CN114119842B - Rendering method and system based on SSIM (structural similarity) and PSNR (Peak Signal to noise ratio) algorithm and computer readable storage medium - Google Patents
Rendering method and system based on SSIM (structural similarity) and PSNR (Peak Signal to noise ratio) algorithm and computer readable storage medium Download PDFInfo
- Publication number
- CN114119842B CN114119842B CN202111407147.9A CN202111407147A CN114119842B CN 114119842 B CN114119842 B CN 114119842B CN 202111407147 A CN202111407147 A CN 202111407147A CN 114119842 B CN114119842 B CN 114119842B
- Authority
- CN
- China
- Prior art keywords
- graph
- rendering
- effect
- parameters
- original image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T15/00—3D [Three Dimensional] image rendering
- G06T15/06—Ray-tracing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformation in the plane of the image
- G06T3/40—Scaling the whole image or part thereof
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Abstract
The application relates to a rendering method, a system and a computer readable storage medium based on SSIM and PSNR algorithms, which comprises the steps of obtaining an original image, and generating a first thumbnail according with scene parameters of the original image according to the original image; rendering the first thumbnail according to preset parameters to obtain a comparison graph; acquiring an effect parameter of a comparison graph, and determining whether the effect parameter of the comparison graph accords with a preset parameter; if so, taking the comparison graph as an effect graph; preliminarily rendering the original image based on the effect image and preset parameters, and generating a conversion image of part of the rendered original image; extracting a rendered part of the conversion graph to obtain a rendered graph; acquiring characteristic parameters of a rendering graph, characteristic parameters of an effect graph and characteristic parameters of a corresponding part of the rendering graph, confirming whether the characteristic parameters of the two graphs are consistent, and outputting judgment information; and determining whether to continue rendering the conversion graph according to preset parameters based on the judgment information, and outputting a completely rendered graph. The method and the device can acquire the rendering condition in real time, check the correctness of the rendering effect of the rendering farm, and save the machine time cost.
Description
Technical Field
The invention relates to the technical field of image rendering, in particular to a rendering method and a rendering system based on SSIM and PSNR algorithms and a computer readable storage medium.
Background
At present, a rendering farm is actually called a "distributed parallel cluster computing system", which is a supercomputer built by using an off-the-shelf CPU, ethernet and an operating system, and which uses mainstream commercial computer hardware equipment to reach or approach the computational power of the supercomputer; the rendering farms are divided into traditional rendering farms only supporting the offline and cloud rendering farms supporting the online, and share huge clusters and strong hardware facilities.
When a user on the market renders on the rendering farm, the formed image needs to be downloaded after the rendering is completed, and then manual inspection is carried out, or the formed image is compared with the user local map by using a rendering algorithm. However, due to the fact that factors such as complex scenes and field noise can cause image rendering deviation, rendering effect check is performed after rendering is completed, intervention cannot be performed in the rendering process, and the time for correcting the image rendering deviation is missed; if the mapping is not correct and the rendering effect is inconsistent with the preset parameters after the rendering is finished, the rendering needs to be carried out again, so that a large amount of time is wasted.
The above-mentioned related art has a drawback of high cost in the case of organic matters.
Disclosure of Invention
In order to solve the problem of high time cost, the invention aims to provide a rendering method based on SSIM and PSNR algorithms.
The above object of the present invention is achieved by the following technical solutions:
a rendering method based on SSIM and PSNR algorithms comprises the following steps,
acquiring an original image, and generating a first thumbnail according with the scene parameters of the original image according to the original image;
rendering the first thumbnail according to preset parameters to obtain a comparison graph;
acquiring the effect parameters of the comparison graph, and determining whether the effect parameters of the comparison graph accord with the preset parameters; if so, taking the comparison graph as an effect graph;
preliminarily rendering the original image based on the effect image and the preset parameters, and generating a conversion image for partially rendering the original image;
extracting a rendered part of the conversion map to obtain a rendered map;
acquiring the characteristic parameters of the rendering graph and the characteristic parameters of the effect graph and the corresponding part of the rendering graph, confirming whether the characteristic parameters of the rendering graph meet the characteristic parameters of the effect graph and the corresponding part of the rendering graph, and outputting judgment information;
and determining whether to continue rendering the conversion graph according to the preset parameters or not based on the judgment information, and outputting a completely rendered graph.
By adopting the technical scheme, whether the effect parameters of the comparison graph meet the preset parameters is determined, the standardized comparison graph is used as the effect graph, the effect graph is used as a comparison sample of the rendering graph, and the accuracy of the farm rendering effect is improved; in the process of rendering the original image, comparing the characteristic parameters of the rendering image with the characteristic parameters of the effect image, and determining whether the characteristic parameters of the rendering image meet the characteristic parameters of the effect image and the corresponding part of the rendering image, so that whether the rendering effect of the original image in the process of rendering meets the requirements is determined, and the correctness of the rendering farm image is ensured; the method and the device have the advantages that the intervention on the original image rendering process is realized, the correctness of the rendering effect can be detected in real time, the rendering condition is obtained in real time, the checking and the comparison are not needed after the image is drawn, the efficiency is improved, the nonstandard drawing is avoided, the waste of resources is reduced, and the machine-time cost is saved.
The present invention in a preferred example may be further configured to:
the method for generating the first thumbnail conforming to the scene parameters of the original image according to the original image comprises the following steps:
obtaining an original image, and generating a primary thumbnail of the original image in an equal-scale scaling manner;
acquiring scene parameters of the original image and scene parameters of the preliminary thumbnail;
determining whether the scene parameters of the preliminary thumbnail accord with the scene parameters of the original picture;
if so, taking the preliminary thumbnail as a first thumbnail;
and if not, re-rendering the preliminary thumbnail.
By adopting the technical scheme, whether the scene parameters of the first thumbnail, such as the three-dimensional model, the optics, the material quality, the dynamics and the like, are in accordance with the scene parameters of the original image is determined, and the standard property of the effect image which is subsequently rendered by rendering the first thumbnail is ensured. If so, taking the preliminary thumbnail as a first thumbnail; and if not, stopping the rendering step, and determining the scene parameters of the preliminary thumbnail again after re-rendering the preliminary thumbnail. The preliminary thumbnail scene parameter determination and the first thumbnail acquisition ensure the correctness of the rendered farm map, thereby saving time.
The present invention in a preferred example may be further configured to:
obtaining an effect parameter of the comparison graph, and determining whether the effect parameter accords with the preset parameter; if yes, the step of using the comparison graph as an effect graph comprises the following steps:
obtaining an effect parameter of the comparison graph according to a ray tracing algorithm;
determining whether the effect parameters of the comparison graph accord with the preset parameters or not;
if so, taking the comparison graph as an effect graph;
if not, re-rendering the comparison graph.
By adopting the technical scheme, the characteristics of the contrast image are extracted by using a ray tracing algorithm to obtain the effect parameters of the contrast image, and then the effect parameters of the contrast image are compared with the preset parameters, so that the standardized effect image is obtained, the effect image can be used as a contrast sample for rendering the original image, the subsequent original image can be rendered according to the standard effect image, and the accuracy of the rendering effect is effectively improved; when the effect parameters of the comparison graph accord with preset parameters, taking the comparison graph as an effect graph; when the effect parameters of the comparison graph do not accord with the preset parameters, the back end outputs the non-conforming place in the comparison graph to the front end to play a warning role, and after the comparison graph is re-rendered, whether the effect parameters of the comparison graph accord with the preset parameters is determined again. Therefore, the information of the rendering effect of the effect graph can be obtained in real time, and the rendering process can be intervened in time; furthermore, the effect icons are standardized, the requirement of serving as a comparison sample is met, the correctness of the rendering effect of the original image is guaranteed, and the machine time cost is saved.
The present invention in a preferred example may be further configured to:
based on the effect graph and the preset parameters, the original graph is rendered preliminarily, and the step of generating a conversion graph for partially rendering the original graph comprises the following steps:
primarily rendering the original image according to the preset parameters after the effect image is obtained;
and acquiring a conversion map for partially rendering the original image based on preset time.
By adopting the technical scheme, after the effect graph is obtained, the original graph is rendered according to the preset parameters, in the original graph rendering process, the partial conversion graph for rendering the original graph is obtained according to the preset time, the rendering condition of the original graph is obtained in real time, and the correctness of rendering the graph of the farm is ensured.
The present invention in a preferred example may be further configured to:
the steps of obtaining the feature parameters of the rendering graph and the feature parameters of the effect graph and the corresponding part of the rendering graph, confirming whether the feature parameters of the rendering graph conform to the feature parameters of the effect graph and the corresponding part of the rendering graph, and outputting judgment information comprise:
acquiring a position value of a pixel of the effect image based on a pixel calculation method;
acquiring a characteristic value and a relative position information value of the effect graph according to the position value, wherein the characteristic value and the relative position information value are used as characteristic parameters of the effect graph;
and acquiring the characteristic parameters of the rendering graph, confirming whether the characteristic parameters of the rendering graph meet the characteristic parameters of the effect graph and the corresponding part of the rendering graph, and outputting judgment information.
By adopting the technical scheme, for each pixel point of the effect image, the product of the neighborhood pixel and the corresponding element of the filter matrix is calculated and then added up to be used as the value of the pixel position, so that the position value of the pixel of the whole effect image is obtained, and then the convolution calculation is carried out on the effect image and the convolution kernel according to the position value, so that the characteristic value and the relative position information value of the pixel can be obtained and used as the characteristic parameter of the effect image, a comparison sample is provided for the characteristic parameter of the rendering image, and the accuracy of the rendering effect of the rendering farm is ensured.
The present invention in a preferred example may be further configured to:
the steps of obtaining the characteristic parameters of the rendering graph, confirming whether the characteristic parameters of the rendering graph meet the characteristic parameters of the effect graph and the corresponding part of the rendering graph, and outputting judgment information comprise:
acquiring the characteristic parameters of the rendering graph by using SSIM and PSNR algorithms according to the characteristic parameters of the effect graph, and generating the similarity between the effect graph and the rendering graph;
and determining whether the characteristic parameters of the rendering graph conform to the characteristic parameters of the corresponding parts of the effect graph and the rendering graph based on the similarity, and outputting judgment information.
By adopting the technical scheme, the similarity of the images is measured by the SSIM algorithm from three aspects of brightness, contrast and structure, and the SSIM algorithm is a full-reference image quality evaluation index; the PSNR algorithm objectively evaluates the image based on the error between corresponding pixel points; according to the two algorithms, the similarity between the screenshot and the effect graph is obtained from the error, the brightness, the contrast and the structure among the corresponding pixel points, so that the contrast effect is better, and the similarity is more accurate; the method ensures the rendering correctness of the original image, reduces the rendering error and further saves the time cost. Scanning the rendering graph by using an SSIM algorithm and a PSNR algorithm, acquiring characteristic parameters of the rendering graph, comparing the characteristic parameters with the characteristic parameters of the effect graph to generate similarity, and determining whether the similarity meets the requirement; whether the characteristic parameters of the conversion diagram meet the characteristic parameters of the effect diagram or not can be determined in real time, so that the rendering effect of the conversion diagram reaches the standard, the correctness of the rendering effect of the rendering farm is ensured, the correctness of the rendering farm drawing is further ensured, and the machine time cost is saved.
The present invention in a preferred example may be further configured to:
determining whether the characteristic parameters of the rendering graph conform to the characteristic parameters of the corresponding parts of the effect graph and the rendering graph based on the similarity, and outputting judgment information comprises the following steps:
if the similarity is in a preset first threshold value, which indicates that the characteristic parameters of the rendering graph accord with the characteristic parameters of the effect graph, continuing to render the conversion graph;
and if the similarity is in a preset second threshold value, which indicates that the characteristic parameters of the rendering graph do not accord with the characteristic parameters of the effect graph, re-rendering the conversion graph.
By adopting the technical scheme, the preset first threshold value and the preset second threshold value are used for judging whether the characteristic parameters of the rendering graph meet the characteristic parameters of the effect graph and the corresponding part of the rendering graph. When the similarity is in a preset first threshold value, the feature parameters of the rendering graph are in accordance with the feature parameters of the effect graph, and then the conversion graph is rendered continuously; and when the similarity is in a preset second threshold value, indicating that the characteristic parameters of the rendering graph do not accord with the characteristic parameters of the effect graph, and re-rendering the conversion graph. Furthermore, whether the characteristic parameters of the rendering graph accord with the characteristic parameters of the effect graph or not can be rendered by presetting the first threshold and the second threshold, the rendering condition is obtained in real time, and when the characteristic parameters do not accord with the characteristic parameters of the effect graph, the error conversion graph can be immediately re-rendered, so that the time cost is saved.
The present invention in a preferred example may be further configured to:
determining whether to continue rendering the conversion graph according to the preset parameters based on the judgment information, wherein the step of outputting the completely rendered graph comprises the following steps:
if the characteristic parameters of the rendering graph accord with the characteristic parameters of the effect graph, continuing to render the conversion graph according to the preset parameters;
obtaining a chart according to the completely rendered conversion chart;
generating a second thumbnail using an image scaling algorithm based on the map;
obtaining a comparison result of the second thumbnail and the effect graph according to SSIM and PSNR algorithms;
and outputting the chart and the comparison result.
By adopting the technical scheme, after the rendering is finished, the image scaling algorithm is used for immediately converting the formed image into the second thumbnail without losing the structural information, and the SSIM and PSNR algorithms are used for comparing the characteristic parameters of the second thumbnail with the characteristic parameters of the effect image to generate a comparison result. According to the arrangement, the second thumbnail corresponds to the effect graph in size, so that comparison is facilitated, the accuracy of a comparison result is improved, the correctness of a farm rendering effect is checked, and the machine-time cost is saved; meanwhile, the output of the comparison result does not need to check whether the preset parameters are met or not after the picture is drawn, and the efficiency is improved.
The invention also aims to provide a system for testing the rendering effect graph, which has the characteristic of saving the machine time cost.
The second aim of the invention is realized by the following technical scheme:
a rendering system based on SSIM and PSNR algorithm comprises,
the generating module is used for acquiring an original image and generating a first thumbnail according with the scene parameters of the original image according to the original image;
the comparison image obtaining module is used for rendering the first thumbnail according to preset parameters to obtain a comparison image;
the effect graph acquisition module is used for acquiring the effect parameters of the comparison graph and determining whether the effect parameters of the comparison graph meet the preset parameters or not; if yes, taking the comparison graph as an effect graph;
the conversion map acquisition module is used for primarily rendering the original map and generating a conversion map for partially rendering the original map based on the effect map and the preset parameters;
the rendering map acquisition module is used for extracting a rendered part of the conversion map to obtain a rendering map;
the rendering graph detection module is used for acquiring the characteristic parameters of the rendering graph and the characteristic parameters of the corresponding part of the effect graph and the rendering graph, confirming whether the characteristic parameters of the rendering graph meet the characteristic parameters of the corresponding part of the effect graph and the rendering graph, and outputting judgment information;
and the formed graph output module is used for determining whether to continue rendering the conversion graph according to the preset parameters based on the judgment information and outputting a completely rendered formed graph.
The third purpose of the invention is to provide a computer readable storage medium, which can store corresponding programs and has the characteristic of saving the time cost.
The third object of the invention is realized by the following technical scheme:
a computer readable storage medium storing a computer program capable of being loaded by a processor and executing any one of the above-mentioned rendering methods based on the SSIM and PSNR algorithms.
Drawings
Fig. 1 is a structural block diagram of a rendering method based on SSIM and PSNR algorithms according to the present application.
Fig. 2 is a schematic flow chart of a rendering method based on SSIM and PSNR algorithms according to the present application before rendering an original image.
Fig. 3 is a schematic flowchart illustrating rendering of an original image in a rendering method based on SSIM and PSNR algorithms according to the present application.
Fig. 4 is a schematic flow chart of an output chart in a rendering method based on SSIM and PSNR algorithms according to the present application.
Fig. 5 is a schematic block diagram of a rendering system based on SSIM and PSNR algorithms according to the present application.
Description of reference numerals: 1. a generation module; 2. a comparison image acquisition module; 3. an effect picture acquisition module; 4. a conversion map acquisition module; 5. a rendering map acquisition module; 6. a rendering map detection module; 7. and a diagram output module.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The embodiments of the present invention will be described in further detail with reference to the drawings attached hereto.
Example one
Referring to fig. 1, an embodiment of the present application provides a rendering method based on SSIM and PSNR algorithms, which is applied to a rendering farm. The existing rendering farm can cause deviation of images in the rendering process due to factors such as image compression, complex scenes, site noise and the like, so that incorrect drawing is caused, and machine time waste is caused. According to the embodiment of the application, through intervening in the rendering process of the image, in the process of converting the original image into the formed image, the converted image is checked and corrected step by step, and the method specifically comprises the following steps before the image is rendered:
referring to fig. 1, in step S100, an original is acquired, and a first thumbnail image corresponding to a scene parameter of the original is generated from the original.
Specifically, the original image is a picture or animation that a user needs to render; and reducing the original image in the same proportion to obtain a first thumbnail which is consistent with the scene parameter of the original image.
Referring to fig. 2, step S100 specifically includes step S101 of acquiring an original image and generating a preliminary thumbnail image obtained by scaling the original image in an equal ratio.
And S102, acquiring the scene parameters of the original image and the scene parameters of the preliminary thumbnail.
Step S103, determining whether the scene parameters of the preliminary thumbnail accord with the scene parameters of the original picture; if so, taking the preliminary thumbnail as a first thumbnail; and if not, re-rendering the preliminary thumbnail.
Specifically, after an original image is obtained, the original image is reduced to a primary thumbnail in an equal proportion; whether scene parameters such as a three-dimensional model, optics, materials and dynamics in the primary thumbnail scene meet the scene parameters of the original image or not is detected through an API provided by 3D software, the scene of the original image can be reflected in the first thumbnail, and the subsequent rendering is more standard. If the first thumbnail is matched with the first thumbnail, taking the preliminary thumbnail as a first thumbnail and starting to execute the next step; and if the original image does not accord with the scene parameter, pausing the rendering program, outputting the place which does not accord with the scene of the original image in the scene of the primary thumbnail to the front end, generating the primary thumbnail from the original image again, and determining the scene parameter of the primary thumbnail and the scene parameter of the original image again. The determination of the scene parameters of the preliminary thumbnail and the acquisition of the first thumbnail are equivalent to the generation of a reduced version of the original image, so that a comparison sample can be obtained by faster rendering, and the time is saved.
Referring to fig. 1, in step S200, the first thumbnail is rendered according to preset parameters to obtain a comparison graph.
Specifically, the preset parameters are configuration parameters selected by a user in the rendering farm according to rendering requirements before rendering, and the original image is rendered according to the preset parameters. Rendering the first thumbnail according with the scene parameters of the original image according to the preset parameters, and obtaining a contrast image after the first thumbnail is rendered.
Step S300, obtaining the effect parameter of the comparison graph, and determining whether the effect parameter of the comparison graph accords with the preset parameter; and if so, taking the comparison graph as an effect graph.
Specifically, the effect parameters include color characteristics of the image, a scene gamma value, a scene peak signal-to-noise ratio, and the like. The effect parameter of the comparison graph is the embodiment of the rendering effect after the rendering of the first thumbnail is finished; and comparing the effect parameters of the comparison graph with preset parameters, and if the effect parameters of the comparison graph are consistent with the preset parameters, taking the comparison graph as an effect graph. For example, the reference index of the scene gamma value is 0-255, the reference index of the scene peak signal-to-noise ratio is larger than 255, and other reference indexes, and if the effect parameters of the comparison graph all accord with the reference indexes, the effect parameters of the comparison graph accord with preset parameters. Through the effect parameter of inspection contrast picture for the effect icon standardization can be as the contrast sample that the original image was rendered, makes follow-up original image can be according to more standard effect picture and come the rendering, thereby the effectual exactness that improves the rendering effect, and then saved the machine-hour cost.
Referring to fig. 2, step S300 specifically includes S301, obtaining an effect parameter of the contrast map according to a ray tracing algorithm.
Step S302, determining whether the effect parameters of the comparison graph accord with the preset parameters; if so, taking the comparison graph as an effect graph; if not, re-rendering the comparison graph.
Specifically, the characteristics are extracted through a comparison graph rendered by using a ray tracing algorithm, so that effect parameters of the comparison graph are obtained; and comparing the preset parameters with the effect parameters of the comparison graph, if the preset parameters are matched with the effect parameters of the comparison graph, taking the comparison graph as an effect graph, and enabling the effect graph to become a comparison standard of a subsequent rendering step. And if the first thumbnail does not accord with the preset parameters, pausing the rendering program, outputting a warning that the comparison graph does not accord with the preset parameters to the front end, updating the rendering working state, re-rendering the first thumbnail to obtain the comparison graph, and determining whether the effect parameters of the comparison graph accord with the preset parameters again. Therefore, the rendering effect information of the effect graph can be acquired in real time, the rendering process can be intervened in time, further, the effect graph is standardized, and the requirement of serving as a sample is met.
Referring to fig. 1, in step S400, the original image is primarily rendered based on the effect map and the preset parameters, and a conversion map for partially rendering the original image is generated.
Specifically, the preliminary rendering is that the original image is not completely rendered. And after the effect graph is obtained, the original graph is subjected to preliminary rendering according to preset parameters, and only a rendering part of the original graph is used as a conversion graph at the time.
Referring to fig. 3, the step S400 specifically includes a step 401 of primarily rendering the original image according to the preset parameters after the effect image is obtained.
And S402, acquiring a conversion map for partially rendering the original image based on preset time.
Specifically, the preset time is an interval time preset by a user, and the conversion chart is obtained every three minutes.
Referring to fig. 1, in step S500, a rendered portion of the conversion map is extracted to obtain a rendered map.
Specifically, a Windows system API is called up by using a python code to find a rendering window, and screenshots are taken on the rendering window every 3 minutes according to a preset time, wherein the rendered part in the screenshots is a rendering graph. Accordingly, the screenshot interval time of the rendering window is not limited to three minutes, and can be debugged according to the requirement. The method and the device have the advantages that screenshot is continuously carried out on the rendering window in the rendering process at intervals, so that the correctness of the rendering effect of the original image is guaranteed step by step.
Step S600, obtaining the characteristic parameters of the rendering graph and the characteristic parameters of the effect graph and the corresponding part of the rendering graph, confirming whether the characteristic parameters of the rendering graph meet the characteristic parameters of the effect graph and the corresponding part of the rendering graph, and outputting judgment information.
Specifically, the characteristic parameter of the rendering map and the characteristic parameter of the effect map are a position value and a relative position information value of the pixel.
Referring to fig. 3, step S600 specifically includes S601, and obtaining a position value of a pixel of the effect map based on a pixel calculation method.
Specifically, the product of the neighborhood pixels of each pixel point of the effect map and the corresponding elements of the filter matrix is calculated, and the result is added as the value of the pixel position, thereby obtaining the position value of the pixel of the entire effect map.
Step S602, according to the position value, obtaining a characteristic value and a relative position information value of the effect graph, and using the characteristic value and the relative position information value as characteristic parameters of the effect graph.
Specifically, the effect graph and the convolution kernel are convolved according to the position value to obtain a characteristic value and a relative position information value of the effect graph, and the characteristic value and the relative position information value are used as identification indexes and can be used as sample characteristics to be compared with the rendering effect of the original graph.
Step S603, acquiring the characteristic parameters of the rendering graph by using SSIM and PSNR algorithms according to the characteristic parameters of the effect graph, and generating the similarity between the effect graph and the rendering graph;
specifically, the similarity of the images is measured by the SSIM algorithm from the three aspects of height, contrast and structure, the PSNR algorithm objectively evaluates the images based on errors among corresponding pixel points, and the two algorithms are combined to comprehensively compare the sample characteristics of the effect graph and the structural parameter information of the rendered part of the screenshot, so that the contrast effect is better, the obtained similarity is more accurate, the rendering accuracy of the original image is further ensured, the rendering error is reduced, and the machine-time cost is further saved. Accordingly, in the embodiment of the present application, the evaluation and comparison of the images are not limited to the SSIM and PSNR algorithms, and other algorithms capable of evaluating the image quality, such as MOVIE and VMAF, may also be used.
And comparing the characteristic parameters of the rendering graph with the characteristic parameters of the effect graph by using an SSIM (structural similarity) and PSNR (Peak Signal to noise ratio) algorithm to obtain the similarity, wherein the similarity is used for judging whether the characteristic parameters of the rendering graph accord with the characteristic parameters of the effect graph or not. The method comprises the steps of presetting a first threshold and a second threshold, wherein the first threshold and the second threshold are continuous and non-overlapping.
Step S604, if the similarity is in a preset first threshold value, which indicates that the characteristic parameters of the rendering graph conform to the characteristic parameters of the effect graph, continuing to render the conversion graph; and if the similarity is in a preset second threshold value, which indicates that the characteristic parameters of the rendering graph do not accord with the characteristic parameters of the effect graph, re-rendering the conversion graph.
Specifically, if the similarity is within a preset first threshold value, which indicates that the feature parameters of the rendering graph conform to the feature parameters of the effect graph, continuing to render the conversion graph; if the similarity is in the preset second threshold value, the characteristic parameters of the rendering graph are not accordant with the characteristic parameters of the effect graph, the rendering program is suspended, the places where the rendering graph is not accordant are output, the conversion graph is re-rendered, and the rendering operation state is updated. Because the conversion graph is subjected to screenshot once every three minutes, the rendered part in the screenshot obtains the rendering graph, and the characteristic parameters of the rendering graph are checked, when the characteristic parameters of the rendering graph do not accord with the characteristic parameters of the effect graph, the rendering is started from the last conversion graph without re-rendering, and the machine-time cost is saved.
In the embodiment of the invention, the total range of the similarity is 0-1, the range of the preset first threshold value is 0.8-1, the range of the preset second threshold value is 0-0.8, and if the range of the similarity is 0.8-1, the characteristic parameters of the rendering graph conform to the characteristic parameters of the effect graph, the conversion graph is continuously rendered; and if the similarity range is between 0 and 0.8, indicating that the characteristic parameters of the rendering graph do not conform to the characteristic parameters of the effect graph, pausing rendering, sending a message to the front end by the rear end of the rendering platform through a websocket service, timely warning and updating the rendering operation state, and continuing rendering from the position where the error occurs after solving the problem. Whether the characteristic parameters of the rendering graph accord with the characteristic parameters of the effect graph or not can be judged through the set similarity range, intervention on the rendering process is achieved, the rendering condition can be obtained in real time, if the effects are inconsistent, rendering can be immediately suspended, and therefore the machine-time cost is saved. Correspondingly, the closer the similarity is to 1, the closer the characteristic parameter of the rendering graph is to the preset parameter; therefore, the range of the preset first threshold and the preset second threshold can be adjusted to adjust the rendering precision of the original image. For example: the preset first threshold is adjusted to be 0.85-1, the preset second threshold is 0-0.85, the similarity needs to be in the range of 0.85-1 to indicate that the effects are consistent, and the better the output rendering effect of the formed image is.
Referring to fig. 1, in step S700, based on the determination information, it is determined whether to continue rendering the conversion map according to the preset parameter, and a completely rendered map is output.
Specifically, if the characteristic parameters of the rendering graph conform to the characteristic parameters of the effect graph, the conversion graph is rendered continuously, when the conversion graph is completely rendered, a graph is obtained, and finally the graph is output.
Referring to fig. 4, the step S700 specifically includes a step S701 of continuing to render the conversion map according to the preset parameter if the feature parameter of the rendering map meets the feature parameter of the effect map.
Step S702, obtaining a chart according to the completely rendered conversion chart;
s703, generating a second thumbnail by using an image scaling algorithm according to the formed image;
step S704, obtaining a comparison result of the second thumbnail and the effect graph according to SSIM and PSNR algorithms;
and step S705, outputting the chart and the comparison result.
In particular, the second thumbnail is scaled equally without losing the patterned structured information. In order to save time and ensure the correctness of the rendering effect of the formed image and avoid the conditions of rendering inconsistency or abnormal program and the like in the last screenshot interval time, after the rendering of the converted image is finished, an image scaling algorithm is used for generating a second thumbnail of the formed image, SSIM and PSNR algorithms are used for comparing the rendering effect of the second thumbnail and the rendering effect of the effect image, and finally the formed image and the comparison result are output. Specifically, in the embodiment of the invention, the image scaling algorithm is used, and the formed image is converted into the second thumbnail reduced in the same proportion on the premise of not losing the structured information of the formed image, so that the second thumbnail can contain the information of the formed image to the maximum extent, and the comparison accuracy is improved. Comparing the effect graph with an SSIM algorithm and a PSNR algorithm, and outputting a graph and a comparison result; and checking the correctness of the rendering effect of the farm, and ensuring the correctness of the rendering effect of the formed image again. Therefore, after the picture is drawn, a user does not need to check whether the picture meets the preset parameters, the time is saved, and the efficiency is improved.
In the embodiment of the invention, the output comparison result is a detection report, the detection report comprises information such as the name of a detected item, the detection result, the reference index of the detection result, the abnormality recently triggered by a program and the like, different marks are used for distinguishing the detection result with the abnormality, and the detected item comprises whether the scene contains dynamics calculation, the material type, the color characteristic of a scene image, the scene gamma value, the scene peak signal-to-noise ratio, the scene similarity and the like. So that the consulting is convenient, and the efficiency is improved.
The rendering method based on SSIM and PSNR algorithm in the embodiment of the invention has the following implementation principle: generating a first thumbnail which accords with the scene parameters of the original image from the original image, rendering the first thumbnail to obtain an effect image with effect parameters which accord with preset parameters, and rendering the original image; the previous two judgments are in the previous stage of rendering, and are all used for obtaining a standard effect graph so as to ensure the accuracy of a subsequent comparison sample, and when inconsistency occurs, intervention can be performed in time, so that the machine time cost is saved.
Continuously screenshot a rendering window in the process of rendering an original image, scanning a rendering image by using SSIM (structural similarity) and PSNR (Peak-to-average rating) algorithms, comparing the rendering image with an effect image to obtain similarity, and detecting the similarity; what this time detected is at the middle stage of rendering, and in the process of rendering, whether the characteristic parameter of rendering map accords with the characteristic parameter of effect picture, compares whole rendering process step by step, has guaranteed the correctness of rendering effect to can acquire the condition of rendering in real time, the condition of not conforming to appears, can in time intervene, has further saved the machine time cost.
After the rendering is finished, converting the formed image into a second thumbnail, performing contrast detection on the formed image and the effect image by using an SSIM (structural similarity model) and PSNR (Peak Signal to noise ratio) algorithm, and finally outputting the formed image and a contrast result; in the post-rendering stage of the detection, the correctness of the farm rendering effect is checked again, the correctness of the mapping rendering effect is ensured, the situation that the mapping does not meet the requirement due to rendering deviation or system abnormity and the like in the final rendering process is avoided, and the machine time cost is saved; meanwhile, the output of the comparison result can be used for judging whether the preset parameters are met or not after the picture is drawn or comparing the preset parameters with the reference picture again, and the efficiency is improved.
Compared with the prior art, the method has the advantages that the intervention of the rendering process is realized, the correctness of the rendering effect can be detected in real time, the rendering condition is obtained, the checking and the comparison after the picture is drawn are not needed, the efficiency is improved, the out-of-standard picture is avoided, the waste of resources is reduced, and the time cost is saved.
Referring to fig. 5, the present application provides a rendering system based on SSIM and PSNR algorithms, which is in one-to-one correspondence with the rendering methods based on SSIM and PSNR algorithms in the first embodiment, and can acquire rendering conditions in real time, check the correctness of rendering farm rendering effects, and save the machine time cost, and the system includes a generation module 1, a comparison diagram acquisition module 2, an effect diagram acquisition module 3, a conversion diagram acquisition module 4, a rendering diagram acquisition module 5, a rendering diagram detection module 6, and a diagram output module 7. The functional modules are explained in detail as follows:
the generation module 1 is configured to acquire an original image, and generate a first thumbnail according with scene parameters of the original image according to the original image.
Specifically, an original image is input, the original image is reduced in the same proportion to obtain a preliminary thumbnail, scene parameters of the preliminary thumbnail are confirmed, and if the scene parameters of the preliminary thumbnail accord with the scene parameters of the original image, the preliminary thumbnail is used as a first thumbnail; and if the scene parameters of the preliminary thumbnail do not accord with the scene parameters of the original picture, re-rendering the preliminary thumbnail and confirming the scene parameters. This ensures that the scene of the artwork is represented in the first thumbnail.
And the comparison image obtaining module 2 is used for rendering the first thumbnail according to preset parameters to obtain a comparison image.
Specifically, a first thumbnail conforming to the scene parameters of the original image is rendered according to preset parameters to obtain a comparison image.
The effect graph obtaining module 3 is configured to obtain an effect parameter of the comparison graph, and determine whether the effect parameter of the comparison graph meets the preset parameter; and if so, taking the comparison graph as an effect graph.
Specifically, comparing the effect parameters of the comparison graph with preset parameters, and if the effect parameters are consistent with the preset parameters, taking the comparison graph as an effect graph; if not, pausing the rendering program, outputting the non-conforming place to the front end, re-rendering the comparison graph, and confirming whether the effect parameters of the comparison graph conform to the preset parameters; the obtained effect graph meets the standard, and the correctness of rendering the graph of the subsequent original graph is improved.
And the conversion map acquisition module 4 is configured to preliminarily render the original map based on the effect map and the preset parameter, and generate a conversion map for partially rendering the original map.
Specifically, after the effect graph is obtained, the original graph is primarily rendered according to preset parameters, and only a rendering part of the original graph is used as a conversion graph at the time.
And the rendering map acquisition module 5 is configured to extract a rendered portion of the conversion map to obtain a rendering map.
Specifically, through a set time interval, continuous screenshot is carried out on a rendering window, and a rendered part in the screenshot is used as a rendering map.
And the rendering graph detection module 6 is configured to acquire the feature parameters of the rendering graph and the feature parameters of the corresponding portion of the effect graph and the rendering graph, determine whether the feature parameters of the rendering graph meet the feature parameters of the corresponding portion of the effect graph and the rendering graph, and output determination information.
Specifically, scanning analysis is performed on the rendering graph by using an SSIM algorithm and a PSNR algorithm, and the characteristic parameters of the rendering graph and the characteristic parameters of the effect graph are compared and analyzed to obtain the similarity. Judging whether the characteristic parameters of the rendering graph accord with the characteristic parameters of the effect graph or not by detecting the similarity, and if so, continuing to render the conversion graph; if not, pausing the rendering program, outputting the non-coincident place in the rendering graph to the front end, re-rendering the conversion graph, and detecting whether the characteristic parameters of the rendering graph accord with the characteristic parameters of the effect graph again.
And the formed graph output module 7 is configured to determine whether to continue rendering the conversion graph according to the preset parameter based on the determination information, and output a completely rendered formed graph.
Specifically, after rendering of the conversion graph is completed, a graph is obtained; and converting the formed image into a second thumbnail by using an image scaling algorithm, scanning and comparing the second thumbnail and the effect image by using SSIM and PSNR algorithms to obtain a comparison result, and finally outputting the formed image and the comparison result. So that manual inspection is not needed after drawing, the waste of time is reduced, and the efficiency is improved.
The present application provides a computer-readable storage medium storing a computer program that can be loaded by a processor and executes the above-mentioned small-area fingerprint image feature extraction method, the computer program realizing the following steps when executed by the processor:
s100, obtaining an original image, and generating a first thumbnail according with the scene parameters of the original image according to the original image;
s200, rendering the first thumbnail according to preset parameters to obtain a comparison graph;
step S300, obtaining the effect parameter of the comparison graph, and determining whether the effect parameter of the comparison graph accords with the preset parameter; if so, taking the comparison graph as an effect graph;
s400, primarily rendering the original image based on the effect image and the preset parameters, and generating a conversion image for partially rendering the original image;
s500, extracting a rendered part of the conversion map to obtain a rendered map;
step S600, acquiring feature parameters of the rendering graph and feature parameters of the effect graph and a part corresponding to the rendering graph, confirming whether the feature parameters of the rendering graph meet the feature parameters of the effect graph and the part corresponding to the rendering graph, and outputting judgment information;
and S700, determining whether to continue rendering the conversion graph according to the preset parameters or not based on the judgment information, and outputting a completely rendered graph.
The computer-readable storage medium includes, for example: various media capable of storing program codes, such as a usb disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk.
The embodiments are preferred embodiments of the present application, and the scope of the present application is not limited by the embodiments, so: all equivalent changes made according to the structure, shape and principle of the present application shall be covered by the protection scope of the present application.
Claims (9)
1. The rendering method based on SSIM and PSNR algorithms is characterized by comprising the following steps:
acquiring an original image, and generating a first thumbnail according with the scene parameters of the original image according to the original image;
rendering the first thumbnail according to preset parameters to obtain a comparison graph;
acquiring the effect parameters of the comparison graph, and determining whether the effect parameters of the comparison graph accord with the preset parameters; if so, taking the comparison graph as an effect graph;
preliminarily rendering the original image based on the effect image and the preset parameters, and generating a conversion image for partially rendering the original image;
extracting a rendered part of the conversion graph to obtain a rendered graph;
acquiring the characteristic parameters of the rendering graph and the characteristic parameters of the effect graph and the corresponding part of the rendering graph, confirming whether the characteristic parameters of the rendering graph meet the characteristic parameters of the effect graph and the corresponding part of the rendering graph, and outputting judgment information;
the steps of obtaining the feature parameters of the rendering graph and the feature parameters of the effect graph and the corresponding part of the rendering graph, confirming whether the feature parameters of the rendering graph conform to the feature parameters of the effect graph and the corresponding part of the rendering graph, and outputting judgment information comprise:
acquiring a position value of a pixel of the effect image based on a pixel calculation method;
acquiring a characteristic value and a relative position information value of the effect graph according to the position value, wherein the characteristic value and the relative position information value are used as characteristic parameters of the effect graph;
acquiring the characteristic parameters of the rendering graph, confirming whether the characteristic parameters of the rendering graph meet the characteristic parameters of the effect graph and the corresponding part of the rendering graph, and outputting judgment information;
and determining whether to continue rendering the conversion graph according to the preset parameters or not based on the judgment information, and outputting a completely rendered graph.
2. The SSIM (Small Scale integration) and PSNR (Peak Signal to network Address translation) algorithm-based rendering method according to claim 1, wherein the step of obtaining an original image, and generating a first thumbnail image according to the original image and the scene parameters of the original image comprises:
acquiring an original image, and generating a primary thumbnail of the original image which is scaled in equal proportion;
acquiring scene parameters of the original image and scene parameters of the preliminary thumbnail;
determining whether the scene parameters of the preliminary thumbnail accord with the scene parameters of the original picture;
if so, taking the preliminary thumbnail as a first thumbnail;
and if not, re-rendering the preliminary thumbnail.
3. The rendering method based on SSIM and PSNR algorithm as claimed in claim 1, wherein an effect parameter of the contrast map is obtained, and it is determined whether the effect parameter conforms to the preset parameter; if yes, the step of using the comparison graph as an effect graph comprises the following steps:
obtaining an effect parameter of the comparison graph according to a ray tracing algorithm;
determining whether the effect parameters of the comparison graph meet the preset parameters;
if so, taking the comparison graph as an effect graph;
if not, re-rendering the comparison graph.
4. The SSIM (Small Scale integration) and PSNR (Peak Signal to network Address translation) algorithm based rendering method as claimed in claim 1, wherein the step of preliminarily rendering the original image based on the effect map and the preset parameters, and the step of generating a transformation map for partially rendering the original image comprises:
primarily rendering the original image according to the preset parameters after the effect image is obtained;
and acquiring a conversion map for partially rendering the original image based on preset time.
5. The SSIM (Small Scale integration) and PSNR (Peak Signal to noise ratio) algorithm-based rendering method as claimed in claim 1, wherein the steps of obtaining the feature parameters of the rendering map, determining whether the feature parameters of the rendering map conform to the feature parameters of the corresponding part of the effect map and the rendering map, and outputting the determination information comprise:
acquiring the characteristic parameters of the rendering graph by using SSIM and PSNR algorithms according to the characteristic parameters of the effect graph, and generating the similarity between the effect graph and the rendering graph;
and determining whether the characteristic parameters of the rendering graph accord with the characteristic parameters of the corresponding parts of the effect graph and the rendering graph or not based on the similarity, and outputting judgment information.
6. The SSIM (Small Scale integration) and PSNR (Peak Signal to noise ratio) algorithm-based rendering method as claimed in claim 5, wherein the step of determining whether the feature parameters of the rendering map are consistent with the feature parameters of the corresponding part of the effect map and the rendering map based on the similarity, and outputting the decision information comprises:
if the similarity is in a preset first threshold value, which indicates that the characteristic parameters of the rendering graph accord with the characteristic parameters of the effect graph, continuing to render the conversion graph;
and if the similarity is in a preset second threshold value, which indicates that the characteristic parameters of the rendering graph do not accord with the characteristic parameters of the effect graph, re-rendering the conversion graph.
7. The SSIM (Small Scale integration) and PSNR (Peak Signal to noise ratio) algorithm-based rendering method as claimed in claim 1, wherein the step of determining whether to continue rendering the transformation graph according to the preset parameters based on the determination information and outputting a fully rendered graph comprises:
if the characteristic parameters of the rendering graph accord with the characteristic parameters of the effect graph, continuing to render the conversion graph according to the preset parameters;
obtaining a chart according to the completely rendered conversion chart;
generating a second thumbnail using an image scaling algorithm based on the map;
obtaining a comparison result of the second thumbnail and the effect graph according to SSIM and PSNR algorithms;
and outputting the chart and the comparison result.
8. Rendering system based on SSIM, PSNR algorithm, applied to rendering method based on SSIM, PSNR algorithm as claimed in any of claims 1 to 7, characterized in that it comprises,
the generation module 1 is used for acquiring an original image and generating a first thumbnail according with the scene parameters of the original image according to the original image;
the comparison image obtaining module 2 is used for rendering the first thumbnail according to preset parameters to obtain comparison;
the effect graph obtaining module 3 is configured to obtain an effect parameter of the comparison graph, and determine whether the effect parameter of the comparison graph meets the preset parameter; if so, taking the comparison graph as an effect graph;
the conversion map obtaining module 4 is configured to preliminarily render the original map based on the effect map and the preset parameter, and generate a conversion map for partially rendering the original map;
a rendering map acquisition module 5, configured to extract a rendered portion of the conversion map to obtain a rendering map;
a rendering graph detection module 6, configured to obtain the feature parameters of the rendering graph and the feature parameters of the corresponding portions of the effect graph and the rendering graph, determine whether the feature parameters of the rendering graph meet the feature parameters of the corresponding portions of the effect graph and the rendering graph, and output determination information;
and the formed graph output module 7 is configured to determine whether to continue rendering the conversion graph according to the preset parameter based on the determination information, and output a completely rendered formed graph.
9. A computer-readable storage medium, in which a computer program is stored which can be loaded by a processor and which executes the method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111407147.9A CN114119842B (en) | 2021-11-24 | 2021-11-24 | Rendering method and system based on SSIM (structural similarity) and PSNR (Peak Signal to noise ratio) algorithm and computer readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111407147.9A CN114119842B (en) | 2021-11-24 | 2021-11-24 | Rendering method and system based on SSIM (structural similarity) and PSNR (Peak Signal to noise ratio) algorithm and computer readable storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114119842A CN114119842A (en) | 2022-03-01 |
| CN114119842B true CN114119842B (en) | 2022-09-06 |
Family
ID=80375489
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111407147.9A Active CN114119842B (en) | 2021-11-24 | 2021-11-24 | Rendering method and system based on SSIM (structural similarity) and PSNR (Peak Signal to noise ratio) algorithm and computer readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114119842B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2008078942A (en) * | 2006-09-20 | 2008-04-03 | Konica Minolta Business Technologies Inc | Device, method and program for generating thumbnail |
| CN108287877A (en) * | 2018-01-02 | 2018-07-17 | 西安交通大学 | A kind of RIB renders compressed file FPGA compression/decompression compression systems and hardware decompressing method |
| CN110910470A (en) * | 2019-11-11 | 2020-03-24 | 广联达科技股份有限公司 | Method and device for generating high-quality thumbnail |
| CN110930307A (en) * | 2019-10-31 | 2020-03-27 | 北京视博云科技有限公司 | Image processing method and device |
| CN112434185A (en) * | 2020-10-26 | 2021-03-02 | 国家广播电视总局广播电视规划院 | Method, system, server and storage medium for searching similar video clips |
| CN112950738A (en) * | 2021-03-30 | 2021-06-11 | 杭州群核信息技术有限公司 | Rendering engine processing method and device, storage medium and electronic equipment |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7069506B2 (en) * | 2001-08-08 | 2006-06-27 | Xerox Corporation | Methods and systems for generating enhanced thumbnails |
| JP2009271831A (en) * | 2008-05-09 | 2009-11-19 | Sharp Corp | Image data processing apparatus |
| US11409791B2 (en) * | 2016-06-10 | 2022-08-09 | Disney Enterprises, Inc. | Joint heterogeneous language-vision embeddings for video tagging and search |
| CN108876912A (en) * | 2018-06-30 | 2018-11-23 | 浙江同创空间技术有限公司 | Three-dimensional scenic physics renders method and its system |
-
2021
- 2021-11-24 CN CN202111407147.9A patent/CN114119842B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2008078942A (en) * | 2006-09-20 | 2008-04-03 | Konica Minolta Business Technologies Inc | Device, method and program for generating thumbnail |
| CN108287877A (en) * | 2018-01-02 | 2018-07-17 | 西安交通大学 | A kind of RIB renders compressed file FPGA compression/decompression compression systems and hardware decompressing method |
| CN110930307A (en) * | 2019-10-31 | 2020-03-27 | 北京视博云科技有限公司 | Image processing method and device |
| CN110910470A (en) * | 2019-11-11 | 2020-03-24 | 广联达科技股份有限公司 | Method and device for generating high-quality thumbnail |
| CN112434185A (en) * | 2020-10-26 | 2021-03-02 | 国家广播电视总局广播电视规划院 | Method, system, server and storage medium for searching similar video clips |
| CN112950738A (en) * | 2021-03-30 | 2021-06-11 | 杭州群核信息技术有限公司 | Rendering engine processing method and device, storage medium and electronic equipment |
Non-Patent Citations (3)
| Title |
|---|
| ExploreMaps: Efficient construction and ubiquitous exploration of panoramic view graphs of complex 3D environments;M. Di Benedetto等;《computer graphics forum》;20140601;第33卷(第2期);459-468 * |
| 基于形状检索的场景图像三维建模;樊亚春等;《高技术通讯》;20130815;第23卷(第08期);781-788 * |
| 基于有机计算的动漫渲染集群系统管理技术的研究与应用;王钰等;《中国优秀博硕士学位论文全文数据库(硕士)信息科技辑》;20120215(第2期);I138-1877 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114119842A (en) | 2022-03-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111028213B (en) | Image defect detection method, device, electronic equipment and storage medium | |
| WO2022170706A1 (en) | Defect detection method and apparatus for mold monitoring, and device and medium | |
| KR20230124713A (en) | Fault detection methods, devices and systems | |
| CN110596120A (en) | Glass boundary defect detection method, device, terminal and storage medium | |
| CN114693633A (en) | Welding defect detection method, device and system | |
| CN112730251B (en) | Device and method for detecting screen color defects | |
| CN114998352A (en) | Production equipment fault detection method based on image processing | |
| CN108871185B (en) | Method, device and equipment for detecting parts and computer readable storage medium | |
| CN116167932A (en) | Image quality optimization method, device, equipment and storage medium | |
| CN114119842B (en) | Rendering method and system based on SSIM (structural similarity) and PSNR (Peak Signal to noise ratio) algorithm and computer readable storage medium | |
| WO2024001309A1 (en) | Method and apparatus for generating and producing template for infrared thermal image analysis report | |
| CN116563262A (en) | Building crack detection algorithm based on multiple modes | |
| US20240005477A1 (en) | Index selection device, information processing device, information processing system, inspection device, inspection system, index selection method, and index selection program | |
| CN115619796A (en) | Method and device for obtaining photovoltaic module template and nonvolatile storage medium | |
| CN115374517A (en) | Testing method and device for wiring software, electronic equipment and storage medium | |
| CN111669575B (en) | Method, system, electronic device, medium and terminal for testing image processing effect | |
| CN112052727B (en) | Portrait recognition analysis system and method based on big data | |
| CN114782822A (en) | Method and device for detecting state of power equipment, electronic equipment and storage medium | |
| CN114092542A (en) | Bolt measuring method and system based on two-dimensional vision | |
| CN113139454B (en) | Road width extraction method and device based on single image | |
| CN113447572B (en) | Steel rail flaw detection method, electronic device, steel rail flaw detection vehicle and readable storage medium | |
| CN117036313B (en) | Computer vision-based infrared intelligent detection method and equipment for power equipment | |
| CN114418941B (en) | Defect diagnosis method and system based on detection data of power inspection equipment | |
| US20230401691A1 (en) | Image defect detection method, electronic device and readable storage medium | |
| CN117058768A (en) | Signature quality inspection method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CP01 | Change in the name or title of a patent holder | ||
| CP01 | Change in the name or title of a patent holder |
Address after: 518051 floor 17, block B, sunshine science and innovation center, No. 2, Shanghua Road, Nanshan street, Nanshan District, Shenzhen, Guangdong Province Patentee after: Shenzhen Ruiyun Technology Co.,Ltd. Address before: 518051 floor 17, block B, sunshine science and innovation center, No. 2, Shanghua Road, Nanshan street, Nanshan District, Shenzhen, Guangdong Province Patentee before: SHENZHEN RAYVISION TECHNOLOGY CO.,LTD. |