CN100452090C - Method and system for implementing high dynamic light range - Google Patents

Method and system for implementing high dynamic light range Download PDF

Info

Publication number
CN100452090C
CN100452090C CNB2006100657606A CN200610065760A CN100452090C CN 100452090 C CN100452090 C CN 100452090C CN B2006100657606 A CNB2006100657606 A CN B2006100657606A CN 200610065760 A CN200610065760 A CN 200610065760A CN 100452090 C CN100452090 C CN 100452090C
Authority
CN
China
Prior art keywords
passage
integer textures
textures
integer
order
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CNB2006100657606A
Other languages
Chinese (zh)
Other versions
CN101038660A (en
Inventor
刘皓
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.)
Tencent Technology Shenzhen Co Ltd
Original Assignee
Tencent Technology Shenzhen Co Ltd
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 Tencent Technology Shenzhen Co Ltd filed Critical Tencent Technology Shenzhen Co Ltd
Priority to CNB2006100657606A priority Critical patent/CN100452090C/en
Publication of CN101038660A publication Critical patent/CN101038660A/en
Application granted granted Critical
Publication of CN100452090C publication Critical patent/CN100452090C/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Landscapes

  • Image Processing (AREA)

Abstract

The invention dislcoses a method and a system for realizing a high undulation illumination range, and used for solve a problem that a video card which is not supported by a DirectX 9.0 standard can not process images by a high undulation illumination range technic. The invention method comprises: compressing a source floating point venation into a high bit integer venation during compressing, and bisecting the high bit integer venation into low bit integer venations; combining the low bit integer venations to the high bit integer venation during decompressing to realize an effect of the high undulation illumination range. With actualization of the invention, the video card which is not supported by the DirectX 9.0 standard can analogously process a 64-bit integer venation, and support an HDR technic, thereby processed images can be exposed to achieve a better shadow effect.

Description

A kind of method and system that realize high dynamic illumination range
Technical field
The present invention relates to image processing field, particularly relate to a kind of method and system that realize high dynamic illumination range.
Background technology
Computing machine and is to come the brightness of differentiate between images with 8bit (256) level during the display display image when presentation video, but obviously is that the gray scale of unit can't reproduce real light conditions with the hundreds of.For example: be limited in without each pixel in the generic graphic file of crossing high dynamic illumination range (HDR) technical finesse in 0 to 255 the brightness range.In the common daytime scenery picture, the white clouds and the sun may all present same brightness (all being pure white), but in fact white clouds can not be the same with the brightness of the sun.
The HDR technology is a kind of image processing techniques.Support the video card of DirectX9.0 standard can adopt the HDR technology source floating-point texture boil down to be not less than 64 integer textures.64 integer textures (each pixel is preserved by 64) have four passages of RGBA, account for 16 respectively, and preserve with round values.Pass through formula: R (16int)=R (float)/16int_max * 65535 pair R passage compresses (other passages are identical), and before CPU was to the compression of source floating-point texture, arbitrary passage all was 0 to 65535 round values in program; After compression and outputing in the video card, represent 0 to 1 brightness value by video card, wherein 0 expression is the darkest, and 1 expression is the brightest, and precision is 1/655535.For example: the value of R passage in program is 18, and then its value is 18/655535 in video card.After video card is to the compressed textures decompress(ion), can on display, show to exceed 0 to 1 brightness range.
The purpose of using the HDR technology is in order to unite the use exposure technique, to reach the better pictures effect.Exposure is the whole color and the brightness of scene to be adjusted exposure E＞0.Expose for non-HDR, then can produce color lump or make the whole scene deepening, the image and the reality that obtain do not match; Exposure is just meaningful for HDR.
Below describe respectively and expose carrying out the image that HDR handles, and the situation that 32 textures are exposed.
The image that carried out the HDR processing is exposed.HDR refers in particular to the HDR special efficacy in recreation, the HDR special efficacy is and image render effects arranged side by side such as Vertex Shader/ShaderModel/Soft shadows/Parallax Mapping.If will realize the HDR special efficacy, the video card of use must be supported the DirectX9.0 standard, and for example: the video card of nVIDIA must be GeForce 6 series or higher, ATI card be at least Radeon 9500 or more than.Referring to shown in Figure 1, for unexposed HDR image,, be the HDR image after the exposure referring to shown in Figure 2, obviously exposure back effect of shadow is better.
At present, also have some video cards not support the DirectX9.0 standard, promptly do not support HDR, can only adopt 32 textures.Each pixel of 32 texture fingerprint reasons is preserved by 32, each pixel has four passages of RGBA, account for 8 respectively, and preserve with round values, promptly wherein arbitrary passage all is 0 to 255 round values in program, and represents 0 to 1 color value in video card, and 0 expression is the darkest, 1 expression is the brightest, and precision is 1/255.For example: the value of R passage in program is 16, and then its value is 16/255 in video card.Referring to shown in Figure 3, in 32 texture images of non-HDR, if E=1 then can normally show.But 32 textures are directly exposed, referring to shown in Figure 4, during E=0.3, the whole image deepening; Referring to shown in Figure 5, during E=5, color lump appears in whole image.
In sum,, then can't adopt the HDR technology that image is handled, and then can't adopt exposure technique to produce effect of shadow more true to nature if user's use is the video card of not supporting the DirectX9.0 standard.
Summary of the invention
The invention provides a kind of method and system that realize high dynamic illumination range, in order to separate the problem that the video card of never supporting the DirectX9.0 standard can't adopt the high dynamic illumination range technology that image is handled.
The inventive method is used to carry out Flame Image Process, comprises the following steps:
When compression,, and should high position integer textures be divided into the low level integer textures with the high-order integer textures of source floating-point texture boil down to; When decompress(ion), each low level integer textures is merged into high-order integer textures, to realize the effect of high dynamic illumination range.
Described integer textures and source floating-point texture include 4 passages, and after high-order integer textures was waited branch, its arbitrary passage included N passage fragment from high to low, and the integer textures of each passage fragment of described arbitrary passage is:
R (0)＝0；
Figure C20061006576000071
Figure C20061006576000072
2≤n≤(N-1)；
Figure C20061006576000073
Wherein, R (H)Integer textures for the passage fragment of most significant digit; R (FLOAT)Arbitrary passage for the source floating-point texture; 16int_max is the maximal value of brightness behind the decompress(ion); H is the figure place of high-order each passage of integer textures; L is the figure place of each passage of low level integer textures; The ascending expression respectively of the value of n passage fragment from high to low; R (L)Integer textures for the passage fragment of lowest order.
After 64 integer textures were divided into 2 32 integer textures, its arbitrary passage included 28 bit ports from high to low, and wherein, the integer textures of the most-significant byte passage fragment of arbitrary passage is:
The integer textures of least-significant byte passage fragment is:
Figure C20061006576000075
After 128 integer textures were divided into 4 32 integer textures, its arbitrary passage included 48 bit ports from high to low, and wherein, the integer textures of each passage fragment of arbitrary passage is:
Figure C20061006576000076
Figure C20061006576000078
Figure C20061006576000079
It is described that each low level integer textures is merged into high-order integer textures is that 4 passages are run jointly respectively.
By following formula 4 passages are merged respectively, and decompress:
Figure C20061006576000081
Wherein, the ascending expression respectively of the value of m passage fragment from high to low.
2 32 integer textures are merged into 64 integer textures, behind arbitrary passage merging of these 64 integer textures and the decompress(ion) are:
Figure C20061006576000082
4 32 integer textures are merged into 128 integer textures, behind arbitrary passage merging of these 128 integer textures and the decompress(ion) are:
Figure C20061006576000083
The system of realization high dynamic illumination range of the present invention comprises:
Compression unit is used for the high-order integer textures of source floating-point texture boil down to;
The decompress(ion) unit is used for high-order integer textures is decompressed;
Cutting unit, be used for the high-order integer textures of described compression unit output is divided into the low level integer textures, described integer textures and source floating-point texture include 4 passages, after high-order integer textures is waited branch, its arbitrary passage includes N passage fragment from high to low
The integer textures of each passage fragment of described arbitrary passage is:
R (0)＝0；
Figure C20061006576000084
Figure C20061006576000085
2≤n≤(N-1)；
Figure C20061006576000086
Wherein, R (H)Integer textures for the passage fragment of most significant digit; R (FLOAT)Arbitrary passage for the source floating-point texture; 16int_max is the maximal value of brightness behind the decompress(ion); H is the figure place of high-order each passage of integer textures; L is the figure place of each passage of low level integer textures; The ascending expression respectively of the value of n passage fragment from high to low; R (L)Integer textures for the passage fragment of lowest order;
Merge cells is used for the low level integer textures of cutting unit output is merged into high-order integer textures, and outputs to the decompress(ion) unit, when described merging, by following formula 4 passages is merged respectively, and decompresses:
Figure C20061006576000091
Wherein, the ascending expression respectively of the value of m passage fragment from high to low.
Described high-order integer textures is minimum to be 64 integer textures.
Beneficial effect of the present invention is as follows:
When the video card that the inventive method is used the user is not supported the DirectX9.0 standard, earlier by CPU with the high-order integer textures (more than 64) of source floating-point texture boil down to; Afterwards, this high position integer textures is divided into the low level integer textures that video card that the user uses can be supported, for example: 64 integer textures are divided into 32 integer textures; When video card decompresses to integer textures, again each low level integer textures is merged, to simulate 64 integer textures.But the video card of not supporting the DirectX9.0 standard like this is 64 integer textures of simulation process also, and support the HDR technology, thereby can the image after handling be exposed, to reach better effect of shadow.
In order to support the inventive method, the present invention also provides a kind of system that realizes high dynamic illumination range, comprising: compression unit is used for the high-order integer textures of source floating-point texture boil down to; The decompress(ion) unit is used for high-order integer textures is decompressed; Cutting unit is used for the high-order integer textures of described compression unit output is divided into the low level integer textures; Merge cells is used for the low level integer textures of cutting unit output is merged into high-order integer textures, and outputs to the decompress(ion) unit.
Description of drawings
Fig. 1 is unexposed HDR image;
Fig. 2 is the HDR image after exposing;
Fig. 3 is 32 unexposed images of texture of non-HDR;
Fig. 4 is 32 textures of non-HDR, with the E=0.3 exposed images;
Fig. 5 is 32 textures of non-HDR, with the E=5 exposed images;
Fig. 6 is a system architecture synoptic diagram of the present invention;
Fig. 7 is the inventive method flow chart of steps;
Fig. 8 is a source floating-point texture design sketch;
Fig. 9 is in the method example one, and 64 integer textures are divided into two 32 integer textures synoptic diagram;
Figure 10 is in the method example one, most-significant byte passage segment design sketch;
Figure 11 is in the method example one, least-significant byte passage segment design sketch;
Figure 12 is in the method example one, behind merging and the decompress(ion), with the design sketch after the E=0.1 exposure;
Figure 13 is in the method example one, behind merging and the decompress(ion), with the design sketch after the E=0.5 exposure;
Figure 14 is in the method example one, behind merging and the decompress(ion), with the design sketch after the E=1.0 exposure;
Figure 15 is in the method example one, behind merging and the decompress(ion), with the design sketch after the E=2.0 exposure;
Figure 16 is in the method example one, behind merging and the decompress(ion), with the design sketch after the E=5.0 exposure.
Embodiment
In order to make the video card of not supporting the DirectX9.0 standard can adopt high dynamic illumination range HDR technology that image is handled, thereby reach better effect of shadow in conjunction with exposure technique, the invention provides a kind of system that realizes high dynamic illumination range, referring to shown in Figure 6, it comprises: the compression unit of Xiang Lianing, cutting unit, merge cells and decompress(ion) unit successively.
Described compression unit is arranged among the CPU, is used for the high-order integer textures of source floating-point texture boil down to.In order to support the HDR technology, described high-order integer textures is 64 integer textures, also can be 128 integer textures or higher.
Described cutting unit can be arranged among the CPU, is used for the high-order integer textures of described compression unit output is divided into the low level integer textures.Described low level integer textures is decided according to the video card that the user uses, and is generally 32 integer textures.
Described merge cells can be arranged in the video card, is used for the low level integer textures of cutting unit output is merged into high-order integer textures, and outputs to the decompress(ion) unit.
Described decompress(ion) unit is arranged in the video card, is used for high-order integer textures is decompressed, to realize the HDR effect.
Use said system, the invention provides a kind of method that realizes high dynamic illumination range,, comprise following key step referring to shown in Figure 7:
S1, with the high-order integer textures of source floating-point texture boil down to.
S2, high-order integer textures is divided into the low level integer textures.
S3, each low level integer textures is merged into high-order integer textures.
S4, high-order integer textures is decompressed.
Below specifically describe the inventive method by two examples.
Method example one: 64 integer textures are divided into 32 integer textures.
S101, with the high-order integer textures of source floating-point texture boil down to.
CPU obtains source floating-point texture (source floating-point texture effect is referring to Fig. 8) from program, and by compression unit with 64 integer textures of source floating-point texture boil down to.Source floating-point texture and integer textures include four passages of RGBA, and the compression process of each passage is all identical, below are example with the R passage.
Compression unit utilizes formula: R (FLOAT)/ 16int_max is with the R passage texture R of source floating-point texture (FLOAT), the R passage integer textures (16) in 64 integer textures of boil down to.Wherein 16int_max is the maximal value of brightness behind the decompress(ion), gets 100 usually.
S102, high-order integer textures is divided into the low level integer textures.
CPU is divided into two 32 integer textures by cutting unit with 64 integer textures, and referring to shown in Figure 9, each 16 bit port of 64 integer textures all are divided into two 8 bit port segments.The segmentation process of each passage is all identical, below is example with the R passage.
Cutting unit utilizes branch R passages such as following formula:
R (0)＝0；
Figure C20061006576000111
Figure C20061006576000121
2≤n≤(N-1)；
Figure C20061006576000122
Wherein, R (H)Integer textures for the passage fragment of most significant digit; R (FLOAT)R passage for the source floating-point texture; 16int_max is the maximal value of brightness behind the decompress(ion); H is the figure place of high-order each passage of integer textures; L is the figure place of each passage of low level integer textures; The ascending expression respectively of the value of n passage fragment from high to low; N is the passage number of fragments that arbitrary passage of high-order integer textures comprises; R (L)Integer textures for the passage fragment of lowest order.
In this example, h=16, l=8, N=2; Owing to only have R in this example (H)And R (L)The integer textures of two passage fragments obtains so bring the value of h, l and N into corresponding formulas:
Figure C20061006576000123
Figure C20061006576000124
Referring to shown in Figure 10, during for 16int_max=100, R (H)Grain effect figure; Referring to shown in Figure 11, during for 16int_max=100, R (L)Grain effect figure.
S103, each low level integer textures is merged into high-order integer textures.
CPU outputs to described two 32 integer textures in the video card.Video card merges respectively 4 passages by merge cells.The merging processing procedure of each passage is all identical, below is example with the R passage.
Merge cells passes through formula:
Figure C20061006576000125
The R passage is merged.Wherein, the ascending expression respectively of the value of m passage fragment from high to low.The R passage of 64 integer textures after merging in this example is
Figure C20061006576000126
So far 4 passages of 64 integer textures have been obtained simulating.
S104, high-order integer textures is decompressed.
Video card decompresses respectively to 4 passages by the decompress(ion) unit.The decompression processing process of each passage is all identical, below is example with the R passage.
The decompress(ion) unit multiplies each other the R passage texture and the 16int_max of merge cells output, promptly
Figure C20061006576000131
Obtain the output texture behind the decompress(ion).
Afterwards, can adopt exposure technique that the texture of output is exposed,, be respectively the design sketch of E=0.1, E=0.5, E=1.0, E=2.0 and E=5.0 referring to Figure 12, Figure 13, Figure 14, Figure 15 and shown in Figure 16.
Method example two: 128 integer textures are divided into 32 integer textures.
S201, with the high-order integer textures of source floating-point texture boil down to.
CPU obtains the source floating-point texture from program, and by compression unit with 128 integer textures of source floating-point texture boil down to.Source floating-point texture and integer textures include four passages of RGBA, and the compression process of each passage is all identical, below are example with the R passage.
Compression unit utilizes formula: R (FLOAT)/ 16int_max is with the R passage texture R of source floating-point texture (FLOAT), the R passage integer textures (32) in 128 integer textures of boil down to.
S202, high-order integer textures is divided into the low level integer textures.
After CPU was divided into 4 32 integer textures by cutting unit with 128 integer textures, each 32 bit port of 128 integer textures included 48 bit port segments from high to low.The segmentation process of each passage is all identical, below is example with the R passage.
Cutting unit utilizes branch R passages such as following formula:
R (0)＝0；
Figure C20061006576000132
2≤n≤(N-1)；
Figure C20061006576000134
Wherein, R (H)Integer textures for the passage fragment of most significant digit; R (FLOAT)R passage for the source floating-point texture; 16int_max is the maximal value of brightness behind the decompress(ion); H is the figure place of high-order each passage of integer textures; L is the figure place of each passage of low level integer textures; The ascending expression respectively of the value of n passage fragment from high to low; N is the passage number of fragments that arbitrary passage of high-order integer textures comprises; R (L)Integer textures for the passage fragment of lowest order.
In this example, h=32, l=8, N=4; So bringing the value of h, l and N into corresponding formulas obtains:
Figure C20061006576000141
Figure C20061006576000142
Figure C20061006576000143
Figure C20061006576000144
S203, each low level integer textures is merged into high-order integer textures.
CPU outputs to described 4 32 integer textures in the video card.Video card merges respectively 4 passages by merge cells.The merging processing procedure of each passage is all identical, below is example with the R passage.
Merge cells passes through formula:
Figure C20061006576000145
The R passage is merged.Wherein, the ascending expression respectively of the value of m passage fragment from high to low.The R passage of 128 integer textures after merging in this example is
Figure C20061006576000146
So far 4 passages of 128 integer textures have been obtained simulating.
S204, high-order integer textures is decompressed.
Video card decompresses respectively to 4 passages by the decompress(ion) unit.The decompression processing process of each passage is all identical, below is example with the R passage.
The decompress(ion) unit multiplies each other the R passage texture and the 16int_max of merge cells output, promptly
Figure C20061006576000147
Obtain the output texture behind the decompress(ion).
Afterwards, can adopt exposure technique that the texture of output is exposed.
Obviously, those skilled in the art can carry out various changes and modification to the present invention and not break away from the spirit and scope of the present invention.Like this, if of the present invention these are revised and modification belongs within the scope of claim of the present invention and equivalent technologies thereof, then the present invention also is intended to comprise these changes and modification interior.

Claims (8)

1, a kind of method that realizes high dynamic illumination range is used to carry out Flame Image Process, it is characterized in that, comprises the following steps:
When compression,, and should high position integer textures be divided into the low level integer textures with the high-order integer textures of source floating-point texture boil down to, described integer textures and source floating-point texture include 4 passages, after high-order integer textures was waited branch, its arbitrary passage included N passage fragment from high to low
The integer textures of each passage fragment of described arbitrary passage is:
R (0)＝0；
Figure C2006100657600002C1
Figure C2006100657600002C2
2≤n≤(N-1)；
Figure C2006100657600002C3
Wherein, R (H)Integer textures for the passage fragment of most significant digit; R (FLOAT)Arbitrary passage for the source floating-point texture; 16int_max is the maximal value of brightness behind the decompress(ion); H is the figure place of high-order each passage of integer textures; L is the figure place of each passage of low level integer textures; The ascending expression respectively of the value of n passage fragment from high to low; R (L)Integer textures for the passage fragment of lowest order;
And
When decompress(ion), each low level integer textures is merged into high-order integer textures, to realize the effect of high dynamic illumination range, when described merging, 4 passages are merged respectively, and decompress by following formula:
R = [ Σ m = 1 N ( R ( m ) / 2 ^ [ h - ( N - m + 1 ) l ] ) ] × 16 int _ max ;
Wherein, the ascending expression respectively of the value of m passage fragment from high to low.
2, the method for claim 1 is characterized in that, 64 integer textures are divided into 2 32 integer textures after, its arbitrary passage includes 28 bit ports from high to low,
Wherein, the integer textures of the most-significant byte passage fragment of arbitrary passage is:
Figure C2006100657600003C1
The integer textures of least-significant byte passage fragment is:
3, the method for claim 1 is characterized in that, 128 integer textures are divided into 4 32 integer textures after, its arbitrary passage includes 48 bit ports from high to low,
Wherein, the integer textures of each passage fragment of arbitrary passage is:
Figure C2006100657600003C3
Figure C2006100657600003C4
Figure C2006100657600003C5
Figure C2006100657600003C6
4, the method for claim 1 is characterized in that, described each low level integer textures is merged into high-order integer textures is that 4 passages are merged respectively.
5, the method for claim 1 is characterized in that, 2 32 integer textures are merged into 64 integer textures, behind arbitrary passage merging of these 64 integer textures and the decompress(ion) is:
R = [ Σ m = 1 2 ( R ( m ) / 2 ^ [ 16 - ( 2 - m + 1 ) 8 ] ) ] × 16 int _ max .
6, the method for claim 1 is characterized in that, 4 32 integer textures are merged into 128 integer textures, behind arbitrary passage merging of these 128 integer textures and the decompress(ion) is:
R = [ Σ m = 1 4 ( R ( m ) / 2 ^ [ 32 - ( 4 - m + 1 ) 8 ] ) ] × 16 int _ max .
7, a kind of system that realizes high dynamic illumination range comprises:
Compression unit is used for the high-order integer textures of source floating-point texture boil down to;
The decompress(ion) unit is used for high-order integer textures is decompressed;
It is characterized in that described system also comprises:
Cutting unit, be used for the high-order integer textures of described compression unit output is divided into the low level integer textures, described integer textures and source floating-point texture include 4 passages, after high-order integer textures is waited branch, its arbitrary passage includes N passage fragment from high to low
The integer textures of each passage fragment of described arbitrary passage is:
R (0)＝0；
Figure C2006100657600004C1
Figure C2006100657600004C2
2≤n≤(N-1)；
Wherein, R (H)Integer textures for the passage fragment of most significant digit; R (FLOAT)Arbitrary passage for the source floating-point texture; 16int_max is the maximal value of brightness behind the decompress(ion); H is the figure place of high-order each passage of integer textures; L is the figure place of each passage of low level integer textures; The ascending expression respectively of the value of n passage fragment from high to low; R (L)Integer textures for the passage fragment of lowest order;
Merge cells is used for the low level integer textures of cutting unit output is merged into high-order integer textures, and outputs to the decompress(ion) unit, when described merging, by following formula 4 passages is merged respectively, and decompresses:
R = [ Σ m = 1 N ( R ( m ) / 2 ^ [ h - ( N - m + 1 ) l ] ) ] × 16 int _ max ;
Wherein, the ascending expression respectively of the value of m passage fragment from high to low.
8, system as claimed in claim 7 is characterized in that, described high-order integer textures is minimum to be 64 integer textures.
CNB2006100657606A 2006-03-14 2006-03-14 Method and system for implementing high dynamic light range Active CN100452090C (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CNB2006100657606A CN100452090C (en) 2006-03-14 2006-03-14 Method and system for implementing high dynamic light range

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CNB2006100657606A CN100452090C (en) 2006-03-14 2006-03-14 Method and system for implementing high dynamic light range

Publications (2)

Publication Number Publication Date
CN101038660A CN101038660A (en) 2007-09-19
CN100452090C true CN100452090C (en) 2009-01-14

Family

ID=38889544

Family Applications (1)

Application Number Title Priority Date Filing Date
CNB2006100657606A Active CN100452090C (en) 2006-03-14 2006-03-14 Method and system for implementing high dynamic light range

Country Status (1)

Country Link
CN (1) CN100452090C (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR101699919B1 (en) * 2011-07-28 2017-01-26 삼성전자주식회사 High dynamic range image creation apparatus of removaling ghost blur by using multi exposure fusion and method of the same
US10334141B2 (en) * 2017-05-25 2019-06-25 Denso International America, Inc. Vehicle camera system

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6560285B1 (en) * 1998-03-30 2003-05-06 Sarnoff Corporation Region-based information compaction as for digital images
CN1483382A (en) * 2003-07-28 2004-03-24 沈阳东软数字医疗系统股份有限公司 Compressing method for original CT image data
CN1595390A (en) * 1998-03-31 2005-03-16 英特尔公司 Executing partial-width packed data instructions
WO2005101309A1 (en) * 2004-04-15 2005-10-27 The University Of British Columbia Methods and systems for converting images from low dynamic range to high dynamic range
US6993200B2 (en) * 2001-11-20 2006-01-31 Sony Corporation System and method for effectively rendering high dynamic range images

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6560285B1 (en) * 1998-03-30 2003-05-06 Sarnoff Corporation Region-based information compaction as for digital images
CN1595390A (en) * 1998-03-31 2005-03-16 英特尔公司 Executing partial-width packed data instructions
US6993200B2 (en) * 2001-11-20 2006-01-31 Sony Corporation System and method for effectively rendering high dynamic range images
CN1483382A (en) * 2003-07-28 2004-03-24 沈阳东软数字医疗系统股份有限公司 Compressing method for original CT image data
WO2005101309A1 (en) * 2004-04-15 2005-10-27 The University Of British Columbia Methods and systems for converting images from low dynamic range to high dynamic range

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
在可编程图形硬件上实现图像高动态范围压缩. 彭韬，胡耀华，李在铭.微机发展，第15卷第9期. 2005 *

Also Published As

Publication number Publication date
CN101038660A (en) 2007-09-19

Similar Documents

Publication Publication Date Title
US9904977B2 (en) Exploiting frame to frame coherency in a sort-middle architecture
US10636336B2 (en) Mixed primary display with spatially modulated backlight
EP2854108B1 (en) Anti-aliasing for graphics hardware
TWI582751B (en) Avoiding sending unchanged regions to display
US8605104B1 (en) Threshold-based lossy reduction color compression
US9754345B2 (en) Compression and decompression of graphics data using pixel region bit values
EP1980998A2 (en) Programmable graphics processing element
CN1870055A (en) High dynamic range material color applying drawing method
US9418471B2 (en) Compact depth plane representation for sort last architectures
WO2014204703A1 (en) Subset based compression and decompression of graphics data
US9741154B2 (en) Recording the results of visibility tests at the input geometry object granularity
Goodnight et al. Computation on programmable graphics hardware
US9582847B2 (en) Color buffer compression
US20180097527A1 (en) 32-bit hdr pixel format with optimum precision
CN100452090C (en) Method and system for implementing high dynamic light range
US8655063B2 (en) Decoding system and method operable on encoded texture element blocks
US20150103252A1 (en) System, method, and computer program product for gamma correction in a video or image processing engine
US9336561B2 (en) Color buffer caching
CN100565576C (en) A kind of method and system that realize high dynamic illumination range
US20070211068A1 (en) Reconfigurable floating point filter
US20180040095A1 (en) Dynamic compressed graphics state references
US9171523B2 (en) GPU-accelerated, two-pass colorspace conversion using multiple simultaneous render targets
US5856831A (en) Clamping system and method for clamping floating point color values from a geometry accelerator in a computer graphics system
CN108648138B (en) High dynamic data processing method for HDR format coding
TW202324292A (en) Non-linear filtering for color space conversions

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C14 Grant of patent or utility model
GR01 Patent grant