CN100414959C - Method for forming preview image - Google Patents
Method for forming preview image Download PDFInfo
- Publication number
- CN100414959C CN100414959C CNB2006100989033A CN200610098903A CN100414959C CN 100414959 C CN100414959 C CN 100414959C CN B2006100989033 A CNB2006100989033 A CN B2006100989033A CN 200610098903 A CN200610098903 A CN 200610098903A CN 100414959 C CN100414959 C CN 100414959C
- Authority
- CN
- China
- Prior art keywords
- preview image
- filtering
- image
- carried out
- decode
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Landscapes
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
The present invention relates to a method for forming a preview image in a video frequency decoding and display device. The present invention divides the necessary filtering operation into a decoding procedure and a display procedure for realization: executing the two-dimension frequency domain filtering for a DCT coefficient in the decoding procedure and executing the smooth filtering in the displaying procedure to eliminate the block effect caused by the two-dimension frequency domain filtering for a DCT coefficient. The present invention greatly reduces the computation amount and the resource cost. Due to the decrease of the computation amount, the real-time of the treatment of each link is improved and consequently the efficiency of forming the preview image is improved.
Description
Technical field
The invention belongs to field of multimedia signal processing, generate the method for preview image in particularly a kind of video decode and the display unit.
Background technology
Present SD and high-definition digital video transmission system, for example DVB-S, DVB-C, DVB-T adopt MPEG-2 as standard.The system layer of MPEG-2 has defined the transport stream of the multi-channel video flow data being utilized the synthetic single channel of time division multiplexing, and the video section of MPEG-2 has defined in the Code And Decode stage that vision signal is handled and adopted discrete cosine transform (DCT) and inverse discrete cosine transformation (IDCT) to carry out the conversion of spatial domain and frequency domain respectively.
Comprise multi-path video stream in single transport stream, the user just relates to switching and the program search problem between the multi-path video stream when watching.Improve the efficient of program search for the convenience of the user, the video equipment terminal can provide the function of video navigation and video preview usually.The mode that video navigation and video preview provide has a lot: the single-unit activity picture adds combination of Video stream information table, a plurality of video still frame, a plurality of video moving frame, a plurality of still frame and moving frame or the like.Here be called " preview image " in static or movable this specification of picture of mentioning.
Existing a kind of method that generates preview image is: the image that directly decoding is obtained according to target size carries out the down-sampled conversion of level and vertical direction, this method realizes simple, but can cause serious spectral aliasing and produce serious image quality loss when making broadcast, such as phenomenons such as interline flickers.
The improved method of existing another kind is: did the filtering operation on the primary space territory before the image that decoding is obtained carries out level and vertical direction down-sampled earlier, can restrain the spectral aliasing after image dwindles like this and improve image quality, but this method needs extra filtering operation and bigger memory space, expend more resource, and have a strong impact on the formation speed of preview image, reduced efficient.
Summary of the invention
Order of the present invention overcomes the deficiencies in the prior art exactly, a kind of method of new generation preview image is proposed, with littler resource overhead, overcome produce in the preview image process because the image quality loss that spectral aliasing causes, further accelerate the formation speed of preview image simultaneously.
For achieving the above object, the present invention has been assigned to decoding to essential filtering operation and has shown that two links realize, at decode phase the DCT coefficient is carried out the two-dimensional frequency characteristic filtering, carry out smothing filtering elimination decode phase in the demonstration stage DCT coefficient carried out the blocking effect that the two-dimensional frequency characteristic filtering causes, specifically comprise following steps:
A. in the IDCT of decoder calculating process, transform domain pixel block behind the dct transform is carried out the two dimensional character frequency domain filtering according to original image size and preview image size, and the process of carrying out the two-dimensional frequency characteristic filtering is that coefficient in the DCT piece and two dimensional character filter template correspondence are multiplied each other;
B. final decode image data is carried out the down-sampled of integer multiple according to foundation original image size and preview image size, the result is stored in the preview image space;
C. from the preview image space, take out the preview image data that step B obtains, according to final preview image display size the preview image data that step B obtains are carried out convergent-divergent, the preview image data that step B is obtained are carried out simple smothing filtering on the spatial domain simultaneously, eliminate blocking effect;
D. according to the application target requirement, the screen message according to current video is play is spliced into complete target image with each preview image in real time;
E. cover Information Level and form final broadcast target image finishing the screen of covering one deck on the target image of splicing in real time and provide background pure color or substrate picture and descriptive information.
Decoder in the described steps A is with the decode system of dct transform as the transition coding algorithm.
Final decode image data among the described step B is to be obtained after IDCT computing and other necessary decode operations by the DCT coefficient through the two-dimensional frequency characteristic filtering in the steps A.
The down-sampled of integer multiple among the described step B is the most down-sampled near the integral multiple of target scaling; At least comprise 1,2,4 etc. 2 integral number power in the horizontal and vertical directions.
A key technology of the present invention is the two-dimensional frequency characteristic filtering in the steps A, because the DCT coefficient itself before the IDCT is just represented the size of frequency domain components, so as long as each DCT coefficient in this coefficient matrix is carried out the two-dimensional frequency characteristic filtering that simple multiplying factor operation can realize enough accuracy, operand is very little, and resource overhead is very little.
Another key technology of the present invention is integer multiple down-sampled among the step B, because being placed among the step C, strict display size coupling and convergent-divergent realize, as long as so carry out the most down-sampledly near the integral multiple of target scaling among the step B, operand is very little, resource overhead is very little.
Another key technology of the present invention is a smothing filtering among the step C, owing to finished the two-dimensional frequency characteristic filtering of enough accuracy in the steps A, radio-frequency component in the preview image is compared desirable inhibition, as long as get final product the blocking effect that removal process A causes so carry out simple smothing filtering among the step C, obviously reduce with respect to the required filter operand of conventional method, resource overhead reduces simultaneously.
The method of generation preview image of the present invention is on the basis of existing generation method, essential filtering operation has been assigned to decoding and has shown that two links realize, and utilize the implication of DCT coefficient self to carry out efficient 2-d frequency domain character filtering more, significantly reduced the operand and the resource overhead of each link at decode phase.Owing to the minimizing of operand, the real-time that each link is handled improves simultaneously, thereby has improved the efficient that generates preview image.
Description of drawings
Fig. 1 is the hardware configuration schematic diagram of one embodiment of the invention;
Fig. 2 is the internal data handling process view among Fig. 1.
Specific implementation method
Following will describe the course of work of the method for generation preview image of the present invention in conjunction with a specific embodiment of the present invention in detail.
Fig. 1 is the hardware configuration schematic diagram of present embodiment.Wherein, video stream data extracts 11 modules are responsible for extracting compression from transport stream original video flow data.This extraction comprises from transport stream extracts multichannel original video flow data and from the target original video flow data of multichannel original video stream extracting data appointment.
SDRAM 12 modules are data switching center of whole system, play the effect that metadata cache and module are isolated.Exchanges data comprises and receives the original video flow data, provides video decode 13 module decoded datas, stores decoded view data, and storage is the preview image data after down-sampled to decoded picture, cover Information Level for screen memory space is provided.
Down-sampled 14 modules receive in real time decoded original image pixel block, according to the information such as evaluation criterion of the suitableeest target preview image size of the target size of the size of original image, preview image and employing, carry out the down-sampled conversion of basic multiple.Because the basic multiple that adopts is 2 power, the resource of down-sampled needs is considerably less, and execution speed is very fast.Data after down-sampled write back to SDRAM.
Reprocessing 15 modules read the preview image corresponding data according to broadcast information from SDRAM, the selection of dimension suitable dimensions matching process according to the target size of preview image and basic multiple after down-sampled carries out the size match of preview image.Another major function of post-processing module is that the ring that produces because of the two dimensional character characteristic filtering in the preview graph and edge block effect are carried out supression and smothing filtering on the spatial domain.
Splice 16 modules and receive the data of coming, finish the operations such as integration of regional size and Position Control, the filling of transition band, stacked treatment of picture, image from reprocessing 15 modules.
Information Level 17 modules that superpose read in the corresponding data that the screen of storing among the SDRAM is covered Information Level according to video playback information, do integration with the preview image data of splicing 16 modules input, generate final playing image.
Illustrate below in conjunction with Fig. 2 how method of the present invention realizes on hardware system shown in Figure 1.Fig. 2 is the MPEG-2 decoder embodiment of a standard, mainly comprise: a variable length code (VLD) decoding unit 21, counter-scanning and inverse quantization processing unit 22, an IDCT processing unit 24, the reference frame storing unit 27 of 25, one stored reference predicted pictures of a motion compensation (MC) processing unit.The original video stream of the compression of extracting from transport stream at first enters VLD decoder 21, decoded data carry out counter-scanning and inverse quantization is handled the DCT piece that is reduced into 8 * 8 sizes, the IDCT processing unit is the pixel block that unit is transformed into the DCT piece 8 * 8 in the spatial domain with the piece, the image that all pixel blocks of 8 * 8 are formed in one frame/field is a residual image, and the image that need be superimposed with after the motion compensation forms final decoded picture.Motion compensation is finished by the another one data path, and motion compensation units 25 is extracted motion vector from VLD decoder 21, extracts corresponding estimation image according to motion vector from reference frame is deposited.Residual image and motion compensated image form final decoded picture by the stack of synthesis unit 26, and the partial decoding of h image will be deposited into during reference image frame deposits, for the decode procedure of back provides reference data.
The present invention has increased very little but the processing unit that specific aim is very strong of some expenses in the video decoding system of standard, and do not influence the decode procedure of primary standard system, comprise: DCT filter processing unit 23, down-sampled processing unit 28, post-processing unit 29, image splicing processing unit 30, information covers processing unit 31 and video playback processing unit 32.
The process object of DCT filter processing unit 23 is through the DCT piece after counter-scanning and the inverse quantization processing, comprises 8 * 8 discrete cosine transform coefficients, and the method for employing is based on the two-dimensional frequency characteristic filtering of 8 * 8 filter template.The process of carrying out the two-dimensional frequency characteristic filtering is that coefficient in the DCT piece and two-dimensional frequency characteristic filtering device template correspondence are multiplied each other.The meaning of " feature " is that this filter is not changeless, need be configured according to the characteristic of original video flow data and the requirement of target preview image.Method that for example can classification according to the requirement of target preview image, is divided into high, medium and low rank with filter coefficient, the bandwidth difference of each rank filter, and the final preview image characteristic that generates is also different.
28 pairs of decoded images of down-sampled processing unit are done the once down-sampled operation of basic multiple simultaneously in the process that writes.So-called basic multiple is the corresponding existing down-sampled process of any multiple that is of a size of target with preview image.Here the basic multiple of Cai Yonging comprises 1,2,4 in the horizontal and vertical directions at least, and they all are 2 power, and the speed that deals with is very fast.The down-sampled substantially multiple that the image of different size is adopted is different, depends on the evaluation criterion of employing, and best interior zone, optimal external zone and optimum Match zone are optionally arranged.
The processing that post-processing unit 29 is done mainly contains two aspects, is the size match of doing preview image on the one hand, on the other hand preview image is carried out simple reprocessing on the spatial domain.The process of the size match of preview image is to be that benchmark mates with the target size of preview image through the picture size of basic multiple after down-sampled, the method of coupling is chosen flexibly according to two sizes and video information, inclusion region filling at least, regional intercepting, area extension and zone compression.Because the two-dimensional frequency characteristic filtering is undesirable in the transform domain, phenomenons such as image border diffusion, edges of regions blocking effect, ring in various degree may appear in filtered image, so processing such as done compensation on spatial domain, restrain and seamlessly transitted.
The processing procedure of image splicing processing unit 30 is real-time, and it is according to the information such as play position of video playback module feedback, and reading corresponding data is spliced processing from the memory block, has saved a large amount of resources.The method that splicing is adopted is decided on the requirement of concrete picture tissue, comprises the methods such as integration of regional size and Position Control, the filling of transition band, stacked treatment of picture, image at least.
Information covers superpose some necessary information and image is done the modification of a little backgrounds of processing unit 31 on spliced image.The process of covering is an in real time corresponding process, and it is according to the information such as play position of video playback module feedback, and reading corresponding data is covered processing from the memory block, has saved a large amount of resources.Screen is covered information and is comprised background colour, substrate picture and descriptive information etc. at least.
Video playback processing unit 32 is that the image demonstration is necessary, and it becomes to meet the data flow of display terminal form according to the display terminal type information of configuration with the playing image digital coding that receives.
As mentioned above, the method of generation preview image of the present invention has been assigned to decoding to essential filtering operation and has shown that two links realize, and utilize the implication of DCT coefficient self to carry out efficient 2-d frequency domain character filtering more, significantly reduced the operand and the resource overhead of each link at decode phase.Owing to the minimizing of operand, the real-time that each link is handled improves simultaneously, thereby has improved the efficient that generates preview image.
Although the present invention is next specifically described with reference to its preferred embodiment, it should be appreciated by those skilled in the art, under the situation that does not break away from the spirit and scope of the present invention that the claims qualification is arranged, can carry out the various modifications of form and details to it.
Claims (4)
1. generate the method for preview image, it is characterized in that this method has been assigned to decoding to essential filtering operation and show that two links realize, at decode phase the DCT coefficient is carried out the two-dimensional frequency characteristic filtering, carry out smothing filtering elimination decode phase in the demonstration stage DCT coefficient carried out the blocking effect that the two-dimensional frequency characteristic filtering causes, specifically comprise following steps:
A. in the IDCT of decoder calculating process, transform domain pixel block behind the dct transform is carried out the two-dimensional frequency characteristic filtering according to original image size and preview image size, and the process of carrying out the two-dimensional frequency characteristic filtering is that coefficient in the DCT piece and two dimensional character filter template correspondence are multiplied each other;
B. final decode image data is carried out the down-sampled of integer multiple according to original image size and preview image size, the result is stored in the preview image space;
C. from the preview image space, take out the preview image data that step B obtains, according to final preview image display size the preview image data that step B obtains are carried out convergent-divergent, the preview image data that step B is obtained are carried out simple smothing filtering on the spatial domain simultaneously, eliminate blocking effect;
D. according to the application target requirement, the screen message according to current video is play is spliced into complete target image with each preview image in real time;
E. cover one deck on the target image of splicing in real time background pure color or substrate picture are provided finishing, and the screen of covering descriptive information is simultaneously covered Information Level and formed final broadcast target image.
2. the method for generation preview image as claimed in claim 1 is characterized in that the decoder in the steps A is with the decode system of dct transform as the transition coding algorithm.
3. the method for generation preview image as claimed in claim 1 is characterized in that the final decode image data among the step B is to be obtained after the IDCT computing by the DCT coefficient through the two dimensional character frequency domain filtering in the steps A.
4. the method for generation preview image as claimed in claim 1 is characterized in that the down-sampled of integer multiple among the step B is the most down-sampled near the integral multiple of target scaling; At least comprise 2 integral number power in the horizontal and vertical directions.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2006100989033A CN100414959C (en) | 2006-07-14 | 2006-07-14 | Method for forming preview image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2006100989033A CN100414959C (en) | 2006-07-14 | 2006-07-14 | Method for forming preview image |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN1878229A CN1878229A (en) | 2006-12-13 |
| CN100414959C true CN100414959C (en) | 2008-08-27 |
Family
ID=37510511
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2006100989033A Active CN100414959C (en) | 2006-07-14 | 2006-07-14 | Method for forming preview image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100414959C (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101552897B (en) * | 2008-07-11 | 2012-12-19 | 深圳市创凯电子有限公司 | Method and apparatus for achieving the display of mosaicing high definition base image by using a U disk or/and a flash memory |
| CN101427924B (en) * | 2008-07-16 | 2010-07-28 | 山东省肿瘤医院 | A Method for Obtaining Complete Anatomical Images by Stitching Segmented Cone Beam CT Images |
| CN102982828B (en) * | 2012-11-22 | 2015-11-25 | 北京百度网讯科技有限公司 | The method of the preview file of generating video file and device |
| CN103049908B (en) * | 2012-12-10 | 2016-08-03 | 北京百度网讯科技有限公司 | The method and apparatus generating stereoscopic video file |
| US11403782B2 (en) * | 2019-12-31 | 2022-08-02 | Alibaba Group Holding Limited | Static channel filtering in frequency domain |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH09182075A (en) * | 1995-12-18 | 1997-07-11 | Lucent Technol Inc | Image transmission method and system |
| CN1422075A (en) * | 2002-12-26 | 2003-06-04 | 上海交通大学 | Parallel image sequence bit rate controlling method for digital TV video coder |
| US20050074182A1 (en) * | 1997-07-16 | 2005-04-07 | Samsung Electronics Co., Ltd. | Signal adaptive filtering method, signal adaptive filter and computer readable medium for storing program therefor |
-
2006
- 2006-07-14 CN CNB2006100989033A patent/CN100414959C/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH09182075A (en) * | 1995-12-18 | 1997-07-11 | Lucent Technol Inc | Image transmission method and system |
| US20050074182A1 (en) * | 1997-07-16 | 2005-04-07 | Samsung Electronics Co., Ltd. | Signal adaptive filtering method, signal adaptive filter and computer readable medium for storing program therefor |
| CN1422075A (en) * | 2002-12-26 | 2003-06-04 | 上海交通大学 | Parallel image sequence bit rate controlling method for digital TV video coder |
Also Published As
| Publication number | Publication date |
|---|---|
| CN1878229A (en) | 2006-12-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| TWI572196B (en) | Projected interpolation prediction generation for next generation video coding | |
| RU2691962C2 (en) | Image processing device and image processing method | |
| CN106385583B (en) | Image processing apparatus and method | |
| JP6163674B2 (en) | Content adaptive bi-directional or functional predictive multi-pass pictures for highly efficient next-generation video coding | |
| DE112007000359B4 (en) | Processing of video data | |
| US20080123750A1 (en) | Parallel deblocking filter for H.264 video codec | |
| TW315570B (en) | ||
| USRE45152E1 (en) | Data processing apparatus, image processing apparatus, and methods and programs for processing image data | |
| US20100232504A1 (en) | Supporting region-of-interest cropping through constrained compression | |
| JP5602367B2 (en) | Scaler structure for image and video processing | |
| JP5230735B2 (en) | Decoding device, decoding method, and receiving device | |
| CN102045578B (en) | Image processing apparatus and image processing method | |
| CN109644269A (en) | Image processing equipment, image processing method and program | |
| KR20220019288A (en) | Apparatus and method for filtering in video coding | |
| KR20050058965A (en) | Decoding device, electronics device, computer, decoding method, program and recording medium | |
| US20130163676A1 (en) | Methods and apparatus for decoding video signals using motion compensated example-based super-resolution for video compression | |
| CN100414959C (en) | Method for forming preview image | |
| KR20220112783A (en) | Block-based compression autoencoder | |
| CN105072353A (en) | Image decoding mosaic control algorism based on multiple GPUs | |
| TWI523498B (en) | An image coding method, an image decoding method, an image coding apparatus, and an image decoding apparatus | |
| EP1147671B1 (en) | Method and apparatus for performing motion compensation in a texture mapping engine | |
| JP2008506294A (en) | Method and system for performing deblocking filtering | |
| JP2002314998A (en) | Method for reducing format concurrently and method for decoding encoded video signal | |
| CN101883276A (en) | Multi-format HD Video Decoder Architecture Based on Joint Decoding of Software and Hardware | |
| Jia et al. | Learning-based conditional image coder using color separation |
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 | ||
| C56 | Change in the name or address of the patentee | ||
| CP03 | Change of name, title or address |
Address after: No. 90, Wensanlu Road, Hangzhou, Zhejiang, Xihu District Patentee after: Hangzhou National Chip Science & Technology Co., Ltd. Address before: 5A, Neusoft building, No. 99 Huaxing Road, Hangzhou, Zhejiang, Xihu District Patentee before: Hangzhou Guoxin Science & Technology Co., Ltd. |