CN101520900A - Method and special equipment for quickening CR/DR/CT graphic display and graphic processing by utilizing GPU - Google Patents

Method and special equipment for quickening CR/DR/CT graphic display and graphic processing by utilizing GPU Download PDF

Info

Publication number
CN101520900A
CN101520900A CN200910131429A CN200910131429A CN101520900A CN 101520900 A CN101520900 A CN 101520900A CN 200910131429 A CN200910131429 A CN 200910131429A CN 200910131429 A CN200910131429 A CN 200910131429A CN 101520900 A CN101520900 A CN 101520900A
Authority
CN
China
Prior art keywords
image
gpu
texture
graphic
medical
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN200910131429A
Other languages
Chinese (zh)
Inventor
李志超
王健
何中市
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Third Military Medical University TMMU
First Affiliated Hospital of TMMU
Original Assignee
First Affiliated Hospital of TMMU
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 First Affiliated Hospital of TMMU filed Critical First Affiliated Hospital of TMMU
Priority to CN200910131429A priority Critical patent/CN101520900A/en
Publication of CN101520900A publication Critical patent/CN101520900A/en
Pending legal-status Critical Current

Links

Images

Landscapes

  • Image Processing (AREA)
  • Image Generation (AREA)

Abstract

The invention relates to a method and special equipment for quickening CR/DR/CT graphic display and graphic processing by utilizing a graphic processing unit (GPU), belonging to the field of medical graphic display and comprising a personal computer (PC), a CR/DR workstation or a picture archiving and communication system (PACS) server. The PC is connected with the CR/DR workstation or the PACS server. The invention utilizes the general computing power of the novel GPU to realize the quick display of CR/DR/CT graphic, accomplishes the graphic display and processing function by utilizing the GPU, and provides a processing method for dividing and splicing graphics. Different from the graphic device interface (GDI) used by the prior PACS, the method sufficiently utilizes the general computing power of the GPU, thereby freeing the GPU from heavy calculation of graphic processing to focus on processing of data communication and data safety and improving the efficiency of the system.

Description

A kind of GPU that utilizes quickens CR/DR/CT image demonstration and image process method and specialized equipment
Technical field
The invention belongs to medical image and show the field, is method and the specialized equipment that a kind of PACS of being used for client utilizes the acceleration CR/DR/CT image of graphic process unit GPU (graphics processing unit) to show and handle.
Background technology
Existing PACS system because image shows that adopt is early stage Graphics Device Interface GDI (GraghicDeice Interface), because GDI is restricted on function, can't be implemented in the hardware-accelerated of the doctor uses always on the PACS client image manipulation; In original CR/DR/CT image demonstration and image processing techniques, use the CPU of PC to carry out Flame Image Process, the video memory of Chuan Songing then, etc. to be shown, but because the operation that CPU also bears other processes simultaneously, and the CR/DR image data amount that the doctor uses always is bigger, therefore CPU is in and operates with full load, reduce the speed of slice, influenced doctor's work efficiency; In addition, the photo-irradiation treatment that the system's fixed pipelines in original technology is given tacit consent to, display speed is slower, the doctor is read sector-meeting produce negative interference; Display body position in original technology is provided with, and lacks dirigibility, and is unfavorable for later stage expansion self-defining operation.
Summary of the invention
The technical problem to be solved in the present invention provides a kind of GPU of utilization and quickens CR/DR/CT image demonstration and image process method and specialized equipment, utilize the processing power of GPU, quicken the demonstration of CR/DR image, utilize GPU to finish works of treatment such as the convergent-divergent, rotation of CR/DR image, local amplifications, window width and window level adjustment, reach and quicken the daily purpose of readding sheet work of processing; Simultaneously, provide corresponding interface, the handling procedure write voluntarily can be applied on the display image, reach the purpose of function performances such as the high speed experimental image is cut apart, registration.
In order to solve the problems of the technologies described above, the invention provides provides a kind of GPU of utilization to quicken demonstration of CR/DR/CT image and image process method, may further comprise the steps:
A) obtain 10 signless DICOM medical images from the image server of PACS system or the workstation of CR/DR/CT;
10 signless DICOM medical images that b) will obtain are converted to the employed R10G10B10A2_UINT form 16 bit image data texturings of GPU;
C) size of judging the image texture data whether be 2n doubly, if then do not adjust image; Otherwise, the image size is adjusted into minimum 2n times of image greater than original size;
D) judge current system DIRECTX version, if less than the DIRECTX10 version, a texture can not be stored the image of CR/DR, with the texture cutting is the little texture array of polylith, texture array according to cutting is provided with the dough sheet array, and little texture section is affixed on respectively on the corresponding dough sheet; If DIRECTX10 or DIRECTX 10 above versions, a texture can be stored the image of CR/DR, and the monolithic dough sheet is set, and image texture is affixed on the surface;
E) according to the processing power of GPU, the setting of concrete display body is to the setting that realizes the medical image window width and window level based on the pixel processor of GPU of texture image data, the negative process of medical image; Perhaps to the medical image of reality, utilize based on the summit handling procedure of GPU and realize the moving of medical image, the convergent-divergent of medical image, the rotation of medical image, the upset of medical image.
For realizing said method, need special equipment, comprising: a) the PC computer is one, and video card and the driving of supporting the DX10 standard are installed, and WindowsXP operating system or WindowsVistaProfessional version operating system possess the DX10 function; Software section includes the medical image read module, and based on the data image pretreatment module of GPU, display body is provided with module, based on the image filtering module of GPU, based on the image texture of GPU module is set, based on the image processing module of GPU; B) CR/DR workstation or PACS server; Wherein, PC computer chain is followed CR/DR workstation or PACS server.
Adopting the above-mentioned this GPU of utilization to quicken the CR/DR/CT image shows and image process method, because in the PC computer software, comprised the medical image read module, data image pretreatment module based on GPU, display body is provided with module, based on the image filtering module of GPU, based on the image texture of GPU module is set, based on the image processing module of GPU; Wherein, GPU (Graphic Processing Unit, graphic process unit) is a notion with respect to CPU, in the computing machine in modern times, particularly at image workstation, in the fan's of part recreation the household PC, the processing of figure becomes more and more important, needs a special graph core processor.GPU is display card " heart ", also just is equivalent to CPU in the operating central role of computer, and it has determined class and most of performance of this video card, also is the differentiation criterion of 2D display card and 3D display hardware simultaneously; The 2D display chip mainly relies on the processing power of CPU when handling 3D rendering and special efficacy, be called " soft acceleration "; The 3D display chip is that 3-D view and special effect processing function are concentrated in the display chip, also promptly so-called " hardware-accelerated " function, and the GPU at initial stage is the simple T﹠amp of support on the hardware; L (Transform and Lighting, polygon conversion is handled with light source) display chip, develop rapidly along with display technique, the function of GPU no longer has been a fixed pipelines, the proposition of shader notion, GPU can be handled summit and pixel according to the program of formulating, and develop into other video card of DX10 level of SM4.0 (shader model) specification, with unified shader architecture summit renderer and pixel rendering device are combined especially, the maximum number of programmed instruction has been increased to 64000 from 512; Working storage quantity also is increased to surprising 4096 from original 32, allows GPU can handle the program that more scripts need CPU to handle, and also allows the software engineer utilize GPU to handle nongraphical data in mode more easily.And the support of texture has also been reached maximum 8096*8096, and having surpassed the size of CR/DR figure, feasible image with a texture storage CR/DR becomes possibility.
For the ease of handling the special instruction of GPU, the high level shading language of some row has appearred, and these language tools can allow the user write the code of similar C, and compiling subsequently converts the shader instruction of GPU to.These instruments have been integrated the simple data multiple programming structure that can map directly to GPU; Likened to become " stream " visually (stream) through the GPU storage with the data of operation, be similar to the array in the standard C; Core (Kernel) is the function of operation on stream, means at stream and has implemented implicit circulation on the element calling a core function on a series of inlet flows, promptly each stream element is called core body.Find the solution at linear algebra, in the processing of fast fourier transform, these problems of ray tracing, utilize the processing power of GPU, under identical high-speed cache, identical cpu instruction optimal conditions, speed promotes up to 7 times more than.
Enforcement of the present invention, replaced the operation of traditional CPU to view data, by directly sending the CR view data to GPU, making the more effective deal with data of GPU energy. enforcement of the present invention has reduced the load of bus traffic CPU, has shortened the time that the map function of CR image display level presents.
Wherein, limit the specific requirement of the data texturing of concrete view data and conversion, be with 10 signless DICOM view data, be converted to the employed R10G10B10A2_UINT format texture of GPU data, because the singularity of medical image, data are stored with 10 unsigned numbers, do not omitted in order to guarantee its details, in such scheme, all data all are stored as the DXGI_FORMAT_R1OG10B10A2_UINT format texture, by above-mentioned concrete restriction, can guarantee the reasonable conversion of view data.
The tissue of image texture is a mode of taking individual texture storage image or many texture difference storage area images, and the Quad Patch quantity and the tissue thereof that determine image texture simultaneously and depended on, can guarantee that image texture is by the processing of the sampling function of system default mistake, in the operational process of program, all abandon the projective transformation in the fixed pipelines, matrixing and illumination conversion adopt self-defining vertex shader and pixelshader that image is moved, convergent-divergent, rotation, summit conversion that reverse turn operation is required and window width and window level adjustment, the required processes pixel of operation such as negative film adjustment are handled.
Self-defining pixel processor (pixel shader), window width and window level to image, the negative film operation is handled, can accelerate display speed, replace the photo-irradiation treatment that system's fixed pipelines is given tacit consent to simultaneously, prevent that the doctor is read sheet produces negative interference, replaced traditional work, help improving entire system efficient based on the CPU of responsible image data processing in the scheme of GDI demonstration with pixel shader.
The setting of concrete display body, use be that self-defining summit handling procedure (vertex shader) is handled common operations such as the moving of medical image, convergent-divergent, rotation, upset; Because adopted the shader program of high flexible, can expand in the later stage into how self-defining operation, as intensify image edge, image region segmentation, functions such as image pseudo-colours demonstration can be used as the experiment porch of extended function.
Description of drawings
Below in conjunction with drawings and Examples technical solution of the present invention is further specified:
Fig. 1 is the particular flow sheet that a kind of GPU of utilization of the present invention quickens demonstration of CR/DR/CT image and image process method;
Fig. 2 is the concrete arrangenent diagram that a kind of GPU of utilization of the present invention quickens the Special Equipment of demonstration of CR/DR/CT image and image process method needs.
Embodiment
As shown in Figure 1 and Figure 2, a kind of GPU that utilizes of the present invention quickens the Special Equipment that the CR/DR/CT image shows and image process method needs, comprise: a) the PC computer is 1 one, PC computer 1 has display screen 11, video card and the driving of supporting the DX10 standard are installed in the PC computer 1, WindowsXP operating system or WindowsVista Professional version operating system possess the DX10 function; Software section includes the medical image read module, and based on the data image pretreatment module of GPU, display body is provided with module, based on the image filtering module of GPU, based on the image texture of GPU module is set, based on the image processing module of GPU; B) CR/ DR workstation 21,22 or PACS server 31,32,33; Wherein, PC computer 1 is connecting CR/ DR workstation 21,22 or PACS server 31,32,33.
Utilize GPU to quicken the CR/DR/CT image and show and image process method, realize as follows:
A) obtain 10 signless DICOM medical images from the image server of PACS system or the workstation of CR/DR/CT;
10 signless DICOM medical images that b) will obtain are converted to the employed R10G10B10A2_UINT form 16 bit image data texturings of GPU;
C) size of judging the image texture data whether be 2n doubly, if then do not adjust image; Otherwise, the image size is adjusted into minimum 2n times of image greater than original size;
D) judge current system DIRECTX version, if less than the DIRECTX10 version, a texture can not be stored the image of CR/DR, with the texture cutting is the little texture array of polylith, texture array according to cutting is provided with the dough sheet array, and little texture section is affixed on respectively on the corresponding dough sheet; If DIRECTX 10 or DIRECTX 10 above versions, a texture can be stored the image of CR/DR, and the monolithic dough sheet is set, and image texture is affixed on the surface;
E) according to the processing power of GPU, the setting of concrete display body is to the setting that realizes the medical image window width and window level based on the pixel processor of GPU of texture image data, the negative process of medical image; Perhaps to the medical image of reality, utilize based on the summit handling procedure of GPU and realize the moving of medical image, the convergent-divergent of medical image, the rotation of medical image, the upset of medical image.
Below will be by the more detailed description of module technical scheme of the present invention:
Data preprocessing module
Read the DICOM data to internal memory, pDICOMImage[iImageIndex] for pointing to the array of pointers of memory image content, iImageIndex is an image index, value is 0 to picture number-1.When display image was CR/DR, iImageIndex got 0 value, because sequence of CR/DR has only a sub-picture usually.IImageHeight, iImageWidth are the height and the width of current display image.
IDirect3DDevice9::GetDeviceCaps (D3DCAPS9*pCaps) function that calls Directx obtains the maximum material size of current GPU equipment.Function parameter is for pointing to the pointer of D3DCAPS9 structure.MaxTextureWidth in the structure and MaxTextureHeight member variable are the full-size of the material of current device support.
Comparison condition 1:iImageWidth<MaxTextureWidth, condition 2iImageHeight<MaxTextureHeight.
Condition 1, condition 2 are true, true time: call CreateTexture2D foundation size and be the 2D texture of MaxTextureWidth*MaxTextureHeight, wherein the RGB component of each texture pixel is the value of DICOM file, and the ALPHA component is 0.Texture image is filled with 0 greater than the DICOM part.Set up 1 quad patch (RectPatch) simultaneously.
Condition 1, condition 2 is true, fictitious time: calling CreateTexture2D foundation size is the 2D texture of MaxTextureWidth*MaxTextureHeight, texture quantity is iImageheight/MaxTextureHeight+1, set up the iImageheight/MaxTextureHeight+1 quad patch simultaneously, dough sheet is vertically arranged.The sub-texture block that is numbered i is corresponding with the dough sheet that is numbered i.The I value is 0 to iImageheight/MaxTextureHeight.
Condition 1, condition 2 is false, true time: calling CreateTexture2D foundation size is the 2D texture of MaxTextureWidth*MaxTextureHeight, texture quantity is iImageWidth/MaxTextureWidth+1, set up the iImageWidth/MaxTextureWidth+1 quad patch simultaneously, dough sheet is transversely arranged.The sub-texture block that is numbered j is corresponding with the dough sheet that is numbered j.The J value is 0 to iImageWidth/MaxTextureWidth.
Condition 1, condition 2 is false, fictitious time: calling CreateTexture2D foundation size is the 2D texture of MaxTextureWidth*MaxTextureHeight, texture quantity is (iImageWidth/MaxTextureWidth+1) * (iImageheight/MaxTextureHeight+1), (the iImageheight/MaxTextureHeight+1 quad patch, dough sheet is grid arrangement to set up (iImageWidth/MaxTextureWidth+1) * simultaneously.Be numbered (j, texture i) be numbered (j, dough sheet correspondence i).The J value is 0 to iImageWidth/MaxTextureWidth.The I value is 0 to iImageheight/MaxTextureHeight.
Display environment is prepared: initial setting up WorldMatrix world transformation matrix, ViewMatirx viewport transform matrix, ProjectMatrix perspective transformation matrix, TransMatrix object transformation matrix.Set up video camera gCamera, wherein gCamera.focus is the video camera initial position, and gCamera.Minfocus and gCamera.Maxfocus are respectively minimum position and the maximum position of video camera on camera direction.Camera points is opposite with the dough sheet normal direction.Close the photo-irradiation treatment of fixed pipelines.The MipFilter=D3DTEXF_NONE of sampler_state is set; MinFilter=D3DTEXF_POINT; MagFilter=D3DTEXF_POINT;
Play up scene, read user's input, call the function corresponding function according to input from system.
The realization of image rotating: the image rotation realizes by following verex shader instruction segment
oPos=mul(Pos,TransMatrix);
oPos=mul(oPos,WorldMatrix);
oPos=mul(oPos,ViewMatrix);
oPos=mul(oPos,ProjectMatrix);
oTex=Tex;
Wherein oPos is the vertex position information of Vertex shader output, and wherein oTex is the material coordinate information of Vertexshader output, and the rotation figure does not change the material coordinate on each summit.WorldMatrix is world's transformation matrix, and ViewMatirx is a viewport transform matrix, and ProjectMatrix is a perspective transformation matrix, and TransMatrix is the object transformation matrix, among the present invention, 2 dimension image rotations are mapped as dough sheet around the rotation of Z axle, its form is { cos θ, sin θ, 0,0;-sin θ, cos θ, 0,0; 0,0,1,0; 0,0,0,1}
The realization of zoomed image: the image rotation realizes by following verex shader instruction segment
oPos=mul(Pos,TransMatrix);
oPos=mul(oPos,WorldMatrix);
oPos=mul(oPos,ViewMatrix);
oPos=mul(oPos,ProjectMatrix);
oTex=Tex;
Wherein oPos is the vertex position information of Vertex shader output, and wherein oTex is the material coordinate information of Vertexshader output, and the rotation figure does not change the material coordinate on each summit.WorldMatrix is world's transformation matrix, and ViewMatirx is a viewport transform matrix, and ProjectMatrix is a perspective transformation matrix, TransMatrix is the object transformation matrix, among the present invention, with 2 dimension images rotations be mapped as video camera towards on move, its amount of movement is Δ D, be converted into the transformation matrix of dough sheet again, because in this programme, displacement only occurs in Z-direction, the TransMatrix form is { 1,0,0,0; 0,1,0,0; 0,0,1,0; 0,0 ,-Δ D, 1}
The realization that window width and window level is adjusted:
The notion of window width and window level: the window width and window level technology is to limited this shortcoming of the resolving power of GTG at human eye, in real work different activities (Raiodensity) is remapped to 256 kinds of grey valencys in various degree, these different grey valencys can redefine pad value according to the different range of window width.
Realize adjusting the window width and window level operation with pixel shader:
Float?fErr=0.000005;
Float4r;
R=tex2D(samTexture,Tex);
If
((r.x+fErr<(WinLev-WinW/2))||(r.x-fErr<(WinLev-WinW/2)))return
{0,0,0,0};
If((r.x+fErr>(WinLev+WinW/2))||(r.x-fErr<(WinLev-WinW/2))
return{256.0,256.0,256.0,0};
R=(r.xyz-(WinLev-WinW/2))*256.0/WinW;
Return?r
Wherein ferr controls the error in the floating-point budget; SamTexture is the pinup picture of sampling, promptly is set to the CR image of texture mapping; R is the sampled result to image texture, promptly is set to a pixel value in the CR image of texture mapping.
The realization of negative process:
Negative process is the GTG value of pixel in the swap image.Among the present invention, abide by the regulation of 256 grades of gray scales, carry out the negative film operation with the method for adjusted pixel value and 256 supplements.
At first realize window width and window level adjustment operation with pixel shader:
Concrete processing procedure is the same.
The operation of negative film:
R*-1+256;
Wherein R is for operating the pixel value that handled 6 the insides.
Support to Mathematical Experiments: this programme opened program inner use play up the parametric variable value, allow the user can use the corresponding shader program of HLSL language compilation, this programme inside compiles this shader program that provides into use by D3DXCreateEffectFromFile, and by the variable of scheme inside is passed to the shader program by ID3DXBaseEffect::Set* series function with state value, thereby realize the user is verified fast the needs of GPU algorithm.

Claims (2)

1. one kind is utilized GPU to quicken demonstration of CR/DR/CT image and image process method, it is characterized in that may further comprise the steps:
A) obtain 10 signless DICOM medical images from the image server of PACS system or the workstation of CR/DR/CT;
10 signless DICOM medical images that b) will obtain are converted to the employed R10G10B10A2_UINT form 16 bit image data texturings of GPU;
C) size of judging the image texture data whether be 2n doubly, if then do not adjust image; Otherwise, the image size is adjusted into minimum 2n times of image greater than original size;
D) judge current system DIRECTX version, if less than DIRECTX 10 versions, a texture can not be stored the image of CR/DR, with the texture cutting is the little texture array of polylith, texture array according to cutting is provided with the dough sheet array, and little texture section is affixed on respectively on the corresponding dough sheet; If DIRECTX 10 or DIRECTX 10 above versions, a texture can be stored the image of CR/DR, and the monolithic dough sheet is set, and image texture is affixed on the surface;
E) according to the processing power of GPU, the setting of concrete display body is to the setting that realizes the medical image window width and window level based on the pixel processor of GPU of texture image data, the negative process of medical image; Perhaps to the medical image of reality, utilize based on the summit handling procedure of GPU and realize the moving of medical image, the convergent-divergent of medical image, the rotation of medical image, the upset of medical image.
2. one kind is utilized GPU to quicken that the CR/DR/CT image shows and the Special Equipment of image process method, it is characterized in that, comprise: a) the PC computer is one, video card and the driving of supporting the DX10 standard are installed, WindowsXP operating system or WindowsVista Professional version operating system possess the DX10 function; Software section includes the medical image read module, and based on the data image pretreatment module of GPU, display body is provided with module, based on the image filtering module of GPU, based on the image texture of GPU module is set, based on the image processing module of GPU; B) CR/DR workstation or PACS server; Wherein, PC computer chain is followed CR/DR workstation or PACS server.
CN200910131429A 2009-03-30 2009-03-30 Method and special equipment for quickening CR/DR/CT graphic display and graphic processing by utilizing GPU Pending CN101520900A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN200910131429A CN101520900A (en) 2009-03-30 2009-03-30 Method and special equipment for quickening CR/DR/CT graphic display and graphic processing by utilizing GPU

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN200910131429A CN101520900A (en) 2009-03-30 2009-03-30 Method and special equipment for quickening CR/DR/CT graphic display and graphic processing by utilizing GPU

Publications (1)

Publication Number Publication Date
CN101520900A true CN101520900A (en) 2009-09-02

Family

ID=41081473

Family Applications (1)

Application Number Title Priority Date Filing Date
CN200910131429A Pending CN101520900A (en) 2009-03-30 2009-03-30 Method and special equipment for quickening CR/DR/CT graphic display and graphic processing by utilizing GPU

Country Status (1)

Country Link
CN (1) CN101520900A (en)

Cited By (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102117490A (en) * 2009-12-30 2011-07-06 上海幻维数码创意科技有限公司 Method and device for constructing role modularized skeleton system based on description data
CN102209191A (en) * 2010-03-30 2011-10-05 联发科技股份有限公司 Method for performing image signal processing with aid of a graphics processing unit and associated apparatus
CN103099635A (en) * 2012-12-27 2013-05-15 李华 Graphics processing unit (GPU) digital image system for microscopic captive test (CT) machine and control method of GPU digital image system for microscopic CT machine
CN103246795A (en) * 2012-02-03 2013-08-14 南京普爱射线影像设备有限公司 Software system for displaying and processing dynamic medical image
CN103942017A (en) * 2013-01-21 2014-07-23 深圳市快播科技有限公司 Image quality regulating method and system
CN105469352A (en) * 2014-08-23 2016-04-06 北京纳米维景科技有限公司 Portable image processing system and method based on mobile GPU
CN106886974A (en) * 2015-10-21 2017-06-23 联发科技股份有限公司 Image accelerator equipment and correlation technique
CN108597590A (en) * 2018-05-08 2018-09-28 上海嘉奥信息科技发展有限公司 The arbitrary layer data display methods of volume data based on GPU operations and system
WO2018205628A1 (en) * 2017-05-12 2018-11-15 京东方科技集团股份有限公司 Display processing apparatus and display processing method thereof, and display apparatus
CN111489281A (en) * 2020-04-09 2020-08-04 常州奥创医疗科技有限公司 Detection method based on GPU and CPU cooperative operation
CN112051285A (en) * 2020-08-18 2020-12-08 大连理工大学 Intelligent nondestructive detection system integrating X-ray real-time imaging and CT (computed tomography) tomography
WO2021013019A1 (en) * 2019-07-19 2021-01-28 华为技术有限公司 Picture processing method and apparatus

Cited By (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102117490A (en) * 2009-12-30 2011-07-06 上海幻维数码创意科技有限公司 Method and device for constructing role modularized skeleton system based on description data
CN102117490B (en) * 2009-12-30 2013-11-13 上海幻维数码创意科技有限公司 Method and device for constructing role modularized skeleton system based on description data
CN102209191A (en) * 2010-03-30 2011-10-05 联发科技股份有限公司 Method for performing image signal processing with aid of a graphics processing unit and associated apparatus
CN103246795A (en) * 2012-02-03 2013-08-14 南京普爱射线影像设备有限公司 Software system for displaying and processing dynamic medical image
CN103099635A (en) * 2012-12-27 2013-05-15 李华 Graphics processing unit (GPU) digital image system for microscopic captive test (CT) machine and control method of GPU digital image system for microscopic CT machine
CN103942017A (en) * 2013-01-21 2014-07-23 深圳市快播科技有限公司 Image quality regulating method and system
CN105469352A (en) * 2014-08-23 2016-04-06 北京纳米维景科技有限公司 Portable image processing system and method based on mobile GPU
CN106886974B (en) * 2015-10-21 2020-10-02 联发科技股份有限公司 Image accelerator apparatus and related methods
CN106886974A (en) * 2015-10-21 2017-06-23 联发科技股份有限公司 Image accelerator equipment and correlation technique
WO2018205628A1 (en) * 2017-05-12 2018-11-15 京东方科技集团股份有限公司 Display processing apparatus and display processing method thereof, and display apparatus
US10681330B2 (en) 2017-05-12 2020-06-09 Boe Technology Group Co., Ltd. Display processing device and display processing method thereof and display apparatus
CN108597590A (en) * 2018-05-08 2018-09-28 上海嘉奥信息科技发展有限公司 The arbitrary layer data display methods of volume data based on GPU operations and system
WO2021013019A1 (en) * 2019-07-19 2021-01-28 华为技术有限公司 Picture processing method and apparatus
CN111489281A (en) * 2020-04-09 2020-08-04 常州奥创医疗科技有限公司 Detection method based on GPU and CPU cooperative operation
CN112051285A (en) * 2020-08-18 2020-12-08 大连理工大学 Intelligent nondestructive detection system integrating X-ray real-time imaging and CT (computed tomography) tomography
CN112051285B (en) * 2020-08-18 2021-08-20 大连理工大学 Intelligent nondestructive detection system integrating X-ray real-time imaging and CT (computed tomography) tomography

Similar Documents

Publication Publication Date Title
CN101520900A (en) Method and special equipment for quickening CR/DR/CT graphic display and graphic processing by utilizing GPU
AU2003203499B2 (en) Systems and methods for providing intermediate targets in a graphics system
US11113788B2 (en) Multi-space rendering with configurable transformation parameters
US9177351B2 (en) Multi-primitive graphics rendering pipeline
US10456666B2 (en) Block based camera updates and asynchronous displays
US8194083B2 (en) Color computation of pixels using a plurality of vertex or fragment shader programs
US11103777B2 (en) Mechanisms for reducing latency and ghosting displays
CN100504923C (en) Grain engine for grain video signal processing, graph processor and method
CN105321143A (en) Control of a sample mask from a fragment shader program
US11010861B2 (en) Fragmented graphic cores for deep learning using LED displays
CN1947156A (en) Graphics processing architecture employing a unified shader
López et al. Accelerating image recognition on mobile devices using GPGPU
EP3200157B1 (en) Frustum rendering
US10733785B2 (en) Graphics processing
US9811875B2 (en) Texture state cache
DE102010025310A1 (en) texture sampling
CA2558657A1 (en) Embedded system with 3d graphics core and local pixel buffer
US9530237B2 (en) Interpolation circuitry and techniques for graphics processing
US6975317B2 (en) Method for reduction of possible renderable graphics primitive shapes for rasterization
CN109448123B (en) Model control method and device, storage medium and electronic equipment
CN106886974A (en) Image accelerator equipment and correlation technique
US10943336B2 (en) Tone-mapping high dynamic range images
TWI401618B (en) Method and computing device for performing texture operation
Harrison et al. Optimising data movement rates for parallel processing applications on graphics processors.
Regan et al. An Architecture for Orientation Mapping Post Rendering

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C02 Deemed withdrawal of patent application after publication (patent law 2001)
WD01 Invention patent application deemed withdrawn after publication

Application publication date: 20090902