a color contribution of the Gaussian body to each of the image pixels, wherein the target pixels in the target pixel set correspond to the same target Gaussian body, and the color contribution of the target Gaussian body to each of the target pixels satisfies a first preset condition; A determination module is configured to determine whether the target image is blurred based on a ratio of a second number of target pixels to the first number of target pixels.

10. An electronic device comprising a processor and a memory, wherein an executable program is stored in the memory, and the processor executes the executable program to perform the image processing method according to any one of claims 1 to 8.

Images