WO2015163720A1 - Procédé de création d'image tridimensionnelle, appareil de création d'image tridimensionnelle mettant en œuvre celui-ci et support d'enregistrement stockant celui-ci - Google Patents

Procédé de création d'image tridimensionnelle, appareil de création d'image tridimensionnelle mettant en œuvre celui-ci et support d'enregistrement stockant celui-ci Download PDF

Info

Publication number
WO2015163720A1
WO2015163720A1 PCT/KR2015/004096 KR2015004096W WO2015163720A1 WO 2015163720 A1 WO2015163720 A1 WO 2015163720A1 KR 2015004096 W KR2015004096 W KR 2015004096W WO 2015163720 A1 WO2015163720 A1 WO 2015163720A1
Authority
WO
WIPO (PCT)
Prior art keywords
node
tree
subspaces
data
sub
Prior art date
Application number
PCT/KR2015/004096
Other languages
English (en)
Korean (ko)
Inventor
박우찬
권혁주
허진석
김우현
Original Assignee
세종대학교 산학협력단
주식회사 실리콘아츠
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 세종대학교 산학협력단, 주식회사 실리콘아츠 filed Critical 세종대학교 산학협력단
Publication of WO2015163720A1 publication Critical patent/WO2015163720A1/fr

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

Definitions

  • the present invention relates to a three-dimensional image generation technology, and more particularly, to a three-dimensional image generating method, a three-dimensional image generating apparatus and a recording medium for storing the same, which can efficiently perform the necessary spatial segmentation in the three-dimensional image will be.
  • KD-tree is a multi-dimensional search, which is a k-dimensional tree that is used to generate 3D images and has a spatial-segmented data structure for organizing points (i.e. data) in k-dimensional space.
  • KD-tree can be useful when searching using multi-dimensional search keys (eg, range search and proximity search), and KD-tree is a kind of BSP (binary space partitioning) tree.
  • Korean Patent Laid-Open No. 10-2011-0124834 relates to a method of generating a KD-tree using hardware, and finds a location where the cost of a node is minimized by using a surface area heuristic (SAH) method and based on the KD-tree. It discloses a method of generating.
  • SAH surface area heuristic
  • the left and right sides of the list data are classified according to the SAH method based on the triangle data read from the memory by referring to the index of the list data, and the cost of the current node is classified. If no minimum location is found, a leaf node can be created.
  • Korean Patent Laid-Open No. 10-2012-0090543 discloses a method for constructing a KD-tree based on the visibility of a voxel.
  • the KD-tree is constructed by constructing a KD-tree by dividing the voxel into two sub-voxels using a plane whose cost function based on the visibility of the voxel is used to construct the KD-tree and minimizing the cost function.
  • the upper node is generated in a binning manner and the lower node is generated in parallel in a sorting manner to generate a three-dimensional image that can generate a tree at high speed while maintaining the quality of the tree.
  • An embodiment of the present invention distinguishes the upper node and the lower node based on the number of raw data, and if the number of the raw data is the number of sorting methods, the sorting method is used. Otherwise, the binning method is used.
  • An embodiment of the present invention is to provide a three-dimensional image generation method that enables burst access by aligning nodes in the search order of the nodes and storing the nodes in a memory, thereby maximizing the memory access effect.
  • the 3D image generating method includes (a) determining a partitioning method of a given space and dividing the given space into a plurality of subspaces (each forming a node of a KD-tree); and (b ) Setting each of the plurality of subspaces to the given space and repeating the step (a), wherein the step (a) is performed if the corresponding subspace corresponds to an internal node of the KD-tree. Checking whether the number of raw data in the node exceeds a threshold.
  • the threshold value may be determined based on the size of the internal memory required when dividing the corresponding subspace in an aligned manner.
  • the inner node may be composed of a partition plane of the given space and a plurality of sub-spaces divided by the partition plane.
  • the step (a) may further include generating the plurality of subspaces by dividing the corresponding subspaces in a binning manner if the number of the raw data exceeds the threshold. .
  • the step (a) may further include generating a plurality of subspaces by dividing the corresponding subspaces in a sorting manner if the number of the raw data does not exceed the threshold. .
  • the step (a) may include generating the plurality of subspaces by performing partitioning of the alignment scheme for the corresponding subspaces in parallel.
  • the step (a) may include converting a division order of the plurality of sub spaces into a visit order of the KD-tree and storing the plurality of sub spaces in a memory.
  • the storing of the plurality of subspaces in a memory may include receiving an address of a node and data of a node generated through partitioning of the given space, and a burst node entry in which the data of the node is to be stored. searching for the location of the burst node entry may comprise comparing the node address with the burst address.
  • the storing of the plurality of subspaces in a memory may include searching for a burst node entry having no space for storing node data.
  • the 3D image generating apparatus may (a) determine a partitioning scheme of a given space and divide the given space into a plurality of subspaces (each forming a node of a KD-tree), and (b) A tree generating unit for repeating the step (a) by setting each of the plurality of sub-spaces to the given space, wherein the tree generating unit includes the internal if the sub-space corresponds to an internal node of the KD-tree; Check whether the number of raw data in the node exceeds the threshold.
  • the threshold value may be determined based on the size of the internal memory required when dividing the corresponding subspace in an aligned manner.
  • the inner node may be composed of a partition plane of the given space and a plurality of sub-spaces divided by the partition plane.
  • the tree generating unit may include a first tree generating unit generating the plurality of sub spaces by dividing the corresponding sub spaces in a binning manner if the number of the raw data exceeds the threshold. have.
  • the tree generating unit may include at least one second tree generating unit that generates a plurality of subspaces by dividing the corresponding subspaces in a sorting manner if the number of the raw data does not exceed the threshold. It may include.
  • the second tree generator may generate the plurality of subspaces by performing partitioning of the sorting method for the corresponding subspaces in parallel.
  • the tree generating unit may include a node scheduler that converts a split order of the plurality of sub spaces into a visit order of the KD-tree and stores the plurality of sub spaces in a memory.
  • the node scheduler receives a node address and data of a node generated through the division of the given space, and a push entry for searching a location of a burst node entry in which the data of the node is to be stored.
  • the push entry selector may include a node address and a burst address.
  • the node scheduler may include a pop entry selector for searching for burst node entries for which there is no space for storing node data.
  • the recording medium on which the computer program relating to the 3D image generating method is recorded may be configured to (a) determine a method of partitioning a given space to form the given space in a plurality of sub-spaces (each of which forms a node of a KD-tree). And (b) setting each of the plurality of subspaces to the given space and repeating step (a), wherein step (a) includes the KD-tree If it corresponds to the internal node of the computer program for a three-dimensional image generation method comprising the function of checking whether the number of raw data in the internal node exceeds a threshold.
  • the upper node is generated in a binning manner and the lower node is generated in parallel in a sorting manner to generate the tree at high speed while maintaining the quality of the tree.
  • a three-dimensional image generating method distinguishes upper nodes and lower nodes based on the number of raw data, and uses a sorting method when the number of raw data is the number of sorting methods. Otherwise, the tree can be generated at high speed while maintaining the quality of the tree by creating the tree using binning.
  • the method for generating a 3D image stores all data necessary for generating a tree in the internal memory, thereby reducing the case of accessing the external memory when the tree is generated.
  • burst access is possible by arranging nodes and storing them in a memory in order of node search, thereby maximizing memory access effects.
  • FIG. 1 is a block diagram illustrating a tree generation unit according to an embodiment of the present invention.
  • Figure 2 (a) is one embodiment of a split surface generated in a binning manner with respect to one axis.
  • FIG. 2 (b) is one embodiment of a split plane created in alignment with respect to one axis.
  • FIG. 3 is a block diagram illustrating a ray tracing apparatus including the tree generation unit of FIG. 1.
  • FIG. 4 is a block diagram illustrating a working memory according to an embodiment of the present invention.
  • 5A is a block diagram illustrating a node scheduler and a node memory according to an embodiment of the present invention.
  • 5B is a block diagram illustrating an embodiment of a node scheduler.
  • FIG. 6 is a block diagram illustrating a process of ray tracing.
  • FIG. 7 is a block illustrating the acceleration structure and geometric data used in the disclosed technique.
  • FIG. 8 is a flowchart illustrating a tree generation unit according to an embodiment of the present invention.
  • first and second are intended to distinguish one component from another component, and the scope of rights should not be limited by these terms.
  • first component may be named a second component, and similarly, the second component may also be named a first component.
  • an identification code (e.g., a, b, c, etc.) is used for convenience of description, and the identification code does not describe the order of the steps, and each step clearly indicates a specific order in context. Unless stated otherwise, they may occur out of the order noted. That is, each step may occur in the same order as specified, may be performed substantially simultaneously, or may be performed in the reverse order.
  • the present invention can be embodied as computer readable code on a computer readable recording medium
  • the computer readable recording medium includes all kinds of recording devices in which data can be read by a computer system.
  • Examples of computer-readable recording media include ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical data storage device, and the like, and are also implemented in the form of a carrier wave (for example, transmission over the Internet). It also includes.
  • the computer readable recording medium can also be distributed over network coupled computer systems so that the computer readable code is stored and executed in a distributed fashion.
  • FIG. 1 is a block diagram illustrating a tree generation unit according to an embodiment of the present invention.
  • the tree build unit 100 may include a first tree generator 110 and at least one second tree generator 120.
  • the first tree generator 110 generates a tree using a binning method.
  • the binning method divides a bounding box space by a predetermined number for three axes (x, y, z), and calculates a surface area heuristic (SAH) value only for the divided surface.
  • SAH is a method of creating an acceleration structure for ray-tracing, where values for finding primitives that collide with any ray (e.g., node visits, intersection with the raw data) And the number of times to calculate whether or not to calculate whether or not) is partitioned based on the partition having the best value among the calculation results.
  • the at least one second tree generator 120 generates a tree using a sorting method.
  • the sorting method extracts the start value and the end value of the entire raw data for the three axes (x, y, z) and divides the three axes into the start value and the end value to generate six list data.
  • the list data is an array consisting of indices of the raw data.
  • the list data is sorted and a split plane is generated from the start value and the end value of the sorted raw data.
  • the space is partitioned based on the partition having the best value by calculating the SAH value in the generated partition.
  • the at least one second tree generator 120 may obtain raw data information through the first tree generator 110 and the plurality of second tree generators 120 may operate in parallel. Is connected to an arbiter 140 and stores node and list data (AS data) received from at least one second tree generator 120 in an external memory 340.
  • AS data node and list data
  • a working memory 130 When the tree is generated in the first tree generator 110 and the at least one second tree generator 120, a working memory 130 is used.
  • the working memory 130 will be described with reference to FIG. 4.
  • Figure 2 (a) is one embodiment of a split surface generated in a binning manner with respect to one axis.
  • the bounding box 210a may include a dividing surface 220a and raw data 230a.
  • the dividing plane 220a is predetermined with five dividing planes, and the SAH value at each dividing plane 220a may be calculated for the six raw data 230a.
  • the bounding box 210a may be divided by selecting the splitting surface having the best value.
  • FIG. 2 (b) is one embodiment of a split plane created in alignment with respect to one axis.
  • the bounding box 210b may include a dividing surface 220b and raw data 230b.
  • the partition plane 220b is generated from the sorted start and end values of the raw data 230b, and the SAH values at each partition plane 220b can be calculated for the raw data 230b.
  • the bounding box 210b may be divided by selecting the splitting surface having the best value.
  • FIG. 3 is a block diagram illustrating a ray tracing apparatus including the tree generation unit of FIG. 1.
  • the ray tracing apparatus 300 may include a central processing unit (CPU) 310, a system memory 320, a dynamic ray tracing accelerator 330, and an external memory. 340 may be included.
  • the CPU 310 may process a 3D application, and may include at least one of an application 311 such as a 3D game engine, an application programming interface (API) 312, and a scene manager 313. Can be.
  • an application 311 such as a 3D game engine
  • API application programming interface
  • scene manager 313. Can be.
  • the system memory 320 may store graphic data information required for a 3D application, and may include a PSS area 321 for storing a primitive static scene and a PDS area for storing a primitive dynamic scene ( 322 and a texture map area 323 that stores a mip map (MIP-MAP) for texture mapping.
  • MIP-MAP mip map
  • the DRTX 330 includes a tree build unit 100 of FIG. 1, and includes a bus interface unit 331, a primitive 332, and accelerated structure data (AS data). (333), working memory (130), accelerated structure cache (AS Cache) 335, texture cache (336), color result buffer (337), stack memory ( The stack memory 338 and a ray tracing unit 339 may be further included.
  • a spatial segmentation structure is constructed based on the graphic data information, ray tracing is performed based on the generated spatial segmentation structure, the results of the ray tracing performed are sent to the CPU 310, and the ray tracing speed Can be used to reconstruct the spatial partitioning structure for the graphical data information.
  • the external memory 340 may temporarily store information processed by the DRTX 330, and may include a geometric information storage area 341, a static scene acceleration structure storage area 342, a dynamic scene acceleration structure storage area 343, and a texture. It may include a map storage area 344 and a color information storage area 345.
  • FIG. 4 is a block diagram illustrating a working memory according to an embodiment of the present invention.
  • the working memory 130 may include primitive data 410, primitive index 420, temporary space 430, and accelerated structure data ( Acceleration Structure data) 440 may include memories.
  • the x, y, z coordinate information of the raw data received from the at least one second tree generator 120 is stored in the raw data 410 memory, and the raw data index 420 is aligned in the process of sorting the raw data in the memory.
  • the index numbers of the raw data are stored in the same order.
  • the index number of the raw data divided left and right while being classified based on the partition plane is temporarily stored in the temporary space 430 memory, and at least one second tree generation unit is performed in the accelerated structure data 440 memory.
  • the generated node and list data are stored.
  • the working memory 130 corresponds to an internal memory.
  • 5A is a block diagram illustrating a node scheduler and a node memory according to an embodiment of the present invention.
  • a node scheduler 510 includes a node data manager 511, a push entry selector 512, and a pop entry selector.
  • Selector 513 and the node memory 520 may include a plurality of data array fields 521.
  • the burst node entry 530 may include a presence field, a full field and a burst field, and a data arrangement field 521 of the node data manager 511.
  • the node data manager 511 manages the information of the data array fields 521 present in the node memory and includes a plurality of existing fields, full fields, and burst fields. It is composed.
  • the presence field has information on whether there is node data in the data array field 521
  • the full field has information on whether it is not possible to store more data in the data array field 521
  • the burst field has The node data has an external memory address for burst access.
  • the push entry selector 512 selects a burst node entry 530 in which node data generated by the first and second tree generators 110 and 120 are to be stored, and selects the node scheduler 510. It receives information about node data and node address from the outside.
  • the node address means an external memory address where node data is to be stored in the node search order.
  • the push entry selector 512 receives information of each burst node entry 530 from a node data manager 511 in the node scheduler 510.
  • the push entry selection unit 512 determines the position of the burst node entry 530 in which node data is to be stored based on the input data, and when the position is selected, the push entry selection unit 512 selects the data of the node address. After checking the data offset, the node data is stored in the data array field 521 of the corresponding burst node entry 530.
  • the node address includes a burst address and a data address
  • the data offset has location information in which the node data is to be stored in the data array field 521.
  • the pop entry selector 513 outputs the data array field 521 of the selected burst node entry 530 to the outside. More specifically, the pop entry selector 513 receives information in the full field of all burst node entries 530 from the node data manager 511 and checks whether valid information exists. If valid information exists, the burst field and data array field 521 of the corresponding burst node entry 530 are output to the outside.
  • the data array field 521 is a memory space capable of storing a plurality of node data and has the same size as the burst access size of the external memory 340.
  • the presence field, the full field, and the burst field of the node data manager 511 together with the data arrangement field 521 of the node memory 520 constitute one burst node entry 530.
  • the node scheduler 510 receives a node address and node data as inputs, and finds a location of the burst node entry 530 in which the node data input by the push entry selector 512 is to be stored.
  • the method of finding the position is as follows. After checking all present fields, look for burst node entries 530 for which node data exists. The burst field of the burst node entries 530 and the burst address of the input node address are then compared to find a burst field having the same address. If the burst field is found, the node data is stored in the corresponding burst node entry 530 in the push data.
  • the burst node entry 530 where no node data exists and change the burst field.
  • the selection method takes precedence over the highest burst node entry 530 for which the presence field is invalid.
  • the changing method is to store the burst address of the node address entered in the burst field of the selected burst node entry 530.
  • the node data is then stored in the data array field 521 of the burst node entry 530 initialized from the push data.
  • the pop entry selecting unit 513 checks all the full field information to find the burst node entry 530 which has no space for storing the node data. If the corresponding burst node entry 530 does not exist, the node scheduler 510 is terminated. Otherwise, the data existing in the burst field and data array field 521 of the corresponding burst node entry 530 is output and the node scheduler 510 is terminated.
  • 5B is a block diagram illustrating an embodiment of a node scheduler.
  • the node memory 520 is stored in the node memory 520 through the process described with reference to FIG. It can be stored in the order of 7, 3, 1, 0, ... according to the search order.
  • the burst access size is 4, and accordingly, four node data may be stored in one data array field 521. That is, the generation order of nodes is 0, 1, 2, 3, ..., but the order stored in the node memory is 7, 3, 1, 0, ... can be seen that is stored in the search order.
  • FIG. 6 is a block diagram illustrating a process of ray tracing.
  • a primary ray P is generated from a camera 610 position for each pixel, and a calculation is performed to find an object 620 that meets the ray P.
  • the object encountered with the ray P is the object 620 having the property of refraction or the objects 631 and 632 having the property of reflection
  • the refraction effect is exerted at the position where the object is met with the ray P.
  • Refraction Ray (F) and Reflection Ray (R) for the reflection effect can be generated, and a Shadow Ray (S) in the direction of the light 650 can be generated.
  • the shadow ray S meets another object 640, the shadow may be generated at the point where the shadow ray S is generated.
  • FIG. 7 is a block illustrating the acceleration structure and geometric data used in the disclosed technique.
  • KD-tree is a type of spatial partitioning structure, and may be used for a ray-triangle intersection test.
  • the KD-tree may include a box node 710, an inner node 720, and a leaf node 730.
  • leaf node 730 may include a triangle list for pointing at least one triangle information included in the geometric data.
  • the triangle information may include vertex coordinates, normal vectors, and texture coordinates for three points of the triangle.
  • the triangle list included in the leaf node may correspond to the array index.
  • the inner node 720 has a bounding box-based spatial region, which may be divided into two regions and allocated to two sub-nodes. That is, the internal node 720 is composed of a partition plane and subtrees of two regions divided through the partition plane, and the leaf node 730 includes only a series of raw data.
  • the dividing plane that partitions the space must find a point where the value for finding the raw data that would encounter an arbitrary ray (for example, the number of node visits, the number of times to calculate whether it intersects the raw data, etc.) is the minimum, and SAH Surface area heuristic may correspond to a method used to find a corresponding point.
  • FIG. 8 is a flowchart illustrating a tree generation unit according to an embodiment of the present invention.
  • step S801 it is checked whether the number of raw data is larger than the threshold (step S801).
  • the threshold is the maximum number of raw data that can be processed by one second tree generator 120.
  • the space is divided by the binning method (step S802), and the process returns to step S801 to confirm again whether the number of the raw data is larger than the threshold value.
  • step S803 If the number of raw data is less than the threshold, the space is divided in a sorting manner (step S803). The division is performed and it is checked whether there is no more space to divide (whether it is a leaf node) (step S804), and if not, it is terminated.
  • dividing the space is to find the dividing plane, which is the location of dividing the space, and the dividing plane is a value for finding the raw data that collides with the ray (e.g., node visits, whether it intersects the raw data or not).
  • the number of calculations, etc. may be the minimum point.
  • the surface is divided using the surface area heuristic (SAH) calculation method, and the SAH method includes binning and alignment methods.
  • SAH surface area heuristic

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Graphics (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Image Generation (AREA)

Abstract

L'invention concerne un procédé de création d'image tridimensionnelle qui comprend les étapes consistant : (a) à déterminer une façon de diviser un espace donné et à diviser l'espace donné en une pluralité de sous-espaces, dont chacun forme un nœud d'un arbre KD ; (b) à définir chacun de la pluralité de sous-espaces comme espace donné et à répéter l'étape (a), si un sous-espace pertinent tombe sous un nœud interne de l'arbre KD, l'étape (a) comprenant une étape consistant à vérifier si le nombre de données brutes dans le nœud interne dépasse ou non un seuil.
PCT/KR2015/004096 2014-04-24 2015-04-24 Procédé de création d'image tridimensionnelle, appareil de création d'image tridimensionnelle mettant en œuvre celui-ci et support d'enregistrement stockant celui-ci WO2015163720A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
KR1020140049254A KR101560283B1 (ko) 2014-04-24 2014-04-24 3차원 영상 생성 방법, 이를 수행하는 3차원 영상 생성 장치 및 이를 저장하는 기록매체
KR10-2014-0049254 2014-04-24

Publications (1)

Publication Number Publication Date
WO2015163720A1 true WO2015163720A1 (fr) 2015-10-29

Family

ID=54332804

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/KR2015/004096 WO2015163720A1 (fr) 2014-04-24 2015-04-24 Procédé de création d'image tridimensionnelle, appareil de création d'image tridimensionnelle mettant en œuvre celui-ci et support d'enregistrement stockant celui-ci

Country Status (2)

Country Link
KR (1) KR101560283B1 (fr)
WO (1) WO2015163720A1 (fr)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2020143233A1 (fr) * 2019-01-07 2020-07-16 平安科技（深圳）有限公司 Procédé et dispositif de construction d'un modèle de carte de score, appareil informatique et support d'informations

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR101807172B1 (ko) 2016-12-28 2017-12-08 세종대학교 산학협력단 레이 트레이싱 장치 및 방법
CN108171785B (zh) * 2018-01-22 2021-12-10 中南大学 用于光线跟踪的sah-kd树设计方法
KR102089269B1 (ko) 2019-04-11 2020-03-17 주식회사 실리콘아츠 포터블 레이 트레이싱 시스템에서의 버퍼링 방법
KR102169799B1 (ko) 2019-04-11 2020-10-26 주식회사 실리콘아츠 포터블 레이 트레이싱 장치

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20100128493A (ko) * 2009-05-28 2010-12-08 박우찬 레이 트레이싱 코어 및 이를 포함하는 레이 트레이싱 칩
KR20110100795A (ko) * 2010-03-05 2011-09-15 인하대학교 산학협력단 이진트리 기반의 공간 분할 기법과 가상 피라미드를 이용한 대용량 항공 레이저 측정자료 처리 방법
KR20120019720A (ko) * 2010-08-26 2012-03-07 삼성전자주식회사 영상 처리 장치 및 방법
KR20120106930A (ko) * 2012-07-24 2012-09-27 세종대학교산학협력단 레이 트레이싱 코어 및 레이 트레이싱 처리 방법

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR20100128493A (ko) * 2009-05-28 2010-12-08 박우찬 레이 트레이싱 코어 및 이를 포함하는 레이 트레이싱 칩
KR20110100795A (ko) * 2010-03-05 2011-09-15 인하대학교 산학협력단 이진트리 기반의 공간 분할 기법과 가상 피라미드를 이용한 대용량 항공 레이저 측정자료 처리 방법
KR20120019720A (ko) * 2010-08-26 2012-03-07 삼성전자주식회사 영상 처리 장치 및 방법
KR20120106930A (ko) * 2012-07-24 2012-09-27 세종대학교산학협력단 레이 트레이싱 코어 및 레이 트레이싱 처리 방법

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2020143233A1 (fr) * 2019-01-07 2020-07-16 平安科技（深圳）有限公司 Procédé et dispositif de construction d'un modèle de carte de score, appareil informatique et support d'informations

Also Published As

Publication number Publication date
KR101560283B1 (ko) 2015-10-15

Similar Documents

Publication Publication Date Title
WO2015163720A1 (fr) Procédé de création d'image tridimensionnelle, appareil de création d'image tridimensionnelle mettant en œuvre celui-ci et support d'enregistrement stockant celui-ci
WO2010137822A2 (fr) Noyau de lancer de rayon et puce de lancer de rayon comprenant ledit noyau
US10192350B2 (en) Compacting results vectors between stages of graphics processing
US8773449B2 (en) Rendering of stereoscopic images with multithreaded rendering software pipeline
US20070132754A1 (en) Method and apparatus for binary image classification and segmentation
US11604654B2 (en) Effective and scalable building and probing of hash tables using multiple GPUs
US20100239185A1 (en) Accelerated Data Structure Optimization Based Upon View Orientation
WO2009117691A2 (fr) Architectures permettant le test et l’ombrage d'intersections en parallèle pour le rendu du tracé de rayons
Meister et al. Parallel locally-ordered clustering for bounding volume hierarchy construction
US8564600B2 (en) Streaming physics collision detection in multithreaded rendering software pipeline
WO2012111864A1 (fr) Noyau de tracé de rayon et procédé de traitement de tracé de rayon
CN111667542B (zh) 适用于人工神经网络的用于处理压缩数据的解压缩技术
WO2012011644A1 (fr) Unité centrale de traitement, son procédé de simulation de gpu et système informatique l'incluant
WO2023106710A1 (fr) Procédé et dispositif de génération de données de visualisation tridimensionnelles en séries chronologiques
WO2016186236A1 (fr) Système et procédé de traitement de couleur pour objet tridimensionnel
US8253738B2 (en) Data processing apparatus and method
WO2020138716A1 (fr) Appareil et procédé de suivi de son
CN116245997A (zh) 基于超级计算机的三维模型动态渲染并行加速方法、系统
WO2022131531A1 (fr) Procédé et dispositif de lancer de rayon à base de concentration pour scène dynamique
WO2016027910A1 (fr) Procédé de traçage de rayons, dispositif de traçage de rayons pour mettre en œuvre ledit procédé, et support d'enregistrement permettant de stocker celui-ci
WO2016093470A1 (fr) Appareil et procédé de traitement d'images et support d'enregistrement
WO2011071273A2 (fr) Générateur d'adresse basée sur sram pour chaque couche et générateur d'adresse le comprenant
WO2022080580A1 (fr) Appareil et procédé de traçage de rayon aux performances améliorées
WO2022092414A1 (fr) Dispositif de traçage de rayon basé sur une pluralité de puces et procédé utilisant une division de trames
WO2012030027A1 (fr) Dispositif de mise en correspondance de chaînes de caractères basé sur un processeur multicœur et procédé de mise en correspondance de chaînes de caractères associé

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: 15783641

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 15783641

Country of ref document: EP

Kind code of ref document: A1