WO2022196200A1 - 情報処理装置、情報処理方法、情報処理プログラム及び情報処理システム - Google Patents
情報処理装置、情報処理方法、情報処理プログラム及び情報処理システム Download PDFInfo
- Publication number
- WO2022196200A1 WO2022196200A1 PCT/JP2022/005356 JP2022005356W WO2022196200A1 WO 2022196200 A1 WO2022196200 A1 WO 2022196200A1 JP 2022005356 W JP2022005356 W JP 2022005356W WO 2022196200 A1 WO2022196200 A1 WO 2022196200A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- rendering
- rendered image
- information processing
- resolution
- restoration
- Prior art date
Links
- 230000010365 information processing Effects 0.000 title claims abstract description 77
- 238000003672 processing method Methods 0.000 title claims description 5
- 238000009877 rendering Methods 0.000 claims abstract description 318
- 230000003044 adaptive effect Effects 0.000 claims abstract description 106
- 238000005070 sampling Methods 0.000 claims description 86
- 230000005540 biological transmission Effects 0.000 claims description 9
- 230000002123 temporal effect Effects 0.000 claims description 6
- 238000013528 artificial neural network Methods 0.000 claims description 3
- 238000000034 method Methods 0.000 description 11
- 238000004364 calculation method Methods 0.000 description 7
- 238000006243 chemical reaction Methods 0.000 description 5
- 230000009467 reduction Effects 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 230000006835 compression Effects 0.000 description 2
- 238000007906 compression Methods 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000000342 Monte Carlo simulation Methods 0.000 description 1
- 238000012614 Monte-Carlo sampling Methods 0.000 description 1
- 241000188156 Tamu Species 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- 239000003086 colorant Substances 0.000 description 1
- 230000007423 decrease Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000009792 diffusion process Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T15/00—3D [Three Dimensional] image rendering
- G06T15/005—General purpose rendering architectures
-
- 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 transformations in the plane of the image
- G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
- G06T3/4046—Scaling of whole images or parts thereof, e.g. expanding or contracting using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T3/00—Geometric image transformations in the plane of the image
- G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
- G06T3/4053—Scaling of whole images or parts thereof, e.g. expanding or contracting based on super-resolution, i.e. the output image resolution being higher than the sensor resolution
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/60—Image enhancement or restoration using machine learning, e.g. neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20004—Adaptive image processing
- G06T2207/20012—Locally adaptive
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2210/00—Indexing scheme for image generation or computer graphics
- G06T2210/36—Level of detail
Definitions
- the present disclosure relates to an information processing device, an information processing method, an information processing program, and an information processing system that render model data by ray tracing.
- General path tracing uses the Monte Carlo method, which randomly calculates a fixed number of samples for each pixel. Typically, there are 1000 ray traces per pixel (1000 SPP), so rendering a 4K image would require computation of 4000 ⁇ 2000 ⁇ 1000 rays.
- Non-Patent Document 1 pre-rendering is first performed with 1SPP to create a noisy image, and the noisy image and its denoised image are used as inputs to learn a DNN that predicts a sampling map.
- This sampling map is an output as an image of how difficult it is to render that pixel (that is, how many SPPs are required to render that pixel well).
- the actual rendering is performed based on the sampling map.
- the number of SPPs at this time is an optimum value for each pixel, which is larger than that for pre-rendering (1 SPP) but far smaller than normal (1000 SPP).
- a DNN that removes noise by inputting the rendered image output last is learned.
- two DNNs, the sampling map prediction DNN and the denoising DNN are optimally interlocked and learned so that the error in the final result is reduced.
- Patent Literature 1 like Non-Patent Literature 1, discloses adaptive sampling configured to learn a DNN that predicts a sampling map from a noisy image created by pre-rendering at a low SPP (1SPP) and an image obtained by denoising it. do.
- Patent Document 2 discloses a method of displaying in low resolution at first and gradually increasing the resolution in rendering via a network.
- Patent Document 2 only the resolution is changed, restoration processing is not performed, and the rendering reduction rate is simply obtained from the transmission bandwidth and the rendering speed.
- Patent Document 3 discloses that in HMD (Head Mounted Display), MSAA (Multi-Sample Anti-Aliasing) rendering method is used when rendering with different resolutions depending on the position from the center of the pixel and the importance of the subject. anti-aliasing.
- MSAA Multi-Sample Anti-Aliasing rendering method
- the restoration is simply an accumulation average, and the rendering reduction rate is not predicted according to the Loss after restoration.
- Non-Patent Document 1 and Patent Document 1 a sampling map that adaptively controls only SPP is predicted. However, even if only the SPP is adaptively changed, there is a limit to the effect of reducing the computation time.
- the purpose of the present disclosure is to render model data with high quality and in a short time by ray tracing.
- An information processing device includes: a pre-rendering unit that pre-renders model data by ray tracing to create a pre-rendered image; a prediction unit that predicts the degree of difficulty of restoration in the pre-rendered image; A rendering condition determination unit that determines rendering conditions that specify the resolution and SPP (Sample Per Pixel) of each element in the pre-rendered image based on the difficulty level of restoration, and creates an adaptive control signal that sets the rendering conditions.
- a pre-rendering unit that pre-renders model data by ray tracing to create a pre-rendered image
- a prediction unit that predicts the degree of difficulty of restoration in the pre-rendered image
- a rendering condition determination unit that determines rendering conditions that specify the resolution and SPP (Sample Per Pixel) of each element in the pre-rendered image based on the difficulty level of restoration, and creates an adaptive control signal that sets the rendering conditions.
- a rendering unit that creates an adaptive rendering image by rendering the model data by ray tracing according to the rendering condition for each element set in the adaptive control signal
- a rendered image restoration unit that restores the adaptive rendered image by super-resolution and denoising to create a final rendered image.
- the SPP not only the SPP but also the resolution are adaptively controlled for each element. This makes it possible to predict the optimal combination of SPP and resolution for each element, adaptively render, restore (denoise and super-resolution), and output the final rendered image at high speed while maintaining image quality. .
- the rendering conditions further specify the number of images, the number of bounces, the number of refractions of internal transmission, random number sequences of noise, bit depth, temporal resolution, on/off of light component, on/off of antialiasing and/or number of subsamples. You may
- the rendering condition determination unit may determine the rendering conditions based on image processing, points of interest, importance of subjects, and/or display information.
- Condition prediction can be performed more easily and lightly. Also, by combining with the condition prediction DNN, the accuracy of the rendering condition can be improved.
- the rendering condition determining unit may determine the rendering condition for each element, for each pixel, for each patch containing a plurality of pixels, or for each object area.
- the accuracy of the adaptive control signal can be improved and the continuity between adjacent pixels can be maintained.
- the rendering conditions for each object it is possible to determine the rendering conditions that follow the edge and have less protrusion. It can improve the prediction accuracy of rendering conditions and reduce computation time.
- the rendering condition determination unit determines a rendering condition for each of the partial elements, Other rendering conditions for each element may be predetermined.
- the calculation time can be reduced, the output processing of each frame can be accelerated, and real-time rendering can be achieved.
- the prediction unit inputs the pre-rendered image to a conditional prediction DNN (Deep Neural Network) to predict the difficulty level of restoration for each element
- the rendered image restoring unit may generate the final rendered image by inputting the adaptive rendered image and the adaptive control signal to a restored DNN learned simultaneously with the conditional prediction DNN.
- conditional prediction coefficients for uniformly outputting the target SPP and target resolution for the entire screen. Then, an image obtained by rendering with the uniform target SPP and target resolution for the entire screen is used as a student, and an image restoration coefficient for restoration (denoising and super-resolution) for predicting a teacher image is learned. As a result, learning of the conditional prediction DNN and the restoration DNN can be performed at the same time.
- the prediction unit predicts a sampling map indicating a difficulty level of restoration for each element in the pre-rendered image
- the rendering condition determining unit may create the adaptive control signal based on the sampling map.
- an appropriate adaptive control signal can be created according to the difficulty of restoration. For example, elements with high recovery difficulty can create adaptive control signals for relatively high SPP and high resolution rendering conditions. As a result, it is possible to output a final rendered image with image quality equivalent to that of the goal image.
- the prediction unit predicts a resolution sampling map and an SPP sampling map
- the rendering condition determining unit may set the rendering condition for the resolution based on the sampling map for the resolution, and set the rendering condition for the SPP based on the sampling map for the SPP.
- the rendering condition determination unit may specify the resolution and SPP without performing any particular conversion.
- the prediction unit predicts the one-dimensional sampling map
- the rendering condition determining unit may set a resolution rendering condition and an SPP rendering condition based on the one-dimensional sampling map.
- the rendering condition determining unit may specify a combination of resolution and SPP according to an arbitrary conversion formula from the predicted one-dimensional sampling map.
- the SPP of the pre-rendered image may be lower than the SPP of the final rendered image.
- the resolution of the pre-rendered image may be lower than the resolution of the final rendered image.
- An information processing method includes: Pre-render the model data by ray tracing to create a pre-rendered image, predicting the difficulty of restoration in the pre-rendered image; Based on the degree of difficulty of restoration, determine a rendering condition that specifies the resolution and SPP (Sample Per Pixel) of each element in the pre-rendered image, and create an adaptive control signal that sets the rendering condition; creating an adaptive rendering image by rendering the model data by ray tracing according to the rendering condition for each element set in the adaptive control signal; The adaptive rendered image is reconstructed by super-resolution and denoising to create a final rendered image.
- SPP Sample Per Pixel
- An information processing program includes the processor of the information processing device, a pre-rendering unit that pre-renders model data by ray tracing to create a pre-rendered image; a prediction unit that predicts the degree of difficulty of restoration in the pre-rendered image; A rendering condition determination unit that determines rendering conditions that specify the resolution and SPP (Sample Per Pixel) of each element in the pre-rendered image based on the difficulty level of restoration, and creates an adaptive control signal that sets the rendering conditions.
- SPP Sample Per Pixel
- a rendering unit that creates an adaptive rendering image by rendering the model data by ray tracing according to the rendering condition for each element set in the adaptive control signal; It operates as a rendered image restoration unit that restores the adaptive rendered image by performing super-resolution and denoising to create a final rendered image.
- An information processing system includes: a pre-rendering unit that pre-renders model data by ray tracing to create a pre-rendered image; a prediction unit that predicts the degree of difficulty of restoration in the pre-rendered image; A rendering condition determination unit that determines rendering conditions that specify the resolution and SPP (Sample Per Pixel) of each element in the pre-rendered image based on the difficulty level of restoration, and creates an adaptive control signal that sets the rendering conditions.
- a pre-rendering unit that pre-renders model data by ray tracing to create a pre-rendered image
- a prediction unit that predicts the degree of difficulty of restoration in the pre-rendered image
- a rendering condition determination unit that determines rendering conditions that specify the resolution and SPP (Sample Per Pixel) of each element in the pre-rendered image based on the difficulty level of restoration, and creates an adaptive control signal that sets the rendering conditions.
- a rendering unit that creates an adaptive rendering image by rendering the model data by ray tracing according to the rendering condition for each element set in the adaptive control signal
- a rendered image restoration unit that restores the adaptive rendered image by super-resolution and denoising to create a final rendered image.
- FIG. 1 shows a configuration of an information processing device according to an embodiment of the present disclosure
- 4 shows an operation flow of the information processing apparatus
- FIG. 1 shows the configuration of an information processing device according to an embodiment of the present disclosure.
- the information processing device 100 is, for example, a device that renders images for displaying on a 3D display capable of displaying 3D images.
- the information processing device 100 is built in or externally connected to a 3D display, for example.
- the information processing apparatus 100 loads an information processing program recorded in a ROM into a RAM and executes it, thereby obtaining a pre-rendering unit 101, a sampling map prediction unit 102, a rendering condition determination unit 103, a rendering unit 104, and a rendered image. It operates as the restoration unit 105 .
- FIG. 2 shows the operation flow of the information processing device.
- the information processing apparatus 100 repeatedly executes the processes from step S101 onward for each frame.
- Step S101 Read model data
- the pre-rendering unit 101 reads model data input as data to be rendered.
- the model data is, for example, 3D CG model data.
- Step S102 Execute pre-rendering
- the pre-rendering unit 101 pre-renders the input model data by ray tracing to create a pre-rendered image.
- the pre-rendering unit 101 pre-renders the model data, for example, at the same resolution as the output resolution (that is, the resolution of the final rendering image to be output) and at a low SPP of about 1 SPP.
- the pre-rendering unit 101 may pre-render the model data at a resolution lower than the output resolution (eg, 1 ⁇ 4 or 1/16) and/or at an SPP greater than one.
- a pre-rendered image created by the pre-rendering unit 101 is a noisy rendered image.
- the pre-rendering unit 101 may also create various other AOV (Arbitrary Output Variable) images (that is, images for each element such as depth, normal, albedo, diffusion, and reflection).
- AOV Arbitrary Output Variable
- the pre-rendering unit 101 inputs the pre-rendering image to the sampling map prediction unit 102 .
- the sampling map prediction unit 102 is a conditional prediction DNN that predicts a sampling map representing the degree of rendering difficulty (difficulty) from the input noisy pre-rendered image.
- Step S103 Cut out patches from the pre-rendered image
- the sampling map prediction unit 102 scans the pre-rendered image and cuts out a plurality of patches from the pre-rendered image.
- the patch size is equal to the input patch size of the conditional predictive DNN.
- the sampling map prediction unit 102 may cut out patches so as to sequentially raster scan from the upper left of the pre-rendered image.
- Step S104 Input the patch to the conditional prediction DNN and predict the sampling map
- the sampling map prediction unit 102 inputs the extracted patch to the conditional prediction DNN and predicts the sampling map.
- the sampling map indicates the rendering difficulty (difficulty). In other words, the sampling map prediction unit 102 predicts the difficulty level of restoration in the pre-rendered image.
- a sampling map prediction unit 102 predicts a sampling map using pre-learned conditional prediction coefficients 106 . When learning is performed using not only pre-rendered images but also various AOV images, the sampling map prediction unit 102 also cuts out patches of the corresponding AOV images and inputs them to the conditional prediction DNN.
- condition prediction coefficient 106 will be explained.
- the conditional prediction coefficients 106 are trained with high SPP and high resolution rendered images created from a large amount of CG models, and are trained with low SPP (eg, 1 SPP) and low resolution (eg, 1/4 and 1/16) rendered images. , and a restored image restored (denoised and super-resolved) by the restored DNN are used as inputs for learning.
- a specific learning procedure for the conditional prediction coefficients 106 will be described.
- a conditional prediction coefficient 106 is set to uniformly output the target SPP (4SPP, for example) and the target resolution (4K, for example) for the entire screen.
- the image restoration (denoising and super-resolution) for predicting the teacher image is learned by learning the image restoration coefficient 107 for the restoration (denoising and super-resolution) for predicting the teacher image by using the result image of rendering with the uniform target SPP and target resolution for the entire screen as a student.
- Loss is the difference between the inference image output as a result of this learning and the teacher image.
- the condition prediction coefficient 106 is learned so as to reduce this loss.
- the conditional prediction coefficient 106 increases the SPP and resolution where the Loss is large and decreases the SPP and resolution where the Loss is small so that the average of the SPP and resolution of the entire screen becomes the target SPP and target resolution. Calculate the sampling map.
- NR also called noise reduction
- SR super resolution
- sampling map output by the sampling map prediction unit 102 is not limited to a one-dimensional map representing the difficulty of rendering. Anything is fine.
- the sampling map prediction unit 102 may predict a one-dimensional sampling map commonly used for resolution and SPP, or may predict the resolution sampling map and the SPP sampling map separately.
- further sampling maps may be predicted for further information (eg, image processing, points of interest, subject importance and/or display information).
- Step S105 Determine rendering conditions from the sampling map
- the rendering condition determination unit 103 determines rendering conditions for each element (for example, each pixel, each patch containing a plurality of pixels, or each object region) in the pre-rendered image based on a sampling map indicating the degree of difficulty of restoration. Rendering conditions specify the resolution and SPP for each element in the pre-rendered image.
- the rendering condition determination unit 103 creates an adaptive control signal that sets the determined rendering conditions. In other words, the rendering condition determination unit 103 calculates an adaptive control signal that sets actual rendering conditions for each element from the sampling map predicted by the sampling map prediction unit 102 .
- the rendering condition determination unit 103 may specify a combination of resolution and SPP according to an arbitrary conversion formula from the predicted one-dimensional sampling map.
- the rendering condition determination unit 103 may specify the resolution and SPP without performing any particular conversion.
- the rendering condition determination unit 103 may convert the sampling map according to an arbitrary conversion formula according to setting conditions (high speed, high resolution, etc.) input by the user (director or viewer).
- Step S106 Determine the rendering conditions for the full screen
- the rendering condition determination unit 103 determines rendering conditions for each element of the full screen of the pre-rendered image.
- Step S107 Execute rendering
- the rendering unit 104 creates an adaptive rendering image by rendering the model data by ray tracing according to the rendering conditions for each element set in the adaptive control signal. That is, the rendering unit 104 creates an adaptive rendering image by rendering the model data for each element with the resolution and SPP specified for each element by the adaptive control signal. In other words, the rendering unit 104 performs the main rendering according to the rendering conditions defined by the calculated adaptive control signal. In short, the rendering unit 104 performs rendering while changing the conditions for each element based on the locally optimum combination of SPP and resolution according to the adaptive control signal.
- the rendering unit 104 basically uses the same renderer (rendering software) as the pre-rendering unit 101, but may use another renderer (such as a renderer that performs more advanced ray calculation).
- Step S108 Cut out patches from the adaptive rendering image
- the rendered image restoration unit 105 scans the adaptive rendered image and cuts out a plurality of patches from the adaptive rendered image.
- the size of the patch is equal to the input patch size of the reconstructed DNN.
- the rendered image restoration unit 105 may cut out patches so as to sequentially perform raster scanning from the upper left of the adaptive rendered image.
- the input patch size of the restoration DNN may be equal to or different from the input patch size of the conditional prediction DNN.
- Step S109 Input the adaptive control signal and patch to the restoration DNN, and predict the output image
- the rendered image restoration unit 105 inputs the adaptive control signal and the patch extracted from the adaptive rendered image to the restoration DNN, and predicts the output image for each patch.
- the rendering conditions set in the adaptive control signals specify the resolution and SPP for each element.
- the decompressor DNN simultaneously handles two tasks of super-resolution and denoising.
- the rendered image restoration unit 105 restores the patch by super-resolution and denoising.
- the image restoration coefficients 107 are learned together with the conditional prediction coefficients 106 .
- the rendered image restoration unit 105 may use not only the adaptive rendered image but also sampling maps and various AOV images. In that case, it is necessary to carry out learning under those conditions.
- Step S110 Predict full-screen output image
- the rendering image restoration unit 105 predicts output images for all patches.
- Step S111 Output the final rendering result
- the rendered image restoration unit 105 connects the super-resolved and denoised output images of all patches to create and output the final rendering result.
- step S105 the rendering condition determination unit 103 determines rendering conditions for each element in the pre-rendered image from the sampling map, and creates adaptive control signals for setting the determined rendering conditions. Various variations of rendering conditions are described.
- the rendering condition determining unit 103 may specify other rendering conditions in addition to the SPP and resolution as rendering conditions.
- the rendering conditions may further include the number of images, the number of bounces, the number of internal transmission refractions, the noise random number sequence, the bit depth, the temporal resolution, the light component on/off, the antialiasing on/off, and/or the number of subsamples. May be specified. Different effects are required for the restored DNN depending on the adaptively changed rendering conditions. As a result, adaptive rendering can be performed more efficiently or under conditions that meet the user's wishes.
- the number of images means dividing the SPP in the direction of the number of images. For example, creating five sheets of 2SPP instead of 10SPP changes the ease of denoising.
- the restoration DNN becomes a DNN that performs not only super-resolution and denoising but also fusion of a plurality of images.
- the number of bounces means how many times a light ray hits an object and is reflected.
- the number of refractions for internal transmission means how many times the light rays are refracted on the way when they enter the transparent object.
- the noise random number sequence means switching between high-frequency-oriented or low-frequency-oriented random numbers.
- Ray tracing uses a technique called Monte Carlo sampling, in which a ray is sampled in a random direction each time it diverges or refracts, and the result is calculated when the ray travels in that direction.
- Monte Carlo sampling in which a ray is sampled in a random direction each time it diverges or refracts, and the result is calculated when the ray travels in that direction.
- white noise is used in random sampling, unevenness and fineness will occur locally.
- more advantageous rendering is performed by switching between high frequency emphasis, low frequency emphasis, and random numbers according to the characteristics of the region such as a flat portion and a complicated portion.
- Bit depth means whether the number of bits is full or reduced. Since collision calculations are performed many times in ray tracing, if the full number of bits is calculated each time, the amount of calculation becomes enormous, so unnecessary bits are reduced for each pixel.
- Temporal resolution means the frame rate during video rendering.
- On/off of light components means on/off of diffused light, reflected light and/or transmitted light components. In ray tracing, various components other than direct light are tracked, and each component is turned on/off for each pixel.
- Antialiasing on/off and the number of subsamples are antialiasing settings.
- ray tracing if rendering is simply performed for each pixel, jaggies occur in oblique lines and the like, so various types of anti-aliasing are performed.
- the representative methods of SSAA (Supersampling Anti-Aliasing) and MSAA (Multi Sample Anti-Aliasing) calculate four sub-sampling points for one pixel and mix them to obtain the final pixel value Ask for The anti-aliasing function is turned on/off and the number of sub-sampling points is changed for each pixel.
- the rendering condition determination unit 103 may determine rendering conditions using a method other than DNN-based learning. For example, the rendering condition determination unit 103 may determine rendering conditions and generate adaptive control signals by model-based signal processing or external settings. Specifically, the rendering condition determination unit 103 may determine rendering conditions based on image processing, points of interest, subject importance, and/or display information. As a result, condition prediction can be performed more easily and lightly. Also, by combining with the condition prediction DNN, the accuracy of the rendering condition can be improved.
- the rendering condition determination unit 103 may determine rendering conditions based on detection (edge, flatness, brightness, depth, motion amount) according to the difficulty level of rendering and restoration. Flat areas, dark areas, and blurred areas (due to out-of-focus, rapid motion, etc.) can be easily restored by super-resolution and denoising, even if the resolution and number of SPPs are small.
- Known model-based image processing can detect such regions and create adaptive control signals to simplify rendering.
- the rendering condition determining unit 103 may determine the rendering condition based on the point of interest and the importance of the subject. Points of interest and important objects can be rendered at high resolution and high SPP, and simplified elsewhere. Such an adaptive control signal can be generated by detecting it by image processing or externally designated by the user.
- the rendering condition determination unit 103 may determine rendering conditions based on user (director or viewer) preferences.
- the rendering condition determination unit 103 can be created by externally designating an adaptive control signal that changes the rendering condition. For example, if the resolution is important, the resolution can be allocated, and if the noise reduction is important, the SPP can be allocated more resources. Priorities can be set for temporal resolution, bit depth, anti-aliasing, and various components. In addition, in the case of 3D, it is possible to make settings such as emphasizing rendering on the effective side.
- the rendering condition determining unit 103 may determine the rendering condition based on various information when displaying 3D or the like.
- viewpoints such as the ends can be rendered with high resolution and high SPP, and viewpoint positions in between can be simplified and rendered.
- the resolution and SPP are increased for rendering in areas such as occlusion areas (areas on the back side of the solid) and screen edges where sufficient information cannot be obtained only from the two-end viewpoints.
- An adaptive control signal reflecting such a viewpoint position and the presence or absence of occlusion can be set by external setting or Visible calculation.
- the rendering condition determining unit 103 may determine the rendering condition according to the display characteristics of the display. For example, in accordance with the resolution of the display, rendering is performed at a resolution that is 1/4 or 1/16 times the resolution. For special displays such as 3D displays, rendering may be performed according to display characteristics other than resolution. For example, in a lenticular 3D display, aliasing and false colors occur depending on the phase relationship between the lenticular lens and the pixels of the panel. Therefore, it is possible to calculate optimal rendering conditions from these characteristics and set adaptive control signals.
- the rendering condition determination unit 103 may determine rendering conditions for each element in the pre-rendered image, for each pixel, for each patch containing a plurality of pixels, or for each object area. This improves the prediction accuracy of the rendering conditions and reduces the computation time. Basically, the rendering condition determining unit 103 may set rendering conditions for each pixel, but instead of pixels, rendering conditions may be set for each rectangular patch, for example.
- the sampling map prediction unit 102 calculates a sampling map using the conditional prediction DNN. Also, the rendering image restoration unit 105 restores (super-resolution, denoises) the adaptive rendering image using the restoration DNN. These conditional prediction DNN and restoration DNN process each rectangular patch extracted from the input image (pre-rendered image and adaptive rendered image).
- the rendering condition determination unit 103 may determine rendering conditions for each rectangular patch as each element in the pre-rendered image. By determining the rendering condition not for each pixel but for each region containing a plurality of pixels in this way, it is possible to improve the accuracy of the adaptive control signal and maintain the continuity between adjacent pixels.
- the size of this rectangular patch may not be the same as the patch size of the conditional prediction DNN and the restoration DNN.
- the patch of the conditional prediction DNN may be further divided into four and aggregated, or a completely different patch size may be adopted. Also, a plurality of sampling maps (adaptive control signals) may be multiplied.
- the rendering condition determining unit 103 may determine rendering conditions for each object region for each element in the pre-rendered image.
- the object area means a meaningful object area such as a person, unlike a mechanically divided rectangular patch.
- the rendering condition determination unit 103 divides a pre-rendered image pre-rendered in 1SPP into a plurality of object regions using an existing semantic segmentation technique, aggregates the rendering conditions for each pixel for each object region, and calculates the object You can set the rendering conditions for each.
- the rendering conditions for each object can be set with high precision, and the rendering conditions that follow the edge and have less protrusion can be determined.
- the rendering condition determining unit 103 may determine rendering conditions for some elements and determine rendering conditions for other elements in advance. Some of the rendering conditions can be pre-computed (before the start of the operational flow of FIG. 2) rather than being computed at pre-rendering time. As a result, the calculation time can be reduced, the speed of output processing for each frame can be increased, and real-time rendering can be realized. Such pre-rendering requires conditions at all times and viewpoints during animation and free-viewpoint generation. This pre-calculation can be performed for all conditions, or can be thinned out at arbitrary intervals. It is also possible to pre-calculate the times and viewpoints of high importance in advance. It is also possible to precalculate only places that do not change with time or viewpoint.
- the information processing apparatus 100 has a pre-rendering unit 101 , a sampling map prediction unit 102 , a rendering condition determination unit 103 , a rendering unit 104 and a rendering image restoration unit 105 .
- the server-side information processing device has a pre-rendering unit 101, a sampling map prediction unit 102, a rendering condition determination unit 103, and a rendering unit 104, and the client-side information processing device has a rendering image restoration unit 105. (not shown).
- the information processing device on the client side is, for example, built in or externally connected to the 3D display at the end user site.
- the video may be compressed during transmission.
- the rendering image is degraded by compression.
- the adaptive control signal at this time can be set not only for the entire screen uniformly according to the band, but also for each area according to the easiness of compression.
- the server-side information processing device has a pre-rendering unit 101, a sampling map prediction unit 102, and a rendering condition determination unit 103, and the client-side information processing device has a rendering unit 104 and a rendered image restoration unit 105.
- a system may be implemented (not shown).
- the information processing device on the client side is, for example, built in or externally connected to the 3D display at the end user site.
- adaptive control signals can also be transmitted at the same time. This allows for optimal adaptive rendering on the client side.
- optimal restoration processing is performed by transmitting adaptive control signals. be able to.
- this also applies to cases where super-resolution only, denoising only, or other signal processing is performed.
- a technique for predicting a sampling map that adaptively controls only the SPP is known.
- the resolution can be reduced to 1/4 or 1/16 to dramatically reduce computation time while maintaining image quality.
- not only the SPP but also the resolution are adaptively controlled for each element.
- the optimal combination of SPP and resolution is locally predicted, adaptively rendered, restored (denoising and super-resolution) by the restoration DNN, and the final rendered image with the same image quality as the goal image is output at high speed. can do.
- the present disclosure may have the following configurations.
- a pre-rendering unit that pre-renders model data by ray tracing to create a pre-rendered image
- a prediction unit that predicts the degree of difficulty of restoration in the pre-rendered image
- a rendering condition determination unit that determines rendering conditions that specify the resolution and SPP (Sample Per Pixel) of each element in the pre-rendered image based on the difficulty level of restoration, and creates an adaptive control signal that sets the rendering conditions.
- SPP Sample Per Pixel
- the rendering conditions further specify the number of images, the number of bounces, the number of refractions of internal transmission, random number sequences of noise, bit depth, temporal resolution, on/off of light component, on/off of antialiasing and/or number of subsamples.
- Information processing equipment (3) The information processing device according to (1) or (2) above, The information processing apparatus, wherein the rendering condition determining unit determines the rendering condition based on image processing, a point of interest, importance of a subject, and/or display information.
- the information processing device determines the rendering condition for each element, for each pixel, for each patch including a plurality of pixels, or for each object region.
- the rendering condition determination unit determines a rendering condition for each of the partial elements, A rendering condition for each of the other elements is determined in advance. Information processing apparatus.
- the prediction unit inputs the pre-rendered image to a conditional prediction DNN (Deep Neural Network) to predict the difficulty level of restoration for each element
- the rendered image restoration unit inputs the adaptive rendered image and the adaptive control signal to a restoration DNN learned simultaneously with the conditional prediction DNN to create the final rendered image.
- the prediction unit predicts a sampling map indicating a difficulty level of restoration for each element in the pre-rendered image
- the information processing apparatus, wherein the rendering condition determination unit creates the adaptive control signal based on the sampling map.
- the prediction unit predicts a resolution sampling map and an SPP sampling map
- the said rendering condition determination part sets the rendering condition of resolution based on the sampling map of said resolution, and sets the rendering condition of SPP based on said sampling map of SPP.
- the prediction unit predicts the one-dimensional sampling map
- the information processing apparatus wherein the rendering condition determination unit sets a resolution rendering condition and an SPP rendering condition based on the one-dimensional sampling map.
- the information processing device according to any one of (1) to (10) above, The information processing apparatus, wherein the resolution of the pre-rendered image is lower than the resolution of the final rendered image.
- (12) Pre-render the model data by ray tracing to create a pre-rendered image, predicting the difficulty of restoration in the pre-rendered image; Based on the degree of difficulty of restoration, determine a rendering condition that specifies the resolution and SPP (Sample Per Pixel) of each element in the pre-rendered image, and create an adaptive control signal that sets the rendering condition; creating an adaptive rendering image by rendering the model data by ray tracing according to the rendering condition for each element set in the adaptive control signal; An information processing method for reconstructing the adaptive rendered image by super-resolution and denoising to create a final rendered image.
- SPP Sample Per Pixel
- a pre-rendering unit that pre-renders model data by ray tracing to create a pre-rendered image
- a prediction unit that predicts the degree of difficulty of restoration in the pre-rendered image
- a rendering condition determination unit that determines rendering conditions that specify the resolution and SPP (Sample Per Pixel) of each element in the pre-rendered image based on the difficulty level of restoration, and creates an adaptive control signal that sets the rendering conditions.
- a rendering unit that creates an adaptive rendering image by rendering the model data by ray tracing according to the rendering condition for each element set in the adaptive control signal;
- An information processing program that operates as a rendered image restoration unit that restores the adaptive rendered image by performing super-resolution and denoising to create a final rendered image.
- a pre-rendering unit that pre-renders model data by ray tracing to create a pre-rendered image; a prediction unit that predicts the degree of difficulty of restoration in the pre-rendered image;
- a rendering condition determination unit that determines rendering conditions that specify the resolution and SPP (Sample Per Pixel) of each element in the pre-rendered image based on the difficulty level of restoration, and creates an adaptive control signal that sets the rendering conditions.
- a rendering unit that creates an adaptive rendering image by rendering the model data by ray tracing according to the rendering condition for each element set in the adaptive control signal;
- a rendered image restoration unit that restores the adaptive rendered image by performing super-resolution and denoising to create a final rendered image.
- the processor of the information processing device a pre-rendering unit that pre-renders model data by ray tracing to create a pre-rendered image; a prediction unit that predicts the degree of difficulty of restoration in the pre-rendered image;
- a rendering condition determination unit that determines rendering conditions that specify the resolution and SPP (Sample Per Pixel) of each element in the pre-rendered image based on the difficulty level of restoration, and creates an adaptive control signal that sets the rendering conditions.
- a rendering unit that creates an adaptive rendering image by rendering the model data by ray tracing according to the rendering condition for each element set in the adaptive control signal
- a non-transitory computer-readable recording medium recording an information processing program that operates as a rendered image restoration unit that restores the adaptive rendered image by super-resolution and denoising to create a final rendered image.
- Pre-rendering unit 101 Sampling map prediction unit 102
- Rendering condition determination unit 103 Rendering unit 104 Rendered image restoration unit 105 condition prediction coefficient 106 Image restoration coefficient 107
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Graphics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Image Generation (AREA)
- Image Processing (AREA)
Abstract
Description
モデルデータをレイトレーシングによりプリレンダリングしてプリレンダリング画像を作成するプリレンダリング部と、
前記プリレンダリング画像内の復元の難易度を予測する予測部と、
前記復元の難易度に基づき、前記プリレンダリング画像内の要素毎の解像度及びSPP(Sample Per Pixel)を指定するレンダリング条件を決定し、前記レンダリング条件を設定するアダプティブ制御信号を作成するレンダリング条件決定部と、
前記アダプティブ制御信号に設定された前記要素毎の前記レンダリング条件に従って、前記モデルデータをレイトレーシングによりレンダリングすることにより、アダプティブレンダリング画像を作成するレンダリング部と、
前記アダプティブレンダリング画像を超解像及びデノイズすることにより復元して最終レンダリング画像を作成するレンダリング画像復元部と
を具備する。
他の前記要素毎のレンダリング条件は、予め決定されてもよい。
前記レンダリング画像復元部は、前記条件予測DNNと同時に学習された復元DNNに前記アダプティブレンダリング画像及び前記アダプティブ制御信号を入力して前記最終レンダリング画像を作成してもよい。
前記レンダリング条件決定部は、前記サンプリングマップに基づき、前記アダプティブ制御信号を作成してもよい。
前記レンダリング条件決定部は、前記解像度のサンプリングマップに基づき解像度のレンダリング条件を設定し、前記SPPのサンプリングマップに基づきSPPのレンダリング条件を設定してもよい。
前記レンダリング条件決定部は、前記1次元のサンプリングマップに基づき解像度のレンダリング条件及びSPPのレンダリング条件を設定してもよい。
モデルデータをレイトレーシングによりプリレンダリングしてプリレンダリング画像を作成し、
前記プリレンダリング画像内の復元の難易度を予測し、
前記復元の難易度に基づき、前記プリレンダリング画像内の要素毎の解像度及びSPP(Sample Per Pixel)を指定するレンダリング条件を決定し、前記レンダリング条件を設定するアダプティブ制御信号を作成し、
前記アダプティブ制御信号に設定された前記要素毎の前記レンダリング条件に従って、前記モデルデータをレイトレーシングによりレンダリングすることにより、アダプティブレンダリング画像を作成し、
前記アダプティブレンダリング画像を超解像及びデノイズすることにより復元して最終レンダリング画像を作成する。
情報処理装置のプロセッサを、
モデルデータをレイトレーシングによりプリレンダリングしてプリレンダリング画像を作成するプリレンダリング部と、
前記プリレンダリング画像内の復元の難易度を予測する予測部と、
前記復元の難易度に基づき、前記プリレンダリング画像内の要素毎の解像度及びSPP(Sample Per Pixel)を指定するレンダリング条件を決定し、前記レンダリング条件を設定するアダプティブ制御信号を作成するレンダリング条件決定部と、
前記アダプティブ制御信号に設定された前記要素毎の前記レンダリング条件に従って、前記モデルデータをレイトレーシングによりレンダリングすることにより、アダプティブレンダリング画像を作成するレンダリング部と、
前記アダプティブレンダリング画像を超解像及びデノイズすることにより復元して最終レンダリング画像を作成するレンダリング画像復元部と
して動作させる。
モデルデータをレイトレーシングによりプリレンダリングしてプリレンダリング画像を作成するプリレンダリング部と、
前記プリレンダリング画像内の復元の難易度を予測する予測部と、
前記復元の難易度に基づき、前記プリレンダリング画像内の要素毎の解像度及びSPP(Sample Per Pixel)を指定するレンダリング条件を決定し、前記レンダリング条件を設定するアダプティブ制御信号を作成するレンダリング条件決定部と、
前記アダプティブ制御信号に設定された前記要素毎の前記レンダリング条件に従って、前記モデルデータをレイトレーシングによりレンダリングすることにより、アダプティブレンダリング画像を作成するレンダリング部と、
前記アダプティブレンダリング画像を超解像及びデノイズすることにより復元して最終レンダリング画像を作成するレンダリング画像復元部と
を具備する。
モデルデータをレイトレーシングによりプリレンダリングしてプリレンダリング画像を作成するプリレンダリング部と、
前記プリレンダリング画像内の復元の難易度を予測する予測部と、
前記復元の難易度に基づき、前記プリレンダリング画像内の要素毎の解像度及びSPP(Sample Per Pixel)を指定するレンダリング条件を決定し、前記レンダリング条件を設定するアダプティブ制御信号を作成するレンダリング条件決定部と、
前記アダプティブ制御信号に設定された前記要素毎の前記レンダリング条件に従って、前記モデルデータをレイトレーシングによりレンダリングすることにより、アダプティブレンダリング画像を作成するレンダリング部と、
前記アダプティブレンダリング画像を超解像及びデノイズすることにより復元して最終レンダリング画像を作成するレンダリング画像復元部と
を具備する情報処理装置。
(2)
上記(1)に記載の情報処理装置であって、
前記レンダリング条件は、さらに、画像枚数、バウンス数、内部透過の屈折回数、ノイズの乱数列、ビット深度、時間解像度、光成分のオン/オフ、アンチエイリアスのオン/オフ及び/又はサブサンプル数を指定する
情報処理装置。
(3)
上記(1)又は(2)に記載の情報処理装置であって、
前記レンダリング条件決定部は、画像処理、注目点、被写体の重要度及び/又は表示情報に基づき前記レンダリング条件を決定する
情報処理装置。
(4)
上記(1)乃至(3)の何れかに記載の情報処理装置であって、
前記レンダリング条件決定部は、前記要素毎として、画素毎、複数の画素を含むパッチ毎又はオブジェクト領域毎の前記レンダリング条件を決定する
情報処理装置。
(5)
上記(1)乃至(4)の何れかに記載の情報処理装置であって、
前記レンダリング条件決定部は、一部の前記要素毎のレンダリング条件を決定し、
他の前記要素毎のレンダリング条件は、予め決定される
情報処理装置。
(6)
上記(1)乃至(5)の何れかに記載の情報処理装置であって、
前記予測部は、条件予測DNN(Deep Neural Network)に前記プリレンダリング画像を入力して前記要素毎の復元の難易度を予測し、
前記レンダリング画像復元部は、前記条件予測DNNと同時に学習された復元DNNに前記アダプティブレンダリング画像及び前記アダプティブ制御信号を入力して前記最終レンダリング画像を作成する
情報処理装置。
(7)
上記(1)乃至(6)の何れかに記載の情報処理装置であって、
前記予測部は、前記プリレンダリング画像内の前記要素毎に復元の難易度を示すサンプリングマップを予測し、
前記レンダリング条件決定部は、前記サンプリングマップに基づき、前記アダプティブ制御信号を作成する
情報処理装置。
(8)
上記(1)乃至(7)の何れかに記載の情報処理装置であって、
前記予測部は、解像度のサンプリングマップと、SPPのサンプリングマップとを予測し、
前記レンダリング条件決定部は、前記解像度のサンプリングマップに基づき解像度のレンダリング条件を設定し、前記SPPのサンプリングマップに基づきSPPのレンダリング条件を設定する
情報処理装置。
(9)
上記(1)乃至(7)の何れかに記載の情報処理装置であって、
前記予測部は、1次元の前記サンプリングマップを予測し、
前記レンダリング条件決定部は、前記1次元のサンプリングマップに基づき解像度のレンダリング条件及びSPPのレンダリング条件を設定する
情報処理装置。
(10)
上記(1)乃至(9)の何れかに記載の情報処理装置であって、
前記プリレンダリング画像のSPPは、前記最終レンダリング画像のSPPより低い
情報処理装置。
(11)
上記(1)乃至(10)の何れかに記載の情報処理装置であって、
前記プリレンダリング画像の解像度は、前記最終レンダリング画像の解像度より低い
情報処理装置。
(12)
モデルデータをレイトレーシングによりプリレンダリングしてプリレンダリング画像を作成し、
前記プリレンダリング画像内の復元の難易度を予測し、
前記復元の難易度に基づき、前記プリレンダリング画像内の要素毎の解像度及びSPP(Sample Per Pixel)を指定するレンダリング条件を決定し、前記レンダリング条件を設定するアダプティブ制御信号を作成し、
前記アダプティブ制御信号に設定された前記要素毎の前記レンダリング条件に従って、前記モデルデータをレイトレーシングによりレンダリングすることにより、アダプティブレンダリング画像を作成し、
前記アダプティブレンダリング画像を超解像及びデノイズすることにより復元して最終レンダリング画像を作成する
情報処理方法。
(13)
情報処理装置のプロセッサを、
モデルデータをレイトレーシングによりプリレンダリングしてプリレンダリング画像を作成するプリレンダリング部と、
前記プリレンダリング画像内の復元の難易度を予測する予測部と、
前記復元の難易度に基づき、前記プリレンダリング画像内の要素毎の解像度及びSPP(Sample Per Pixel)を指定するレンダリング条件を決定し、前記レンダリング条件を設定するアダプティブ制御信号を作成するレンダリング条件決定部と、
前記アダプティブ制御信号に設定された前記要素毎の前記レンダリング条件に従って、前記モデルデータをレイトレーシングによりレンダリングすることにより、アダプティブレンダリング画像を作成するレンダリング部と、
前記アダプティブレンダリング画像を超解像及びデノイズすることにより復元して最終レンダリング画像を作成するレンダリング画像復元部と
して動作させる情報処理プログラム。
(14)
モデルデータをレイトレーシングによりプリレンダリングしてプリレンダリング画像を作成するプリレンダリング部と、
前記プリレンダリング画像内の復元の難易度を予測する予測部と、
前記復元の難易度に基づき、前記プリレンダリング画像内の要素毎の解像度及びSPP(Sample Per Pixel)を指定するレンダリング条件を決定し、前記レンダリング条件を設定するアダプティブ制御信号を作成するレンダリング条件決定部と、
前記アダプティブ制御信号に設定された前記要素毎の前記レンダリング条件に従って、前記モデルデータをレイトレーシングによりレンダリングすることにより、アダプティブレンダリング画像を作成するレンダリング部と、
前記アダプティブレンダリング画像を超解像及びデノイズすることにより復元して最終レンダリング画像を作成するレンダリング画像復元部と
を具備する情報処理システム。
(15)
情報処理装置のプロセッサを、
モデルデータをレイトレーシングによりプリレンダリングしてプリレンダリング画像を作成するプリレンダリング部と、
前記プリレンダリング画像内の復元の難易度を予測する予測部と、
前記復元の難易度に基づき、前記プリレンダリング画像内の要素毎の解像度及びSPP(Sample Per Pixel)を指定するレンダリング条件を決定し、前記レンダリング条件を設定するアダプティブ制御信号を作成するレンダリング条件決定部と、
前記アダプティブ制御信号に設定された前記要素毎の前記レンダリング条件に従って、前記モデルデータをレイトレーシングによりレンダリングすることにより、アダプティブレンダリング画像を作成するレンダリング部と、
前記アダプティブレンダリング画像を超解像及びデノイズすることにより復元して最終レンダリング画像を作成するレンダリング画像復元部と
して動作させる情報処理プログラム
を記録した非一過性のコンピュータ読み取り可能な記録媒体。
プリレンダリング部101
サンプリングマップ予測部102
レンダリング条件決定部103
レンダリング部104
レンダリング画像復元部105
条件予測係数106
画像復元係数107
Claims (14)
- モデルデータをレイトレーシングによりプリレンダリングしてプリレンダリング画像を作成するプリレンダリング部と、
前記プリレンダリング画像内の復元の難易度を予測する予測部と、
前記復元の難易度に基づき、前記プリレンダリング画像内の要素毎の解像度及びSPP(Sample Per Pixel)を指定するレンダリング条件を決定し、前記レンダリング条件を設定するアダプティブ制御信号を作成するレンダリング条件決定部と、
前記アダプティブ制御信号に設定された前記要素毎の前記レンダリング条件に従って、前記モデルデータをレイトレーシングによりレンダリングすることにより、アダプティブレンダリング画像を作成するレンダリング部と、
前記アダプティブレンダリング画像を超解像及びデノイズすることにより復元して最終レンダリング画像を作成するレンダリング画像復元部と
を具備する情報処理装置。 - 請求項1に記載の情報処理装置であって、
前記レンダリング条件は、さらに、画像枚数、バウンス数、内部透過の屈折回数、ノイズの乱数列、ビット深度、時間解像度、光成分のオン/オフ、アンチエイリアスのオン/オフ及び/又はサブサンプル数を指定する
情報処理装置。 - 請求項1に記載の情報処理装置であって、
前記レンダリング条件決定部は、画像処理、注目点、被写体の重要度及び/又は表示情報に基づき前記レンダリング条件を決定する
情報処理装置。 - 請求項1に記載の情報処理装置であって、
前記レンダリング条件決定部は、前記要素毎として、画素毎、複数の画素を含むパッチ毎又はオブジェクト領域毎の前記レンダリング条件を決定する
情報処理装置。 - 請求項1に記載の情報処理装置であって、
前記レンダリング条件決定部は、一部の前記要素毎のレンダリング条件を決定し、
他の前記要素毎のレンダリング条件は、予め決定される
情報処理装置。 - 請求項1に記載の情報処理装置であって、
前記予測部は、条件予測DNN(Deep Neural Network)に前記プリレンダリング画像を入力して前記要素毎の復元の難易度を予測し、
前記レンダリング画像復元部は、前記条件予測DNNと同時に学習された復元DNNに前記アダプティブレンダリング画像及び前記アダプティブ制御信号を入力して前記最終レンダリング画像を作成する
情報処理装置。 - 請求項1に記載の情報処理装置であって、
前記予測部は、前記プリレンダリング画像内の前記要素毎に復元の難易度を示すサンプリングマップを予測し、
前記レンダリング条件決定部は、前記サンプリングマップに基づき、前記アダプティブ制御信号を作成する
情報処理装置。 - 請求項1に記載の情報処理装置であって、
前記予測部は、解像度のサンプリングマップと、SPPのサンプリングマップとを予測し、
前記レンダリング条件決定部は、前記解像度のサンプリングマップに基づき解像度のレンダリング条件を設定し、前記SPPのサンプリングマップに基づきSPPのレンダリング条件を設定する
情報処理装置。 - 請求項1に記載の情報処理装置であって、
前記予測部は、1次元の前記サンプリングマップを予測し、
前記レンダリング条件決定部は、前記1次元のサンプリングマップに基づき解像度のレンダリング条件及びSPPのレンダリング条件を設定する
情報処理装置。 - 請求項1に記載の情報処理装置であって、
前記プリレンダリング画像のSPPは、前記最終レンダリング画像のSPPより低い
情報処理装置。 - 請求項1に記載の情報処理装置であって、
前記プリレンダリング画像の解像度は、前記最終レンダリング画像の解像度より低い
情報処理装置。 - モデルデータをレイトレーシングによりプリレンダリングしてプリレンダリング画像を作成し、
前記プリレンダリング画像内の復元の難易度を予測し、
前記復元の難易度に基づき、前記プリレンダリング画像内の要素毎の解像度及びSPP(Sample Per Pixel)を指定するレンダリング条件を決定し、前記レンダリング条件を設定するアダプティブ制御信号を作成し、
前記アダプティブ制御信号に設定された前記要素毎の前記レンダリング条件に従って、前記モデルデータをレイトレーシングによりレンダリングすることにより、アダプティブレンダリング画像を作成し、
前記アダプティブレンダリング画像を超解像及びデノイズすることにより復元して最終レンダリング画像を作成する
情報処理方法。 - 情報処理装置のプロセッサを、
モデルデータをレイトレーシングによりプリレンダリングしてプリレンダリング画像を作成するプリレンダリング部と、
前記プリレンダリング画像内の復元の難易度を予測する予測部と、
前記復元の難易度に基づき、前記プリレンダリング画像内の要素毎の解像度及びSPP(Sample Per Pixel)を指定するレンダリング条件を決定し、前記レンダリング条件を設定するアダプティブ制御信号を作成するレンダリング条件決定部と、
前記アダプティブ制御信号に設定された前記要素毎の前記レンダリング条件に従って、前記モデルデータをレイトレーシングによりレンダリングすることにより、アダプティブレンダリング画像を作成するレンダリング部と、
前記アダプティブレンダリング画像を超解像及びデノイズすることにより復元して最終レンダリング画像を作成するレンダリング画像復元部と
して動作させる情報処理プログラム。 - モデルデータをレイトレーシングによりプリレンダリングしてプリレンダリング画像を作成するプリレンダリング部と、
前記プリレンダリング画像内の復元の難易度を予測する予測部と、
前記復元の難易度に基づき、前記プリレンダリング画像内の要素毎の解像度及びSPP(Sample Per Pixel)を指定するレンダリング条件を決定し、前記レンダリング条件を設定するアダプティブ制御信号を作成するレンダリング条件決定部と、
前記アダプティブ制御信号に設定された前記要素毎の前記レンダリング条件に従って、前記モデルデータをレイトレーシングによりレンダリングすることにより、アダプティブレンダリング画像を作成するレンダリング部と、
前記アダプティブレンダリング画像を超解像及びデノイズすることにより復元して最終レンダリング画像を作成するレンダリング画像復元部と
を具備する情報処理システム。
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202280019857.0A CN116964635A (zh) | 2021-03-17 | 2022-02-10 | 信息处理装置、信息处理方法、信息处理程序和信息处理系统 |
US18/549,324 US20240169657A1 (en) | 2021-03-17 | 2022-02-10 | Information processing apparatus, information processing method, information processing program, and information processing system |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP2021043193A JP2024063786A (ja) | 2021-03-17 | 2021-03-17 | 情報処理装置、情報処理方法、情報処理プログラム及び情報処理システム |
JP2021-043193 | 2021-03-17 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2022196200A1 true WO2022196200A1 (ja) | 2022-09-22 |
Family
ID=83322210
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2022/005356 WO2022196200A1 (ja) | 2021-03-17 | 2022-02-10 | 情報処理装置、情報処理方法、情報処理プログラム及び情報処理システム |
Country Status (4)
Country | Link |
---|---|
US (1) | US20240169657A1 (ja) |
JP (1) | JP2024063786A (ja) |
CN (1) | CN116964635A (ja) |
WO (1) | WO2022196200A1 (ja) |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20200380763A1 (en) * | 2019-06-03 | 2020-12-03 | Nvidia Corporation | Bayesian machine learning system for adaptive ray-tracing |
-
2021
- 2021-03-17 JP JP2021043193A patent/JP2024063786A/ja active Pending
-
2022
- 2022-02-10 CN CN202280019857.0A patent/CN116964635A/zh active Pending
- 2022-02-10 WO PCT/JP2022/005356 patent/WO2022196200A1/ja active Application Filing
- 2022-02-10 US US18/549,324 patent/US20240169657A1/en active Pending
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20200380763A1 (en) * | 2019-06-03 | 2020-12-03 | Nvidia Corporation | Bayesian machine learning system for adaptive ray-tracing |
Non-Patent Citations (1)
Title |
---|
ALEXANDR KUZNETSOV; NIMA KHADEMI KALANTARI; RAVI RAMAMOORTHI: "Deep Adaptive Sampling for Low Sample Count Rendering", COMPUTER GRAPHICS FORUM : JOURNAL OF THE EUROPEAN ASSOCIATION FOR COMPUTER GRAPHICS, WILEY-BLACKWELL, OXFORD, vol. 37, no. 4, 20 July 2018 (2018-07-20), Oxford , pages 35 - 44, XP071489107, ISSN: 0167-7055, DOI: 10.1111/cgf.13473 * |
Also Published As
Publication number | Publication date |
---|---|
US20240169657A1 (en) | 2024-05-23 |
CN116964635A (zh) | 2023-10-27 |
JP2024063786A (ja) | 2024-05-14 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105374005B (zh) | 数据处理系统及其操作方法、计算机可读存储介质 | |
US7916934B2 (en) | Method and system for acquiring, encoding, decoding and displaying 3D light fields | |
US8860790B2 (en) | Rendering improvement for 3D display | |
US8044994B2 (en) | Method and system for decoding and displaying 3D light fields | |
JP2022523335A (ja) | 層状シーン分解コーデックシステムおよび方法 | |
EP2092488B1 (en) | Image compression and/or decompression | |
EP2797054B1 (en) | Rendering of an indirect illumination data buffer | |
US20080074438A1 (en) | Sampling methods suited for graphics hardware acceleration | |
EP1906359B1 (en) | Method, medium and system rendering 3-D graphics data having an object to which a motion blur effect is to be applied | |
US20070013696A1 (en) | Fast ambient occlusion for direct volume rendering | |
JP2012527005A (ja) | 表示装置およびそのための方法 | |
KR20050004124A (ko) | 흐림 생성 방법 | |
US10868969B1 (en) | Method and apparatus for accelerated tonemapping and display | |
CN114514746B (zh) | 用于作为对视频编码的预处理的运动自适应滤波的系统和方法 | |
WO2022196200A1 (ja) | 情報処理装置、情報処理方法、情報処理プログラム及び情報処理システム | |
EP3818495B1 (fr) | Procede et dispositif d'affichage de donnees sonar ou radar a haute dynamique | |
WO2023217867A1 (en) | Variable resolution variable frame rate video coding using neural networks | |
US20050007379A1 (en) | Matched texture filter design for rendering multi-rate data samples | |
CN113256785B (zh) | 图像处理方法、装置、设备及介质 | |
EP4242960A1 (en) | High efficiency dynamic contrast processing | |
CN116681594B (zh) | 图像处理方法及装置、设备、介质 | |
Nguyen | Adaptive Windowed Sinc Filter for Image Interpolation | |
US20240169654A1 (en) | Information processing apparatus, information processing method, and information processing program | |
US20230325977A1 (en) | Adaptive tile based super resolution | |
Ritschel et al. | Perceptual Rasterization for Head-mounted Display Image Synthesis |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 22770959 Country of ref document: EP Kind code of ref document: A1 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 18549324 Country of ref document: US |
|
WWE | Wipo information: entry into national phase |
Ref document number: 202280019857.0 Country of ref document: CN |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 22770959 Country of ref document: EP Kind code of ref document: A1 |
|
NENP | Non-entry into the national phase |
Ref country code: JP |