CN1166110A - Method and apparatus for digital image decoding - Google Patents

Method and apparatus for digital image decoding Download PDF

Info

Publication number
CN1166110A
CN1166110A CN97109507A CN97109507A CN1166110A CN 1166110 A CN1166110 A CN 1166110A CN 97109507 A CN97109507 A CN 97109507A CN 97109507 A CN97109507 A CN 97109507A CN 1166110 A CN1166110 A CN 1166110A
Authority
CN
China
Prior art keywords
data
pixel
frame memory
decoding
compression
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.)
Granted
Application number
CN97109507A
Other languages
Chinese (zh)
Other versions
CN1158876C (en
Inventor
大平英雄
浅野研一
田敏明
浅井光太郎
村上笃道
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.)
Mitsubishi Electric Corp
Original Assignee
Mitsubishi Electric Corp
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 Mitsubishi Electric Corp filed Critical Mitsubishi Electric Corp
Publication of CN1166110A publication Critical patent/CN1166110A/en
Application granted granted Critical
Publication of CN1158876C publication Critical patent/CN1158876C/en
Anticipated expiration legal-status Critical
Expired - Fee Related legal-status Critical Current

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
    • H04N19/503—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving temporal prediction
    • H04N19/51—Motion estimation or motion compensation
    • H04N19/577—Motion compensation with bidirectional frame interpolation, i.e. using B-pictures
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
    • H04N19/115—Selection of the code volume for a coding unit prior to coding
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
    • H04N19/124—Quantisation
    • H04N19/126—Details of normalisation or weighting functions, e.g. normalisation matrices or variable uniform quantisers
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/134—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
    • H04N19/136—Incoming video signal characteristics or properties
    • H04N19/14—Coding unit complexity, e.g. amount of activity or edge presence estimation
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/134—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
    • H04N19/146—Data rate or code amount at the encoder output
    • H04N19/15—Data rate or code amount at the encoder output by monitoring actual compressed data size at the memory before deciding storage at the transmission buffer
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/134—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
    • H04N19/157—Assigned coding mode, i.e. the coding mode being predefined or preselected to be further used for selection of another element or parameter
    • H04N19/159—Prediction type, e.g. intra-frame, inter-frame or bidirectional frame prediction
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
    • H04N19/17—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object
    • H04N19/172—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object the region being a picture, frame or field
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
    • H04N19/177—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being a group of pictures [GOP]
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/42—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals characterised by implementation details or hardware specially adapted for video compression or decompression, e.g. dedicated software implementation
    • H04N19/423—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals characterised by implementation details or hardware specially adapted for video compression or decompression, e.g. dedicated software implementation characterised by memory arrangements
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/42—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals characterised by implementation details or hardware specially adapted for video compression or decompression, e.g. dedicated software implementation
    • H04N19/423—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals characterised by implementation details or hardware specially adapted for video compression or decompression, e.g. dedicated software implementation characterised by memory arrangements
    • H04N19/426—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals characterised by implementation details or hardware specially adapted for video compression or decompression, e.g. dedicated software implementation characterised by memory arrangements using memory downsizing methods
    • H04N19/428—Recompression, e.g. by spatial or temporal decimation
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/60—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding
    • H04N19/61—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding in combination with predictive coding
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
    • H04N19/124—Quantisation
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/134—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or criterion affecting or controlling the adaptive coding
    • H04N19/146—Data rate or code amount at the encoder output

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Compression Or Coding Systems Of Tv Signals (AREA)
  • Compression, Expansion, Code Conversion, And Decoders (AREA)

Abstract

A method and apparatus of digital image decoding is provided for reducing compression-related deterioration of an image to a minimum with a reduced storage capacity. The digital image decoding apparatus is equipped with a compression rate judging section for judging an optimal rate of compression for effecting the least deterioration to the image based upon the size of image in connection with the storage capacity of a frame memory. A compressing section compresses decoded data based upon the optimal rate of compression and sends the compressed data to a predictive/display frame memory for storage. An expanding A section expands the compressed data based upon the optimal rate of compression and sends the expanded data to a decoding section when the expanded data is required. An expanding B section reads out the compressed data of a display frame from the predictive/display frame memory and expands the compressed data based upon the optimal rate of compression and sends the expanded data to a display unit for display.

Description

Digital image code translator and digital image interpretation method
The present invention relates to a kind of digital image receiving system or employed digital image code translator and digital image interpretation methods such as digital CATV or digit broadcasting system.
Figure 54 and Figure 55 for example represent the block diagram of the LSI that put down in writing in the service manual by the STi3500 of the image processing LSI of SGS-Thomson Microeletronics company distribution and the memory sign of external memory storage.
In Figure 54, the 501st, microcomputer interface, the 502nd, FIFO (first-in first-out) memory, the 503rd, opening code detector, the 504th, memory I/O (I/O) unit, the 505th, variable length decoder, the 506th, decoding processor, the 507th, video-stream processor, the 508th, external memory storage, the 550th, the microcomputer interface line, the 551st, microcomputer bus, the 552nd, data wire, the 553rd, data wire, the 554th, external memory bus, the 555th, input-output line.
In Figure 55, the 601st, bit buffering district, the 602nd, OSD (screen display) district, the 603rd, predictive frame storage 1 district, the 604th, predictive frame storage 2 districts, the 605th, display frame memory block.
The action of this device is described below.The coded data of being stored in the bit buffering district 601 of external memory storage 508 is delivered to opening code detector 503, detect opening code here; Detect after this opening code, through FIFO the coded data of the opening code that continues is offered variable length decoder 505, variable length decoder 505 carries out variable length decoding; After this variable length decoding, carry out image decoding by decoding processor 506 and handle, and, the decoding image is written in the external memory storage 508 via memory I/O unit 504.
Place as storage decoding image, this external memory storage 508 has predictive frame storage 1 district 603, predictive frame storage 2 districts 604 and display frame memory block 605, other are frame predictions and show that being written to predictive frame with pictorial data stores in 1 district 603 or predictive frame storage 2 districts 604, only is written in the display frame memory block 605 as the pictorial data that shows usefulness.
Synchronously read being written to the horizontal vertical synchronizing signal of predictive frame storage 1 district 603, predictive frame storing data in 2 districts 604 and the display frame memory block 605 and television image etc., and output to video-stream processor 507 by external memory bus 554.
Demonstrations such as character data are distinguished in 602 with the OSD (screen display) that data are written in the external memory storage 508, with predictive frame storage 1 district 603, that predictive frame is stored 2 districts 604 is the same with display frame memory block 605, carry out access as required, and offer video-stream processor 507 through external memory bus 554.If the data in OSD district 602 are effective, video-stream processor 507 just covers the data of reading from predictive frame storage 1 district 603, predictive frame storage 2 districts 604 and display frame memory block 605 to the data in OSD district 602, and outputs to the outside.
Like this, just obtain based on the displayed image that is stored in the video data in the external memory storage 508.
According to above-mentioned digital image code translator, when decoding, must all store data necessary in the external memory storage 508, that is: under the situation of the coded data of carrying out interframe encode, must all store the data of employed other frames of this frame coding for the pictorial data of certain frame being deciphered just.
Therefore, existing problem is that the capacity of external memory storage 508 is big more for the just essential huge memory data output of encoding, and the scale of hardware is also just big more.
The present invention provides a kind of and can compress the capacity of memory as much as possible for a short time and realize the digital image code translator and the digital image interpretation method of hardware downsizing for addressing the above problem the technical scheme of proposition, its objective is.
Even the purpose of this invention is to provide a kind ofly under the situation of compressing the capacity of memory as far as possible for a short time, also can realize coming image degradation is suppressed to minimum digital image code translator and digital image interpretation method according to image size information.
Be characterised in that according to digital image code translator of the present invention the coded data that the picture coding of dimension of picture of regulation is crossed deciphered that this device has following key element:
(a) with the frame be unit the coded data of encoding is deciphered and obtain with the frame be the decoding data of unit decoder,
(b) with the frame be unit capacity storage pictorial data in accordance with regulations frame memory,
(c) decoding data crossed with above-mentioned decoder for decoding of compression and the decoding data that compressed as packed data be stored in compressor reducer in the above-mentioned frame memory,
(d) take out the decompression machine that is stored in interior packed data of above-mentioned frame memory and decompress(ion) output packed data.
Above-mentioned frame memory has the prediction frame memory of the above-mentioned decoder of storage for the packed data of the predictive frame of coded data being deciphered reference; The decoding data of said compressor compression predictive frame also is stored in the prediction frame memory; Above-mentioned decompression machine has decompress(ion) and is stored in the packed data of the predictive frame in the prediction frame memory and outputs to the decompression machine that the predictive frame of above-mentioned decoder is used; Above-mentioned frame memory has memory image and shows the display frame memory of the demonstration of usefulness with the packed data of frame; The said compressor compression shows with the decoding data of frame and is stored in and shows with in the frame memory; Above-mentioned decompression machine has decompress(ion) and is stored in the packed data of the display frame in the display frame memory and exports the decompression machine that the display frame of the data that decompress(ion) crosses is used.
The coded data that is input to above-mentioned decoder is the data that in accordance with regulations coded system was carried out interframe encode, above-mentioned digital image code translator is provided with input encoded data and judges the mode determinant of coded system, and said compressor is handled according to the coded system change compression of being judged by the aforesaid way determinant.
Above-mentioned decoder is exported the decoding data of M pixel * N pixel * r position as a data block; When the decoding data of each data block was transformed to packed data, said compressor was obtained the coefficient relevant with image quality, and the line translation of going forward side by side is handled, and big bit length is distributed to important coefficient, and little bit length is distributed to unessential coefficient; Said compressor does to become fixed length to the figure place S of the packed data that the result of the conversion process of a data block takes place.
Said compressor is characterised in that the optimal quantization table selection circuit of the quantized result quantization table that also selection is best that is provided with the different a plurality of quantizers of quantization table, each quantizer of comparison and selects from the selector of use by the output of the quantizer of the selected quantization table of optimal quantization table selection circuit.
Above-mentioned digital image code translator also is provided with the dimension of picture information of the dimension of picture of the above-mentioned image of input expression, compresses above-mentioned decoding data according to the capacity of above-mentioned dimension of picture and above-mentioned frame memory simultaneously and is judged to be the compression ratio determinant of the compression ratio that is stored in the above-mentioned frame memory.
The decoding data that result's compression that said compressor is judged according to above-mentioned compression ratio determinant is crossed by above-mentioned decoder for decoding, and a decoding data of compression is stored in the above-mentioned frame memory as packed data.
Above-mentioned decompression machine takes out the packed data that is stored in the above-mentioned frame memory, and according to the compression ratio decompress(ion) output packed data of judging by above-mentioned compression determinant.
Said compressor has the different a plurality of compact models of compression ratio; The size that above-mentioned compression determinant is selected packed data among a plurality of compact models is less than the capacity of above-mentioned frame memory and be maximum compact model.
Said compressor is compressed the decoding data of M pixel * N pixel * r position according to the quantification treatment of using one dimension differential pulse coding modulation (1D-DPCM) as a data block.
Above-mentioned digital image code translator has input encoded data, and judges the mode determinant of coded system.
Said compressor is handled according to the coded system change compression of being judged by the aforesaid way determinant; The aforesaid way determinant judge be by past and the following bi-directional predicted interframe coded system of predicting as coded system, still the single directional prediction inter-frame encoding of only predicting from the frame in past as coded system.
Under the situation of the coded data of single directional prediction inter-frame encoding, said compressor is not compressed decoding data; And under the situation of the coded data of bi-directional predicted interframe coded system, said compressor compression decoding data.
Said compressor is provided with a quantification treatment device that the decoding data of M pixel * N pixel quantizes as data block decoding data; Above-mentioned decompression machine has packed data is carried out re-quantization and decompress(ion) is the decompression processing device of the decoding data of M pixel * N pixel; Above-mentioned quantification treatment utensil has the different a plurality of quantizers of quantized character.
Said compressor be provided with the decoding data that detects above-mentioned M pixel * N pixel predetermined characteristic the characteristic searcher and from a plurality of quantizers of above-mentioned quantification treatment device, select a quantizer and in selected quantizer, the quantizer that the M pixel * decoding data of N pixel quantizes selected device according to the predetermined characteristic that above-mentioned characteristic searcher detects; Above-mentioned quantizer selects utensil to have the input M pixel * decoding data of N pixel to obtain and export the peaked maximum value detector of the difference value between pixel adjacent, the minimum detector of the minimum value of the difference value between pixel adjacent is obtained and exported to the input M pixel * decoding data of N pixel, maximum that characteristic quantization table that the maximum of being obtained by above-mentioned maximum value detector and the minimum value obtained by above-mentioned minimum detector are quantized and input are obtained by above-mentioned maximum value detector and the minimum value of being obtained by above-mentioned minimum detector also quantize above-mentioned maximum and minimum value with above-mentioned characteristic quantity table, are re-used as the characteristic quantizer that quantized value is exported.
Above-mentioned quantizer selector has according to above-mentioned quantized value to be selected the option table of a quantizer and selects the selector of optimum quantizer according to above-mentioned option table for compressing decoding data among a plurality of quantizers of above-mentioned quantification treatment device.
Above-mentioned decompression machine has and is provided with a plurality of inverse quantizers that have corresponding to the re-quantization characteristic of each quantized character of a plurality of quantizers that possess above-mentioned quantification treatment device; Above-mentioned digital image code translator also has the controller of the re-quantization characteristic of the quantized character of setting said compressor and above-mentioned decompression machine; Above-mentioned a plurality of quantizer is the variable quantizer of quantized character; Above-mentioned a plurality of inverse quantizer is the variable inverse quantizer of re-quantization characteristic.
When above-mentioned controller is provided with and sets quantized character respectively for above-mentioned a plurality of quantizers, for above-mentioned a plurality of inverse quantizers corresponding to the setting of above-mentioned quantized character set respectively the re-quantization characteristic the quantized character setting apparatus, according to the setting of the quantized character of above-mentioned quantized character setting apparatus above-mentioned option table is set to the option table setting apparatus in the quantizer selector and according to the setting of the quantized character of above-mentioned quantized character setting apparatus above-mentioned characteristic quantization table is set to characteristic quantization table setting apparatus in the characteristic quantizer.
According to digital image interpretation method of the present invention, comprise the steps: in the digital image interpretation method of digital image code translator with decoder and frame memory
The coded data that interframe/interior coding is crossed is decoded as the decoding data of the data block unit of M pixel * N pixel;
The decoding data of being deciphered the data block unit of the M pixel * N pixel that obtains by above-mentioned decoding step is carried out quantification treatment, and the compression of the row decoding data of going forward side by side is handled;
For the coded data that interframe/interior coding is crossed being deciphered, the prediction storing step more than the frame in the prediction frame memory that remains on frame memory by above-mentioned compression step compressed and supplied data with above-mentioned decoding step;
Remain on display frame storing step in the display frame memory of frame memory being used for the subsequently displaying transmitted image data of displayed image;
From above-mentioned prediction frame memory, read packed data, and carry out re-quantization and handle, carry out the decompress(ion) of packed data again, then data necessary is offered the 1st decompress(ion) step of above-mentioned decoding step; And
From above-mentioned display frame memory, read packed data, and carry out re-quantization and handle, carry out the decompress(ion) of packed data again, the 2nd decompress(ion) step of carrying out the output of visual video data then.
Above-mentioned digital image interpretation method also has judges dimension of picture and the compression ratio of judging the compression decoding data according to the capacity and the dimension of picture of frame memory, is re-used as the compression ratio determination step that compression ratio information used in the above-mentioned compression step is notified.
Above-mentioned digital image interpretation method has the controlled step of the re-quantization characteristic that the re-quantization of the quantized character of the quantification treatment of setting the above-mentioned compression step of change and the above-mentioned the 1st and the 2nd decompress(ion) step handles.The preferred embodiments of the present invention (to call embodiment in the following text) below are described with reference to the accompanying drawings.
Brief Description Of Drawings Fig. 1 is the formation block diagram according to the digital image code translator of one embodiment of the present of invention.Fig. 2 is the classification figure of frame.Fig. 3 is the bit flag figure of frame memory.Fig. 4 is the action diagram of frame memory.Fig. 5 is the flow chart of an example of compression processing.Fig. 6 is the flow chart of an example of compression processing.Fig. 7 is quantification figure.Fig. 8 is the bit flag figure of prediction display frame memory.Fig. 9 is compression processing mode Hull conversion figure.Figure 10 is necessary data area of decompress(ion) and the necessary data area figure of decoding.Figure 11 is the formation block diagram of decompression machine A.Figure 12 is the formation block diagram of decompression machine B.Figure 13 is the figure of the processing of expression decompression machine B.Figure 14 is the encode series familygram.Figure 15 is the encode series familygram.Figure 16 is that action flow chart is handled in compression.Figure 17 is the summary bit flag figure of prediction frame memory.Figure 18 is the formation block diagram of the digital image code translator of one embodiment of the present of invention.Figure 19 is the flow chart of digital image interpretation method of the present invention.Figure 20 is the figure of expression with the compress mode of the compressor reducer of digital image code translator of the present invention.Figure 21 is the figure of expression with the compress mode of the compressor reducer of digital image code translator of the present invention.Figure 22 is the figure of expression with the compress mode of the compressor reducer of digital image code translator of the present invention.Figure 23 is the figure of expression with the compress mode of the compressor reducer of digital image code translator of the present invention.Figure 24 is the compressor reducer pie graph of digital image code translator of the present invention.Figure 25 is the compress mode figure of expression with the compressor reducer of digital image code translator of the present invention.Figure 26 is the compress mode figure of expression with the compressor reducer of digital image code translator of the present invention.Figure 27 is the compress mode figure of expression with the compressor reducer of digital image code translator of the present invention.Figure 28 is the compress mode figure of expression with the compressor reducer of digital image code translator of the present invention.Figure 29 is the compressor reducer pie graph of digital image code translator of the present invention.Figure 30 is the other pie graph of the compressor reducer of digital image code translator of the present invention.Figure 31 is the other pie graph of the compressor reducer of digital image code translator of the present invention.Figure 32 is the formation block diagram of the digital image code translator of one embodiment of the present of invention.Figure 33 is the memory marking pattern of the bi-directional predicted frame memory in the digital image code translator.Figure 34 is the memory marking pattern of the single directional prediction frame memory in the digital image code translator.Figure 35 is the compressor reducer pie graph of digital image code translator of the present invention.Figure 36 is the other pie graph of digital image code translator of the present invention.Figure 37 is the other pie graph of digital image code translator of the present invention.Figure 38 is the other pie graph of digital image code translator of the present invention.Figure 39 is the interior block diagram of compressor reducer of the present invention.Figure 40 is the interior block diagram of quantification treatment device of the present invention.Figure 41 is the quantized character figure of quantizer of the present invention.Figure 42 is the quantized character figure of quantizer q2 of the present invention.Figure 43 is the quantized character figure of quantizer q15 of the present invention.Figure 44 is the packed data of a pixel of the present invention.Figure 45 is the interior block diagram of characteristic searcher of the present invention and quantizer selector.Figure 46 is a characteristic quantization table of the present invention.Figure 47 is an option table of the present invention.Figure 48 is the interior block diagram of decompression machine B of the present invention.Figure 49 is the other pie graph of digital image code translator of the present invention.Figure 50 is the interior block diagram of controller of the present invention and compressor reducer.Figure 51 is the interior block diagram of quantification treatment device of the present invention.Figure 52 is the interior block diagram of characteristic searcher of the present invention and quantizer selector.Figure 53 is the interior block diagram of decompression machine B of the present invention.Figure 54 is the block diagram of original example.Figure 55 is the bit flag figure of the frame memory of original example.
Embodiment 1 Fig. 1 is the general block diagram according to the device of embodiments of the invention.
Among the figure, the 101st, image coded data is carried out decoders for decoding; The 102nd, the compressor reducer that decoding data is compressed; The 103rd, the prediction display frame memory that constitutes by prediction frame memory and display frame memory; The 104th, read packed data and carry out the decompress(ion) of packed data from prediction display frame memory, again the data through decompress(ion) are outputed to the decompression machine A (predictive frame decompression machine) of decoder; The 105th, read packed data and carry out the decompress(ion) of packed data from prediction display frame memory, again the data through decompress(ion) are outputed to the decompression machine B (show and use decompression machine) of display unit (not shown).
The 150th, coded data, the 151st, decoding data, the 152nd, packed data, the 153rd, packed data, the 154th, video data, the 155th, decompressed data.
The following describes the action of Fig. 1 device.
Decoder 101 usefulness decompressed datas 155 are deciphered processing as prediction data to the coded data that is provided, and more resulting decoding data 151 is handled in decoding and carried out that non-reversible compression is handled or reversible compression is handled, thus the reduction amount of information.So-called non-reversible compression is handled the data and the complete inconsistent compression of the data after the compression that are meant before the compression and is handled; So-called reversible compression processing is meant data and the on all four compression processing of the data after the compression before the compression.For use the packed data 152 that compressed by compressor reducer 102 as future corresponding decoded frame prediction data show, the packed data 152 by compressor reducer 102 compressions is written to prediction display frame memory 103.And no frame packed data is written to the display frame memory block in the prediction, and the frame packed data of prediction and demonstration usefulness is written to the predictive frame memory block.As hereinafter described, not necessarily compressing total data also is fine.
Show that in order to carry out image the packed data that is written into carries out decompression processing in decompression machine B105, the description detailed as the back, by display unit just the order of the grating of usefulness data that decompress(ion) is crossed be presented on the display unit.
On the other hand, the predictive frame memory block of decompression machine A104 access prediction display frame memory 103, and the packed data that obtains like this carried out decompression processing, and in decoder 101, decipher when handling, its decompressed data 155 (prediction data) as necessity is provided out.
For stores compressed data, prediction display frame memory 103 can be by constituting than the little capacity of amount of information that has as the pictorial data of object.
Among Fig. 2, the predictive frame that 301a～301c uses when being other picture frame decodings; 302a～302d is the display frame of only using when image shows.In Fig. 3,310a is the predictive frame memory block of storage the 1st predictive frame; 310b is the predictive frame memory block of storage the 2nd predictive frame; The 311st, the display frame memory block of storage display frame.
In the code device (not shown), generate coded data 150 by following order, and be input in the decoder 101.
(1) predictive frame 301a
(2) predictive frame 301b
(3) display frame 302a (from predictive frame 301a and predictive frame 301b prediction)
(4) display frame 302b (from predictive frame 301a and predictive frame 301b prediction)
(5) predictive frame 301c
(6) display frame 302c (from predictive frame 301b and predictive frame 301c prediction)
(7) display frame 302d (from predictive frame 301b and predictive frame 301c prediction)
Decoding data 151 and packed data 152 also are imported into prediction display frame memory by the order of above-mentioned (1)～(7), and 103 storages show packed data 152 to prediction display frame memory by order shown in Figure 4.
Predictive frame 301a～301c is stored in predictive frame memory block 310a, among the 310b, is used for showing, and when being used for the decoding processing of other predictive frames, the decoding that is used for display frame 302a～302d is handled.On the other hand, display frame 302a～302d is stored in the display frame memory block 311 of prediction display frame memory 103, and only is used for showing.
The data of display frame only are used to show, therefore, in the compressor reducer 102 of Fig. 1, compress the data of display frame with non-reversible manner, and when decompress(ion), even the data behind the generation decompress(ion) are to the inconsistent mistake of data before compressing, because other frames are with reference to display frame 302a～302d, so the mistake that is taken place when not transmitting the display frame decompress(ion).As previously mentioned, even so-called non-reversible manner is the compress mode that compression back decompress(ion) also can not restore the data before compressing fully, that is: obliterated data takes place, and the compress mode of the inconsistent mistake of data after preceding data of compression and the compression takes place.
On the other hand, predictive frame memory block 310a, the data of the predictive frame that is write in the 310b are used to the decoding of other picture frame, therefore, under the situation that predictive frame 301a～301c is compressed with non-reversible compress mode, because the mistake that this compression produced just is sent in other picture frame.And in compressor reducer 102 under the situation of the reversible compress mode of employed compress mode right and wrong, because predictive frame 301a～301c is not compressed, and the decoding data former state is stored in predictive frame memory block 310a, among the 310b, so the mistake that is produced by compression can not be sent in other frames.
On the other hand, with compressor reducer 102 according to compression before with compression after the compression of the on all four reversible compress mode of data situation about handling under, because the data before the compression can be restored fully,, two sides of predictive frame 301a～301c and display frame 302a～302d cut down amount of information so being compressed processing.
Fig. 5 is the flow chart of an example of expression compression treatment step.Like this, be to be used to the predicted frame data predicting and show to judge it by the decoding picture frame of decoder 101 output, still only be used to the display frame data that show.Predicted frame data is not compressed, and be written in predictive frame memory block 310a, the 310b of prediction display frame memory 103.On the other hand, the display frame data that only are used to show are compressed processing, be written to then in predictive frame memory block 310a, the 310b of prediction display frame memory 103.Because compression is handled other frames is not exerted an influence, so under the situation that the compression of the non-reversible compress mode of compressor reducer 102 employings is handled, this mode is best.
As shown in Figure 6, in compression is handled, adopt under the situation of reversible compress mode, predictive frame and display frame two sides are all compressed processing, and the predicted frame data that is used to predict and show is written in predictive frame memory block 310a, the 310b, and the display frame data that only are used for showing are written to display frame memory block 311.
Also can only compress processing according to the kind of picture frame to predictive frame.
Represented among Fig. 7 is the summary of compression treatment step.In the decoder 101 of Fig. 1, the data of the data block 201 of the M pixel * N capable (pixel) that encoded are accepted the conversion process of regulation, at this, owing to each pixel is represented with the r position, so the amount of information of a data block is M * N * r position.Live through in the capable data of the M pixel * N of conversion process of discrete cosine transform, upper leftly be low territory signaling zone 292, middle body is a middle territory signaling zone 293, and the bottom right is a high territory signaling zone 294.
Fig. 8 is the storage mark of 1 frame data of the compression in the prediction display frame memory 103.
Among this figure, the 210th, the memory location of the frame information that compressed, the 211st, the storage position information of the t data block in 1 frame that compressed.
In the processing of compressor reducer 102, according to the characteristic of image the data block 201 of M * N pixel unit is carried out conversion process, and the data block after the conversion is divided into low territory signaling zone 292, middle territory signaling zone 293 and high territory signaling zone 294.The pixel count of low territory signal is that the unit picture element figure place of r1, low territory signal is s1 position/pixel; The pixel count of middle territory signal is that the unit picture element figure place of r2, middle territory signal is s2 position/pixel; The pixel count of high territory signal is that the unit picture element figure place of r3, high territory signal is s3 position/pixel (s1＞s2＞s3 wherein, r1+r2+r3=M * N).Like this, big figure place is distributed in low territory, is because low territory signal is big to the influence of image.Therefore, reduce the compression reduction that the influence of image is realized data volume.
If come a minute ligancy like this, the information generation S=r1 * s1+r2 * s2+r3 * s3 of data block is just always certain.
Therefore, can obtain the address of data block unit regularly, and the compressed image frame of from data block readout memory arbitrarily, being stored.For example: as shown in Figure 8, the beginning address number of condensed frame is under A number the situation, and the address of the t data block of condensed frame is exactly (A+ (t-1) * S) address number～(address number of A+t * S-1).Therefore, under the situation of the access of handling for the decoding of carrying out the t data block, because an arbitrarily individual condensed frame is all known its memory location, so just can carry out the access of data block unit.
Fig. 9 is illustrated in the example of the situation of a Hull conversion (Haar conversion) of using non-inverible transform in the transition coding algorithm.
H among the figure represents the coefficient matrix of 8 pixels * 8 row to conversion.
The pixel of getting the preceding data block of 1 dimension Haar conversion is X, and the pixel of the data block after the conversion is B, so
B＝HX
When B is quantized, and the data block after compressing is taken as B ' time, the data block Y that carries out behind the resulting decompress(ion) of decompress(ion) is exactly
Y＝H -1B
According to such computing, can carry out Compress softwares.
Because in this processing, after the conversion, because of quantification makes the figure place reduction, so just become non-reversible compression.At this, though represented the example of Haar conversion,, with the mode beyond the Haar also can, do not limit the present invention.
Figure 10 represents the relation in the various zones in the frame image data.
Among this figure, the 220th, picture frame, the 221st, the capable decoding prediction data piece of K pixel * L of necessity during decoding, the 222nd, the decompress(ion) of decompress(ion) necessity is used the data block group.
Still with Fig. 1 this Figure 10 is described.
Decoder 101 utilizes from all deciphering the capable decoding prediction data piece of K pixel * L that the arbitrfary point of handling and being stored in the picture frame 220 in the prediction display frame memory 103 obtains as the prediction pictorial data.On the other hand, be that unit compression is handled the data in the prediction display frame memory 103 and stored with the data block.Therefore, under the situation that the necessary capable decoding prediction data piece 221 of K pixel * L strides across between the data block in decoding is handled, only data block of decompress(ion) just can not get data necessary.
So, take out a plurality of decompress(ion)s that comprise decoding prediction data piece 221 with decompression machine A104 from prediction display frame memory 103 and use data block group 222, and each data block unit carried out the data extract of the decoding prediction data piece 221 of decoder 101 necessity being come out to deliver to decoder 101 after the decompression processing.When prediction display frame memory 103 extracted packed data, the prediction display frame memory address of packed data was undertaken by above-mentioned addressing at decompression machine A104.
Like this, just can access by data block unit and compress and the decoding prediction data blocks of data of the arbitrary region of the data of storing, and, in the data block store device (not shown) that the decompress(ion) that carries out after the decompression processing is used with data block group 222 the decompressed data of storage in decompression machine A104, just can carry out next data block to decoder 101 and handle necessary prediction and upgrade its necessary part again with pictorial data.Particularly, the position of the decoding prediction data piece of necessity is to predict according to the motion vector of each interframe during decoding, so, the probability height that utilizes again between adjacent data block.When the data block store behind the decompress(ion) of stated number in decompression machine A104, and when other data blocks are essential data block in next data block, be that unit upgrades the data of being stored, so that improve the efficient of decompression processing with the data block.
And, the memory of the pictorial data of a plurality of data blocks of crossing with the same storage side by side of picture frame decompress(ion) is set, and from according to the rules for example every horizontal line sense data of order here, preferably only wherein necessary partly extracted with gate circuit, at this moment, memory can be done to become and the same formation of data chunk line memory described later.
The data that also can only read necessary scope from the memory of the pictorial data of having stored a plurality of data blocks offer decoder 101.That is: only read the capable data of above-mentioned K pixel * L in order, and it is offered decoder 101.
Figure 11 handles sequential chart.Among this figure, the data block decoding time of being spent is handled in 280 decodings that are illustrated in a data block in the decoder 101, the compression processing time of being spent is handled in 281 compressions that are illustrated in a data block in the compressor reducer 102, and the decompression processing time of being spent is handled in 282 decodings that are illustrated in decoder 101 data necessary (K pixel * L is capable) among the decompression machine A104.
In decoder 101, in data block decoding time 280 to deciphering by the data of data block unit encoding, at this moment, must be from any original position of prediction display frame memory 103 K pixel * L line data as prediction data.At this, decompression machine A104 takes out data necessary according to the requirement from decoder 101 from prediction display frame memory 103, carry out decompress(ion) and supply with decoder 101.The decompression processing time 282 is decoders 101 requiring to output to the time that can supply with data decoder 101 behind the decompression machine A104.The decoding data of being handled by decoder 101 151 is sent to compressor reducer 102, in the time of the decoding data 151 that transmits next data block from decoder 101, finishes compression processing, and be written in the prediction display frame memory 103 data that are transmitted.
Do the decoding of the motion video of can in time encoding like this and handle, and compression decoding image, be then written in the frame memory, also can move without a doubt even cut down amount of information.
Figure 12 represents the formation of decompression machine B105.
Among this figure, the 270th, decompression processing device, the 271st, data chunk line memory.
Be imported into the decompression machine B105 from the data of the data block unit that prediction display frame memory 103 is read, at first, data by 270 pairs of data block units that imported of decompression processing device are carried out decompression processing, and the data after the decompression processing are pressed the sequential storage of each data block on the position of the regulation of data chunk line memory 271.Here, data chunk line memory 271 has the capacity of the data block (data chunk line) of the horizontal direction that can store whole picture frames 220.For example: the horizontal direction of picture frame 220 is made of the T pixel, and data block unit is made of J data block, and data chunk line memory 271 just has the capacity of J data block.
This data block is read unlike shown in Figure 13 and is undertaken by data block unit, but is undertaken by each pixel along the scan-line direction (striding data block direction from left to right) that constitutes image.That is: read the data of whole pixels of 1 horizontal scanning line in order.When reading of 1 horizontal scanning line finished, read each pixel data of next horizontal scanning line, repeat.
According to such formation, once carry out the storage of data chunk line and just can read the data of compressing by data block along grating orientation, and the output displayed image.For example:, just can access the signal that shows usefulness with the data that the horizontal-drive signal of 1 horizontal scanning line of regulation display frame is synchronously read 1 horizontal scanning line.
Figure 14, the kind of Figure 15 presentation code sequence, Figure 16 is the flow chart of expression compressor reducer action, Figure 17 is the summary bit flag that keeps the prediction frame memory of packed data.
As Figure 14, shown in Figure 15, in the coded data row, two kinds of bi-directional predicted and single directional prediction are arranged, that is: before the handlebar frame and back frame two sides' data as prediction data, the bi-directional predictive coding data rows that image is deciphered and only preceding frame data are used as 2 kinds of the single directional prediction coded data row that prediction data deciphers image.
As shown in figure 16, judge the kind of coded data row,, the data through decoding are written to the predictive frame memory block 310a that predicts display frame memory 103, among the 310b in compressor reducer 102 if single directional prediction coded data row do not compress; On the other hand, if the coded data row are bi-directional predictive coding data rows, just compress, and 2 frame data that compressed are written to respectively among predictive frame memory block 310a and the predictive frame memory block 310b.
Like this, as shown in figure 17, store data.That is: under the bi-directional predictive coding data conditions, 2 frame packed datas of prediction usefulness are stored in respectively among the predictive frame memory block 310a and predictive frame memory block 310b of prediction display frame memory 103.Therefore, decoder 101 just utilizes them to decipher processing.And under the situation of single directional prediction coded data, utilizing predictive frame memory block 310a, 1 frame data of being stored among the 310b are deciphered.
Like this, can not decipher, so can not produce the deterioration of the image that causes because of compression because under the situation of single directional prediction coded data, just do not compress.On the other hand, owing under the bi-directional predictive coding data conditions, utilize 2 predictive frames that therebetween frame is carried out predictive coding, so can carry out more high efficiency coding.In compressor reducer 102 compressed storage in prediction display frame memory 103, so just can suppress the increase of memory span.
Embodiment 2
This because not taking dimension of picture, aforesaid digital image code translator does not compress processing, so because to compressing processing less than the pictorial data of storable dimension of picture in the frame memory, just might produce unnecessary image degradation.For example: fixing in compression ratio, be compressed under the only about half of situation, is under about 1.1 times situation of frame storage content as dimension of picture, carries out only about half of compression, and this will involve and surpass the image degradation that requires.In this embodiment, the digital image code translator that illustrate is according to the dimension of picture information that is comprised in the coded data, changes compression ratio according to the capacity and the dimension of picture of frame memory, thereby can be the degradation inhibiting of image to minimum.
" form the bit width of 1 frame pixel count * per 1 pixel of image " with formula and obtain dimension of picture.That is: dimension of picture is obtained dimension of picture with formula " the bit width r of 1 row pixel count T * 1 frame line number U * per 1 pixel ".
Figure 18 is the general block diagram of an embodiment of digital image code translator.
Among this figure, the 101st, to the image coded data decoders for decoding, the 103rd, the prediction display frame memory that constitutes by prediction frame memory and display frame memory, 107a is the compressor reducer that carries out the decoding data compression, the 108th, carry out the decompression machine A of the decompress(ion) of the packed data read from frame memory, the 109th, by the decompression machine B of raster order dateout, the 106th, compression determinant, the 150th, coded data, the 151st, decoding data, the 152nd, packed data, the 153rd, packed data, the 154th, video data, the 155th, decompressed data, the 156th, dimension of picture information, the 157th, compression ratio information.At this, the content of prediction display frame memory 103 is identical with Fig. 3.
To be unit with the frame decipher the coded data 150 of the pictorial information coding that comprises dimension of picture for decoder 10, thereby to obtain with the frame be the decoding data 151 of unit.
Prediction display frame memory 103 capacity in accordance with regulations is unit memory image data with the frame, below simply prediction display frame memory is called frame memory.
In the time of the dimension of picture information of the above-mentioned dimension of picture of compression ratio determinant 106 input expression, compress above-mentioned decoding data 151 according to the capacity of above-mentioned dimension of picture and above-mentioned frame memory 103, and judge the compression ratio of storage usefulness in above-mentioned frame memory 103.Dimension of picture data 156 can be dimension of pictures itself, also can be the identification codes of recognition image size.Dimension of picture information 156 extracts from the pictorial information that comprises dimension of picture, also can be input to compression ratio determinant 106 to dimension of picture information 156 from outside specially.
Compressor reducer 107a compresses the decoding data of deciphering 151 according to the compression ratio of being judged by above-mentioned compression ratio determinant 106 with decoder 101, and the packed data that compressed is stored in the above-mentioned frame memory 103 as packed data 152.
Decompression machine A108 and decompression machine B109 are referred to as decompression machine altogether, and decompression machine extracts the packed data 152 that is stored in the above-mentioned frame memory 103, according to the compression ratio depressurizing compression data of being judged by above-mentioned compression ratio determinant 106.
Above-mentioned frame memory 103 has prediction frame memory, being used for storing 101 pairs of coded datas of above-mentioned decoder deciphers and the predictive frame decoding data of reference, the decoding data of said compressor 107a compression predictive frame also is stored in the prediction frame memory, and above-mentioned decompression machine has the decoding data of the predictive frame of being stored in the decompress(ion) prediction frame memory and outputs to the decompression machine A108 of above-mentioned decoder 101.
Above-mentioned frame memory 103 has memory image and shows the display frame memory of the demonstration of usefulness with the decoding data of frame, said compressor 107a compression shows with the decoding data of frame and is stored in the display frame memory, and above-mentioned decompression machine has the demonstration of being stored in the decompress(ion) display frame memory with the decoding data of frame and export the decompression machine B109 of decompressed data.
Flow chart with Figure 19 illustrates action below.
Decoder 101 is by coded data 150 and carry out the decoding processing (S1) of image as the decompressed data 155 of prediction usefulness; On the other hand, compression ratio determinant 106 is according to the dimension of picture information 156 that comprises in the coded data 150, and the capacity of consideration prediction display frame memory 103 is judged optimal compression value (S3).Compression ratio in this example is defined as:
Compression ratio=(amount of information before the compression)/(amount of information after the compression)
Compression ratio is from 1 1～1 n(n: natural number, 1 m〉=1, select in the n class of 1≤m≤n).For example: dimension of picture is that T pixel * U is capable, and the bit width of establishing per 1 pixel is the r position, and predictive frame memory block 310a, the predictive frame memory block 310b of prediction display frame memory and the capacity of display frame memory block 311 are respectively the Z position, so just with T * U * r/1 m1 of≤Z mSelect 1 of minimum m
Compressor reducer 107a compresses processing according to the 157 pairs of decoding datas 151 from decoder 101 of compression ratio information from compression ratio determinant 106, thereby cuts down amount of information (S4).Compression ratio information 157 can be compression ratio itself, also can be the compression ratio typical value of each scope of scope division by size, can also be the identifying information of each scope of expression.This compression ratio or typical value or identifying information become the information of the pattern corresponding to compressor reducer 107a described later.For the prediction data of the frame of the following decoding of opposing by packed data 152 usefulness of compressor reducer 107a compression and as video data, and carry out to predict the writing of display frame memory 103 (S6, S8).The packed data that is write carries out decompress(ion) (S9) by being used for carrying out the visual decompression machine B109 that shows; And read and show (S10) according to raster order.According to the decompression processing of carrying out decompression machine B109 from the compression ratio information 157 of compression ratio determinant 106.
For data necessary is supplied with decoder 101, decompression machine A108 carries out the access of necessary data from prediction display frame memory 103, and after the decompression processing of carrying out packed data, necessary data is supplied with decoder 101 (S7).About the decompression processing of decompression machine A108, also the same with decompression machine B109, carry out according to compression ratio information 157 from compression ratio determinant 106.
According to such formation, because compression and storage data, prediction display frame memory 103 just can be by constituting less than the memory capacity as the amount of information that pictorial data had of object, and, because the dimension of picture according to coded data changes compression ratio, thereby selected optimal compression value just can not cause above the image degradation that requires so carry out the compression of pictorial data.
What Figure 20～Figure 23 represented is the action of compressor reducer.
Figure 20 represents the 1st compression processing mode in the compressor reducer, and Figure 21 represents the 2nd compression processing mode in the compressor reducer, and Figure 22 represents the 3rd compression processing mode in the compressor reducer, and Figure 23 represents the 4th compression processing mode in the compressor reducer.
Figure 24 represents an example of the formation of compressor reducer 107a, the 120th, and subtracter, the 121st, quantizer, the 122nd, inverse quantizer, 123a, 123b are selectors, 124 is 1 pixel delays, the 125th, selector is selected signal generator, and the 159th, select signal.
The following describes Figure 20～Figure 23.
In Figure 20～Figure 23, represented is an example of being compressed by compressor reducer 107a.In the example of this figure, 8 * 8 pixels of block size M=8, N=8 (r=8 position/pixel), compress mode has 4 kinds of compact models with 1 dimension DPCM (1Dimention-Differential Pulse Code Modulation:1D-DPCM).1 dimension DPCM is the quantification of the differential data of adjacent pixels being carried out 4 (p=4).That is: have following 4 patterns:
(1) carries out the pattern 1 (compression ratio=1.78) of 1D-DPCM by 8 pixel units (L=8)
(2) carry out the pattern 2 (compression ratio=1.6) of 1D-DPCM by 4 pixel units (L=4)
(3) carry out the mode 3 (compression ratio=1.3) of 1D-DPCM by 2 pixel units (L=2)
(4) pattern of not compressing 4 (L=1, compression ratio=1)
In pattern 1, owing to be the 1D-DPCM of 8 pixel units, the 1st pixel of the amount of information after the compression is 8, the 2nd～7 pixel is 4, and repeats N=8 time, and its amount of information just becomes (8+4 * 7) * 8, and the amount of information before the compression is 8 * 8 * 8, and therefore, compression ratio is:
Compression ratio=(amount of information before the compression)/(amount of information after the compression)
＝(8×8×8)/((8+4×7)×8)＝1.78
In pattern 2, compression ratio=(8 * 8 * 8)/((8+4 * 3) * 16)=1.6;
In mode 3, compression ratio=(8 * 8 * 8)/((8+4 * 1) * 32) ≈ 1.3;
In pattern 4, compression ratio=(8 * 8 * 8)/(8 * 8 * 8)=1.
Figure 24 is described
Compression ratio information 157 is the information corresponding to the pattern of compressor reducer 107a, pointing-type according to compression ratio information 157, initial pixel to DPCM, 8 decoding datas, 151 former states are input to 1 pixel delays 124, to decoding data 151 in addition, with the difference of subtracter 120 calculating, further, carry out 4 quantifications with 121 pairs of differential datas of quantizer with the output of 1 pixel delays 124.
Packed data 152 through quantizing outputs to the outside of compressor reducer 107a, simultaneously, carries out partial decode with inverse quantizer 122, and is input to 1 pixel delays 124.
This is the compressed action according to 1D-DPCM, under the situation of pattern 1, compressor reducer 107a selects to use 8 decoding datas 151 of per 8 pixel selection of selector 123a the signal generator 125 from selector according to compression ratio information 157,7 remaining pixels is selected 4 quantized datas of quantizer 121 with selector 123a.
Equally, at pattern 2, mode 3 and pattern 4 difference per 4 pixels, 2 pixels and 8 decoding datas 151 of 1 pixel selection.
In aforesaid example, the 1st pixel is according to the output of r=8 position former state, even the 1st pixel also cannot be quantified as t position (t≤r), and in precedent, 8 * 8 pixels as 1 data block, still, also can M * N pixel (M=N or M ≠ N) as 1 data block as 1 data block.In precedent, along continuous straight runs carries out 1D-DPCM, also can be vertically (L≤N) is that unit carries out 1D-DPCM with the L pixel.
As mentioned above, the compressor reducer of this embodiment is characterised in that the L of unit (L≤M or the L≤N that compresses with the compression ratio decision, L is the common divisor of M or N) value, with the L pixel be unit (L≤M or L≤when N) amount of information of per 1 pixel being taken as the r position, 1 amount of pixels of the benchmark that becomes the L pixel is turned to the t position, to in addition pixel by the difference of pixel adjacent carry out the p position (quantification treatment of p≤r), thus cut down the generation amount of information of the data block of M * N pixel according to compression ratio.
Embodiment 3
What Figure 25～Figure 28 represented is the action of compressor reducer.
Figure 25 represents the 1st compression processing mode in the compressor reducer, and Figure 26 represents the 2nd compression processing mode in the compressor reducer, and Figure 27 represents the 3rd compression processing mode in the compressor reducer, and Figure 28 represents the 4th compression processing mode in the compressor reducer.
Figure 29 represents the formation of compressor reducer 107b, 123c, and 123d, 127a, 127b are selectors, the 129th, selector is selected signal generator, and the 160th, select signal.
The following describes Figure 25～Figure 28.
In Figure 25～Figure 28, represented is an example of being compressed by compressor reducer 107b.In the example of this figure, 8 * 8 pixels of block size M=8, N=8 (r=8 position/pixel), compress mode has 4 kinds of compact models with 1 dimension DPCM.1 dimension DPCM is the differential data displacement adaptively to adjacent pixels, quantizes.That is: have following 4 patterns:
(1) carries out the pattern 1 that 1D-DPCM, 4 (p=4) quantize by 8 pixel units (L=8)
(2) carry out the pattern 2 that 1D-DPCM, 5 (p=5) quantize by 8 pixel units (L=8)
(3) carry out the mode 3 that 1D-DPCM, 6 (p=6) quantize by 8 pixel units (L=8)
(4) carry out the pattern 4 that 1D-DPCM, 7 (p=7) quantize by 8 pixel units (L=8)
That is: in pattern 1, owing to be the 1D-DPCM of 4 quantizers, so the amount of information after the compression be (8+4 * 7) * 8, and the preceding amount of information of compression is 8 * 8 * 8, so compression ratio is:
Compression ratio=(amount of information before the compression)/(amount of information after the compression)
＝(8×8×8)/((8+4×7)×8)＝1.78
Equally, in pattern 2, compression ratio=(8 * 8 * 8)/((8+5 * 7) * 8) ≈ 1.49;
In mode 3, compression ratio=(8 * 8 * 8)/((8+6 * 7) * 8)=1.28;
In pattern 4, compression ratio=(8 * 8 * 8)/((8+7 * 7) * 8) ≈ 1.12.
Figure 29 is described
Initial pixel to DPCM, 8 decoding datas, 151 former states are input to 1 pixel delays 124, to decoding data 151 in addition, with the difference of subtracter 120 calculating with the output of 1 pixel delays 124, further, with quantizer 121a～121d differential data is carried out 4 quantifications.
When quantizing, under the situation of pattern 1, under the situation of pattern 2, under the situation of mode 3 and under the situation of pattern 4, according to the pattern of compression ratio information 157, select respectively by the data of 4 quantizer 121a quantifications, by the data of 5 quantizer 121b quantifications, by the data of 6 quantizer 121c quantifications and the data that quantize by 7 quantizer 121d with selector 127a.
The data that quantized are outputed to the outside of selector 127b as packed data 152 through selector 123, simultaneously, carry out partial decode, select by selector 127, be input to 1 pixel delays 124 then according to pattern with inverse quantizer 122a～122d.
When carrying out re-quantization, the same with quantification treatment, under the situation of pattern 1, under the situation of pattern 2, under the situation of mode 3 and under the situation of pattern 4, select respectively by the data of 4 inverse quantizer 122a re-quantizations, by the data of 5 inverse quantizer 122b re-quantizations, by the data of 6 inverse quantizer 122c re-quantizations with by the data of 7 inverse quantizer 122d re-quantizations.
As mentioned above, no matter the compressor reducer of this embodiment is characterised in that compression ratio, and the L of unit that fixedly compresses (L≤M or L≤N, L is the common divisor of M or N) value, 1 amount of pixels of the benchmark that becomes the L pixel is turned to the t position, in addition pixel is carried out the quantification treatment of p position by the difference of pixel adjacent, and according to the value change t of compression ratio and the value of p, thereby cut down the amount of information of the data block of M * N pixel according to compression ratio.
Embodiment 4
Figure 30 represents the formation of compressor reducer 107ac, and the label identical with Figure 24 and Figure 29 represented identical or equal content.
Below its action of explanation.
Among the figure, initial pixel to DPCM, 8 decoding datas, 151 former states are input to 1 pixel delays 124, to decoding data 151 in addition, with the difference of subtracter 120 calculating with the output of 1 pixel delays 124, further, with quantizer 121a～121d differential data is carried out 4 quantifications.
When quantizing, selector 127a selects according to the model selection of compression ratio information 157 that 4 quantizer 121a, 5 quantizer 121b, 6 quantizer 121c and 7 quantizer 121d's is some.The packed data 152 that quantized is outputed to the outside of selector 127c, simultaneously, carry out partial decode, be input to 1 pixel delays 124 then with inverse quantizer 122a～122d.When carrying out re-quantization, the same with quantification treatment, selector 127b selection 4 inverse quantizer 122a, 5 inverse quantizer 122b, 6 inverse quantizer 122c and 7 inverse quantizer 122d's is some.
Select signal generator 125 to select 8 pixel units, 4 pixel units, 2 pixel units and 1 pixel unit according to the pattern of compression ratio information 157 by selector again, and select to select 8 decoding data 151 or some by the selected quantized data of selector 127a by selector 123a.Carry out exquisiteness quantizes like this.
As mentioned above, the compressor reducer of this embodiment is characterised in that with compression ratio and changes the L of unit that compresses, 1 amount of pixels of the benchmark that becomes the L pixel is turned to the t position, pixel is in addition carried out the quantification treatment of p position by the difference of pixel adjacent, and change the value of t and p, thereby cut down the amount of information of the data block of M * N pixel according to compression ratio according to compression ratio.
Embodiment 5
Figure 31 represent compressor reducer 102a quantizer quantization table an example and select the best table of its best table to select an example of circuit.
Compressor reducer 102a shown in Figure 31 is used for replacing the compressor reducer 102 shown in Fig. 1, under the situation of using compressor reducer 102a shown in Figure 31, does not need compression ratio determinant 106.
230a～230n is the different n of a quantization table quantizer, 231a～231n is a delay circuit, 232a～232n is a difference channel, 233a～233n is an absolute value circuit, 234a～234n is a summation circuit, the 235th, each quantizer 230a～230n of comparison also selects the best table of optimal quantization table to select circuit, and the 128th, the selector of the output of the quantizer of the quantization table of the next personal best table selection circuit of selection 235 selections.
Below explanation action.
Quantized from the decoding data 151 of decoder 101 outputs by each quantizer 230a～230n respectively, in order to distinguish n quantization table, getting under the situation of distributing the e position, quantization table is prepared to be less than 2 e(n≤2 e) individual table.
With each quantizer 230a～230n the decoding data 151 before quantizing is quantized, decoding data 250a～250n after in difference channel 232a～232n, getting quantification with quantize before the difference of decoding data 151, again in absolute value circuit 233a～233n and summation circuit 234a～234n by DPCM unit (L) get difference absolute value with.
Best table select circuit 235 usefulness selectors 128 from quantized data 250a～250n, select to fetch data block unit absolute value and the quantized data of minimum, like this, owing to prepare a plurality of quantization tables, just the optimal quantization table can be selected, thereby deterioration quantized data still less can be in the L of unit of each DPCM, selected.
Embodiment 6
Figure 32 is the block diagram of an embodiment of digital image code translator.The 110th, input encoded data is also judged the mode determinant of coded system, the 111st, and the compressor reducer of handling according to the coded system change compression of judging by aforesaid way determinant 110.Aforesaid way determinant 110 is judged the bi-directional predicted interframe coded system predicted by the past and following frame and some positions coded system of the single directional prediction inter-frame encoding only predicted from the frame in past, said compressor 111 is under the situation of the coded data of bi-directional predicted interframe coded system, than the compression ratio that improves decoding data under the situation of the coded data of single directional prediction inter-frame encoding.In Figure 32, the 158th, the mode information of the aforementioned coded system of expression, the label identical with Figure 18 represented identical or suitable content.
Figure 33, the storage mark of the frame memory of the two-way and single directional prediction of an embodiment of 34 expression digital image code translators, the 126th, the selector different with Figure 24 selected signal generator.
Below explanation action
Decoder 101 carries out image decoding processing from coded data 150 with as the decompressed data of predicting 155, on the other hand, compression ratio determinant 106 is judged optimal compression value according to the dimension of picture information 156 that comprises in the coded data 150 from the capacity of prediction display frame memory 103.Compression ratio is from 1 1～1 n(n: natural number, 1 m〉=1, select in the n class of 1≤m≤n).For example: dimension of picture is that T pixel * U is capable, and the bit width of establishing per 1 pixel is the r position, and a frame capacity of prediction display frame memory is the Z position, so just with T * U * r/1 m1 of≤Z mSelect 1 of minimum m
In mode determinant 110, judge that coded data 150 is only from the single directional prediction inter-frame encoding of past prediction or use from the coded data of the bi-directional predicted interframe coded system of past and following two direction predictions, is sent to compressor reducer 111 to this mode information 158 then.
Compressor reducer 111 is according to compressing processing from the compression ratio information 157 of compression ratio determinant 106 with from 158 pairs of decoding datas 151 from decoder 101 of mode information of mode determinant 110, thereby carries out the reduction of amount of information.
For example, under the compress mode situation identical, in selector selection signal generator 126, set the unit of 1D-DPCM by compression ratio information 157 and mode information 158 with the compress mode shown in the embodiment 2.
In Figure 14 and Figure 15, represented unidirectional and bi-directional predicted summary situation, in order to predict from the past and following picture frame, 2 picture frames of bi-directional predicted necessary storage, and under the situation of single directional prediction, in order only to predict, can 1 picture frame be stored from the picture frame in past.
Therefore, as Figure 33 and shown in Figure 34, because the situation of single directional prediction is compared with bi-directional predicted situation, with predictive frame memory block 310a, 310b compares among the 310c of predictive frame memory block, mode information 158 can be used 2 times memory block, so even same dimension of picture is compared with bi-directional predicted compression ratio X, under the situation of single directional prediction, compression ratio is taken as X/2, also compresses processing.
Compression ratio when bi-directional predicted is less than ' 2 ' time, under the situation of single directional prediction, just can not compress processing, for the prediction data of following decoded frame that packed data 152 usefulness of compressor reducer 111 compressions are opposed, so it is written in the prediction display frame memory 103.
Like this, only carry out the coded data of single directional prediction with compressor reducer and compare, reduce compression ratio compression (or not compressing) with the bi-directional predicted interframe coded data of same dimension of picture.
Show in order to carry out image, the packed data that is write with decompression machine B109 decompress(ion), and is read demonstration by raster order, according to the decompression processing of carrying out decompression machine B109 from the compression ratio information 157 of compression ratio determinant 106.
For data necessary is supplied with decoder 101, in decompression machine A108, carry out the data necessary access, and carry out after the decompression processing of packed data necessary data being supplied with decoder 101 from prediction display frame memory 103.About the decompression processing of decompression machine A108, also the same with decompression machine B109, carry out according to compression ratio information 157 from compression ratio determinant 106.
According to such formation, because compress, so prediction display frame memory 103 can constitute with the capacity that is less than as the amount of information of the pictorial data of object, and, owing to select the optimal compression value of carrying out the compression ratio change according to the dimension of picture of coded data, can not cause the deterioration that surpasses the image that requires so carry out the compression of image.
Under situation shown in Figure 32, represented to have the situation of compression ratio determinant 106, but also can not have compression ratio determinant 106.Under the situation that does not have compression ratio determinant 106,111 of compressor reducers compress processing according to mode information 158.
Figure 36 represents a situation of Compress softwares predictive frame.
Figure 37 represents a situation of Compress softwares display frame.
Like this, also can a Compress softwares predictive frame and a certain side of display frame.
Embodiment 7
Figure 38 is the block diagram of an embodiment of digital image code translator.
In Figure 38, the constitutive characteristic part is that compressor reducer 112, decompression machine A113 and decompression machine B114 are different with embodiment 1.
Figure 39 is the interior block diagram of compressor reducer 112.
Quantification treatment device 703 has the different a plurality of quantizers of quantization characteristic, characteristic searcher 701 input decoding datas 151, and detect the maximum and the minimum value of difference value of adjacent pixel of the decoding data of M pixel * N pixel data blocks as the characteristic of regulation.The characteristic signals 751 of the characteristic of the regulation of the maximum that quantizer selector 702 is exported from characteristic searcher 701 according to expression, minimum value etc. selects optimum quantizer from quantification treatment device 703, and signal 752 is selected in output
Figure 40 is the interior block diagram of quantification treatment device 703.
Quantification treatment device 703 has 16 quantizers from quantizer q0 to quantizer q15, sets each quantizer and quantizes in scope shown in Figure 41.For example: quantizer q2 is set and quantizes 0～255 scope, and quantizer q15 is set the scope of quantification-255～+ 255.
Figure 42 represents the quantized character of quantizer q2.
Quantizer q2 is the quantizer that quantizes 0～255 value by 0～9 10 grades.
Figure 43 is the quantized character figure of quantizer q15.Quantizer q15 is the quantizer that quantizes-255～+ 255 value by 0～9 10 grades.
By Figure 42 and Figure 43 more as can be known, quantizer q2 and quantizer q15 relatively, it can quantize with 2 times precision.
Like this, as shown in figure 41, quantizer q0～q15 shown in Figure 40 just has the quantized character that has nothing in common with each other in advance, under compressor reducer 112 is quantized M pixel * N pixel (for example 8 pixels * 8 pixels) as the decoding data 151 of 1 data block situation, be in and select one in a plurality of quantizers in the quantification treatment device 703, quantize.
Figure 44 represents the example by the packed data form of the packed data 152 of quantification treatment device 703 outputs.
Figure 44 represents the packed data of corresponding 1 pixel, and the form of packed data shown in Figure 44 is a general form in 16 quantizers.The y bit representation of packed data quantizes the quantizer of usefulness in quantification treatment device 703.In this embodiment, owing to can discern 16 quantizers, so the y position is 4.As the quantized result of 1 pixel, the quantization index sign indicating number of output z position, as Figure 42 and shown in Figure 43, under the situation of carrying out 10 grades of quantifications, the z position be 4 just enough.Like this, the packed data output of the z position both of the y position of expression quantizer and expression quantization index sign indicating number as 1 pixel.
Carry out the selection of quantizer according to following method.
Figure 45 is the interior block diagram of characteristic searcher 701 and quantizer selector 702.
Maximum value detector 704 is imported the decoding data of M pixel * N pixels, and obtains the maximum of the difference value of adjacent pixel.Minimum detector 705 is imported the decoding data of M pixel * N pixels, and obtains the minimum value of the difference value of adjacent pixel.Characteristic quantizer 706 is imported maximum of being obtained by maximum value detector 704 and the minimum value of being obtained by minimum detector 705, and quantizes with 781 pairs of maximums of characteristic quantization table and minimum value.
One example of Figure 46 characterization quantization table 781.
The example of Figure 46 is that decoding data is got-255～+ 255 value (that is: 9), quantizes progression and is taken as 10.From the maximum of maximum value detector 704 outputs are n, from the minimum value of minimum detector 705 outputs are m, when satisfy A2≤n＜A3, (A2)≤m＜(A1) time, peaked quantification typical value is that AD8, maximum quantized value 770 are that the quantification typical value of S8, minimum value is that AD2, maximum quantized value 771 are S2.
Like this, characteristic quantizer 706 is exported maximum quantized value 770 and maximum quantized value 771 with reference to characteristic quantization table 781 as characteristic signals 751.
Quantizer selector 702 input characteristics signals 751 are also selected optimum quantizer by selector 783 with reference to option table 782.
Figure 47 represents an example of option table 782.
Option table 782 shown in Figure 47 is that the characteristic according to each quantizer shown in Figure 41 makes.At maximum quantized value 770 is that S8, minimum value quantized value 771 are under the situation of S2, selects quantizer q14 from Figure 47.As shown in figure 41, quantizer q14 be-A3～+ quantizer that quantizes in the data area of A3.As other example, be that S9, minimum value quantized value 771 are under the situation of S5 at maximum quantized value 770, use quantizer q2.Option table 782 is that expression has in 16 quantizers of predetermined regulation quantized character can optimal quantization maximum quantized value 770 and the table of the quantizer of the data of minimum value quantized value 771 expressions.From the selection signal 752 of selector output is the signal of expression with which quantizer, and as shown in figure 40, this signal is imported into quantification treatment device 703.703 of quantification treatment devices make the chosen quantizer action in it, do not make not chosen quantizer action.Like this, quantification treatment device 703 is exported packed data 152 with regard to input decoding data 151.
Figure 48 is the interior block diagram of decompression machine B114.
Decompression machine B114 has decompression processing device 270 and line data block storage 271, in decompression processing device 270, is provided with inverse quantizer r0～r15.Inverse quantizer r0～r15 is provided with corresponding to quantizer q0～q15, that is: inverse quantizer r0～r15 carries out re-quantization in the data area of each quantizer shown in Figure 41, for example: inverse quantizer r0 is corresponding to quantizer q0, input packed data 153b, and carry out the re-quantization of packed data, the decoding data in output 0～A3 scope.More particularly, the packed data that decompression processing device 270 input is shown in Figure 44 makes corresponding to the inverse quantizer action by the quantizer of y bit representation, and input is carried out re-quantization by the quantization index sign indicating number of z bit representation.Corresponding to being failure to actuate by the inverse quantizer beyond the inverse quantizer of the quantizer of y bit representation.The decoding data that carried out decompression processing with decompression processing device 270 is imported into line data block storage 271.Later action is because identical with embodiment 1, so, omitted its explanation at this.Though diagram is not come out, in decompression machine A113, still have and the decompression processing device shown in Figure 48 270 same processors that constitute, make the some inverse quantizer actions in a plurality of inverse quantizers, carry out the decoding of packed data.
As mentioned above, for compressed image data and store, can constitute prediction display frame memory 103 with the amount of information that pictorial data was had that is less than as object.
Calculate the characteristic of data for the unit that compresses with each, and carry out quantification treatment with the quantizer of characteristic the best, can optimal compression be written to the data of predicting display frame memory 103, and the amount of information that pictorial data had that can be less than as object constitutes, and, can do one's utmost to suppress the image degradation that causes owing to compression.
Because the miniaturization of prediction display frame memory 103, not only cut down the capacity of memory block device, and reduced memory and read and write the bit width of necessary address space and data, thereby equipment miniaturization is also reduced cost significantly.
Embodiment 8
Figure 49 is the example according to the digital image code translator of present embodiment.
The difference of Figure 49 and Figure 48 is to be provided with controller 700, and controller 700 is the devices that are used for controlling the quantized character of compressor reducer 112a.Controller 700 is also controlled the re-quantization characteristic of decompression machine A113a and decompression machine B114a.
Figure 50 is the interior block diagram of controller 700 and compressor reducer 112a.
Figure 51 is the interior block diagram of quantification treatment device 703a.
Figure 52 is the interior block diagram of characteristic searcher 701a and quantizer selector 702a.
As shown in figure 50, in controller 700, be provided with characteristic quantization table setting apparatus 784, option table setting apparatus 785 and quantized character setting apparatus 786.Shown in Figure 52, characteristic quantization table setting apparatus 784 uses control line 760 that characteristic quantization table 781a is set among the characteristic searcher 701a; 785 pairs of quantizer selectors of option table setting apparatus 702a sets option table 782a with control line 761; Each quantizer of 786 pairs of quantification treatment devices of quantized character setting apparatus 703a is set the scope of quantized data with control line 762.Quantizer q0～q15 can be according to the data area change quantized character of being formulated by control line 762.For example: shown in Figure 51, quantized character setting apparatus 786 usefulness control lines 762 the setting range of 0～A3 in quantizer q0, quantize; Scope for quantizer q1 setting-A～0 quantizes.
Figure 53 is the interior block diagram of decompression machine B114a.
The decompression processing device 270a of decompression machine B114a has a plurality of inverse quantizers, and with the same with the situation of Figure 51 explanation, each inverse quantizer like this, has just been prepared inverse quantizer corresponding and each quantizer through the scope of control line 762 input re-quantization data.
Though embodiment 7,8 is not done special explanation, can tie up DPCM with 1 dimension DPCM or 2 as compress mode.
From embodiment 2 to embodiment 6,1 dimension DPCM is described, also can be to use the situation of 2 dimension DPCM, or also can compress with other compress modes in addition.
From embodiment 1 to embodiment 7, the situation of the coded data of interframe encode has been made special instruction, even but under the situation of the coded data of intraframe coding, also can compress processing.
As mentioned above, according to the pictorial data of the present invention,, the frame memory of the enough low capacities of energy constitutes the digital image code translator thereby being small-scale hardware owing to the condensed frame memory.
According to the present invention, owing to select optimal compression value from the relation of dimension of picture and frame storage content, thus can be the degradation inhibiting of image to minimum degree.
According to the present invention, be the unit that changes 1D-DPCM according to compression ratio owing in compressor reducer, compress, and change compression ratio according to dimension of picture with 1D-DPCM, thus can be the degradation inhibiting of image to minimum degree.
According to the present invention, be the quantization digit that changes 1D-DPCM according to compression ratio owing in compressor reducer, compress, and change compression ratio according to dimension of picture with 1D-DPCM, thus can be the degradation inhibiting of image to minimum degree.
According to the present invention, owing to prepare a plurality of quantization tables, and select quantization error the best table still less, thus can be the degradation inhibiting of image to minimum degree.
According to the present invention owing to differentiate single directional prediction and bi-directional predicted coded data, even and same dimension of picture also change compression ratio, and maximally utilise frame memory, thus can be the degradation inhibiting of image to minimum degree.
As mentioned above, according to the present invention, because the pictorial data of condensed frame memory, thereby the frame memory of the enough low capacities of energy constitutes the digital image code translator.
According to the present invention, owing to use the different a plurality of quantizers of quantized character in the compressor reducer selectively, this just can carry out the optimum data compression, thereby can compress the few image compression of deterioration.
Owing to grasp the pictorial data characteristic of each compression unit by the characteristic searcher, thereby can carry out the high compression of data efficiency.
According to the present invention, owing to calculate the maximum and the minimum value of the differential data between adjacent pixels, further this max min is quantized, thereby enough simple circuit of energy or module are grasped the compression of pictorial data characteristic easily.
According to the present invention,, thereby can carry out correct decoding owing to a plurality of inverse quantizers that have corresponding to a plurality of quantizers.
According to the present invention, owing to change quantized character with controller, thus can change compression property lentamente in conjunction with the characteristic of decoding data, and the frame memory of the enough low capacities of energy constitutes the digital image code translator.

Claims (14)

1. digital image code translator that the coded data of the picture coding of regulation dimension of picture is deciphered has:
(a) coded data of picture coding is deciphered with frame dimension unit, and obtain the decoding data of frame unit decoder,
(b) capacity in accordance with regulations with the frame be unit memory image data frame memory,
(c) compression is with the decoding data of above-mentioned decoder for decoding, and the decoding data that compresses as packed data be stored in the above-mentioned frame memory compressor reducer and
(d) take out the packed data of being stored in the above-mentioned frame memory, and the decompression machine of packed data decompress(ion) output.
2. according to the digital image code translator of claim 1, it is characterized in that it is to be used for deciphering the data of encoding and the prediction frame memory of the packed data of the predictive frame of reference that described frame memory has the described decoder of storage; The decoding data of described compressor compresses predictive frame also stores in the prediction frame memory; Described decompression machine has the packed data of the predictive frame of being stored in the decompress(ion) prediction frame memory, and outputs to the predictive frame decompression machine of described decoder; Described frame memory has the display frame memory of packed data that memory image shows the display frame of usefulness; Described compressor compresses shows with the decoding data of frame and stores in the display frame memory; The demonstration of being stored in the described decompression machine decompress(ion) display frame memory packed data of frame, and the demonstration decompression machine of output decompressed data.
3. according to the digital image code translator of claim 1, the coded data that it is characterized in that being input to described decoder is the data of coded system interframe encode according to the rules, and described digital image code translator has input encoded data and judges the mode determinant of coded system; The coded system change compression that described compressor reducer is judged according to described mode determinant is handled.
4. according to the digital image code translator of claim 1, it is characterized in that described decoder is output as 1 data block to the decoding data of M pixel * N pixel * r position; Obtain coefficient when described compressor reducer is transformed to packed data to the decoding data of each data block, carry out conversion process, big bit length is distributed to important coefficient, little bit length is distributed to unessential coefficient about image quality; Described compressor reducer does to become fixed length to the figure place S of the resultant packed data of the conversion process of 1 data block.
5. according to the digital image code translator of claim 1, it is characterized in that described compressor reducer has the quantized result of the different a plurality of quantizers of quantization table, each quantizer of comparison, and select the best table of optimal quantization table to select circuit, select to use the selector of output of selecting the quantizer of the quantization table that circuit selects by the best table.
6. according to the digital image code translator of claim 1, when it is characterized in that also having the dimension of picture information of dimension of picture of the described image of input expression, compress described decoding data and judge according to the capacity of described dimension of picture and described frame memory and be used for the compression ratio determinant of the compression ratio of in described frame memory, storing; The compression ratio that described compressor reducer is judged according to described compression ratio determinant is compressed the decoding data with described decoder for decoding, and the decoding data of compression is stored in the described frame memory as packed data; Described decompression machine takes out the packed data that is stored in the described frame memory, and the compression ratio decompress(ion) of being judged according to described compression ratio determinant again and export packed data.
7. according to the digital image code translator of claim 6, it is characterized in that described compressor reducer has the different a plurality of compact models of compression ratio; Described compression ratio determinant is selected packed datas from a plurality of compact models size is less than the capacity of described frame memory and be the compact model of maximum.
8. according to the digital image code translator of claim 6, it is characterized in that described compressor reducer the decoding data of M pixel * N pixel as 1 data block, and according to compressing with the quantification treatment of 1 dimension differential pulse coding modulation (1D-DPCM).
9. according to the digital image code translator of claim 3, it is characterized in that described digital image code translator has input encoded data, and judge the mode determinant of coded system; The coded system change compression that described compressor reducer is judged according to described mode determinant is handled; Described mode determinant a certain mode of bi-directional predicted interframe coded system of predicting from past and following frame and the single directional prediction inter-frame encoding only predicted from past frame as coded system; Under the situation of the coded data of single directional prediction inter-frame encoding, described compressor reducer does not compress decoding data, under the situation of the coded data of bi-directional predicted interframe coded system, and described compressor compresses decoding data.
10. according to the digital image code translator of claim 1, it is characterized in that described compressor reducer has the quantification treatment device that M pixel * N pixel is quantized as the decoding data of 1 data block; Described decompression machine carries out re-quantization to packed data, and decompress(ion) is the decompression processing device of the decoding data of M pixel * N pixel;
Described quantification treatment utensil has the different a plurality of quantizers of quantized character;
The predetermined characteristic that described compressor reducer has the characteristic searcher of the predetermined characteristic of the decoding data that detects described M pixel * N pixel, detected according to described characteristic searcher is selected 1 quantizer and the quantizer selector of the decoding data of quantification M pixel * N pixel in the quantizer of selecting of institute from a plurality of quantizers of described quantification treatment device;
Described quantizer selector has the decoding data of input M pixel * N pixel and obtains the maximum value detector that the maximum of the difference value between pixel adjacent is exported again, the decoding data of input M pixel * N pixel is also obtained the minimum detector that the minimum value of the difference value between pixel adjacent is exported again, the characteristic quantization table of the minimum value that is used for quantizing the maximum that detects by described maximum value detector and detects by described minimum detector, maximum that input is detected by described maximum value detector and the minimum value that is detected by described minimum detector also quantize described maximum and minimum value is re-used as the characteristic quantizer that quantized value is exported with described quantized character table;
Described quantizer selector has according to described quantized value to be selected the option table of 1 quantizer and selects to be used for to compress the selector of the optimum quantizer of decoding data according to described option table from a plurality of quantizers of described quantification treatment device.
11. according to the digital image code translator of claim 1, it is characterized in that described decompression machine has a plurality of inverse quantizers, described a plurality of inverse quantizers have the re-quantization characteristic corresponding to each quantized character of a plurality of quantizers that are provided with in the described quantification treatment device; Described digital image code translator also has the controller of the re-quantization characteristic of the quantized character of setting described compressor reducer and described decompression machine;
Described a plurality of inverse quantizer is the variable inverse quantizer of re-quantization characteristic;
Described controller has the quantized character setting apparatus of corresponding to the setting of described quantized character described a plurality of inverse quantizers being set the re-quantization characteristic when described a plurality of quantizers are set each quantized character respectively, corresponding to by the setting of the quantized character of described quantized character setting apparatus and described option table is set to the option table setting apparatus of quantizer selector, corresponding to by the setting of the quantized character of described quantized character setting apparatus and described characteristic quantization table is set to the characteristic quantization table setting apparatus of characteristic quantizer.
12. the digital image interpretation method with digital image code translator of decoder and frame memory comprises the steps:
The coded data that interframe/interior coding is crossed is decoded as the decoding step of decoding data of the data block unit of M pixel * N pixel;
The decoding data of being deciphered the data block unit of the M pixel * N pixel that obtains by above-mentioned decoding step is carried out quantification treatment, and the compression of the row decoding data of going forward side by side is handled;
For the coded data that interframe/interior coding is crossed being deciphered, the prediction storing step more than the frame in the prediction frame memory that remains on frame memory by above-mentioned compression step compressed and supplied data with above-mentioned decoding step;
Remain on display frame storing step in the display frame memory of frame memory being used for the subsequently displaying transmitted image data of displayed image;
From above-mentioned prediction frame memory, read packed data, and carry out re-quantization and handle, carry out the decompress(ion) of packed data again, then data necessary is offered the 1st decompress(ion) step of above-mentioned decoding step; And
From above-mentioned display frame memory, read packed data, and carry out re-quantization and handle, carry out the decompress(ion) of packed data again, the 2nd decompress(ion) step of carrying out the output of visual video data then.
13. digital image interpretation method according to claim 12, it is characterized in that described digital image interpretation method also has the compression ratio that above-mentioned digital image interpretation method also has the judgement dimension of picture and judges the compression decoding data according to the capacity and the dimension of picture of frame memory, is re-used as the compression ratio determination step that compression ratio information used in the above-mentioned compression step is notified.
14., it is characterized in that described digital image interpretation method has the controlled step of the re-quantization characteristic that the re-quantization of the quantized character of the quantification treatment of setting the above-mentioned compression step of change and the above-mentioned the 1st and the 2nd decompress(ion) step handles according to the digital image interpretation method of claim 12.
CNB971095078A 1996-03-04 1997-03-04 Method and apparatus for digital image decoding Expired - Fee Related CN1158876C (en)

Applications Claiming Priority (10)

Application Number Priority Date Filing Date Title
JP4634696 1996-03-04
JP46346/1996 1996-03-04
JP46346/96 1996-03-04
JP202492/96 1996-07-31
JP202492/1996 1996-07-31
JP20249296 1996-07-31
CA002185753A CA2185753C (en) 1996-03-04 1996-09-17 Digital image decoding apparatus
JP350305/96 1996-12-27
JP350305/1996 1996-12-27
JP8350305A JPH1098731A (en) 1996-03-04 1996-12-27 Device and method for decoding digital image

Publications (2)

Publication Number Publication Date
CN1166110A true CN1166110A (en) 1997-11-26
CN1158876C CN1158876C (en) 2004-07-21

Family

ID=27427293

Family Applications (1)

Application Number Title Priority Date Filing Date
CNB971095078A Expired - Fee Related CN1158876C (en) 1996-03-04 1997-03-04 Method and apparatus for digital image decoding

Country Status (4)

Country Link
JP (1) JPH1098731A (en)
CN (1) CN1158876C (en)
CA (1) CA2185753C (en)
FR (1) FR2745679B1 (en)

Families Citing this family (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP3271585B2 (en) * 1998-06-19 2002-04-02 日本電気株式会社 Video decoding device
US6807311B1 (en) * 1999-07-08 2004-10-19 Ati International Srl Method and apparatus for compressing and storing image data
EP1298937A1 (en) * 2001-09-26 2003-04-02 Chih-Ta Star Sung Video encoding or decoding using recompression of reference frames
JP3680845B2 (en) 2003-05-28 2005-08-10 セイコーエプソン株式会社 Compressed video decompression device and image display device using the same
JP3680846B2 (en) 2003-05-28 2005-08-10 セイコーエプソン株式会社 MOVING IMAGE COMPRESSION DEVICE AND IMAGING DEVICE USING THE SAME
JP4589709B2 (en) * 2004-12-10 2010-12-01 パナソニック株式会社 Video playback device
JP5234241B2 (en) * 2004-12-28 2013-07-10 日本電気株式会社 Moving picture encoding method, apparatus using the same, and computer program
US20090052536A1 (en) * 2006-02-09 2009-02-26 Nec Corporation Video Decoding Device, Decoded Image Recording Device, Their Method and Program
CN101395924A (en) * 2006-03-03 2009-03-25 皇家飞利浦电子股份有限公司 Differential coding with lossy embedded compression
KR101270167B1 (en) 2006-08-17 2013-05-31 삼성전자주식회사 Method and apparatus of low complexity for compressing image, method and apparatus of low complexity for reconstructing image
KR100837410B1 (en) * 2006-11-30 2008-06-12 삼성전자주식회사 Method and apparatus for visually lossless image data compression
US8761531B2 (en) * 2009-07-09 2014-06-24 Qualcomm Incorporated Image data compression involving sub-sampling of luma and chroma values
WO2012035766A1 (en) * 2010-09-16 2012-03-22 パナソニック株式会社 Image decoding method, image encoding method, image decoding device and image encoding device
CN103329527B (en) * 2011-01-12 2017-03-01 西门子公司 The compression of reference picture and decompression in video encoder

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5455629A (en) * 1991-02-27 1995-10-03 Rca Thomson Licensing Corporation Apparatus for concealing errors in a digital video processing system
US5614952A (en) * 1994-10-11 1997-03-25 Hitachi America, Ltd. Digital video decoder for decoding digital high definition and/or digital standard definition television signals
US5450130A (en) * 1994-03-30 1995-09-12 Radius Inc. Method and system for cell based image data compression
ATE246864T1 (en) * 1994-06-06 2003-08-15 Sci Worx Gmbh METHOD FOR CODING/DECODING A DATA STREAM

Also Published As

Publication number Publication date
FR2745679B1 (en) 2003-03-14
CA2185753A1 (en) 1998-03-18
CA2185753C (en) 2000-09-12
CN1158876C (en) 2004-07-21
FR2745679A1 (en) 1997-09-05
JPH1098731A (en) 1998-04-14

Similar Documents

Publication Publication Date Title
CN1135845C (en) Image transmitting method, image processing method, image processor, data storage medium
CN1194550C (en) Method and apparatus for changing resolution by direct DCT mapping
CN1202650C (en) Image processing method, image processing device, and data storage medium
CN1076932C (en) Method and apparatus for coding video signal, and method and apparatus for decoding video signal
CN1199472C (en) Variable length decoder for decoding digitally encoded video signals
CN1227909C (en) System for editing compressed image sequences
CN1166110A (en) Method and apparatus for digital image decoding
CN1242620C (en) Transcoder-based adaptive quantization for digital video recording
CN1203679C (en) Method and device used for automatic data converting coding video frequency image data
CN1638484A (en) Image encoding apparatus, image encoding method, image encoding program, image decoding apparatus, image decoding method, and image decoding program
CN1950832A (en) Bitplane coding and decoding for AC prediction status and macroblock field/frame coding type information
CN1366776A (en) Decoder and decoding method, recorded medium and program
CN1550110A (en) Moving picture signal coding method, decoding method, coding apparatus, and decoding apparatus
CN1168322C (en) Image coding/decoding method and recorded medium on which program is recorded
CN1428742A (en) Device and method for coding and decoding keyword data
CN1543223A (en) Image encoder and image decoder
CN1197252C (en) Image coding and decoding by using conversion coefficient of grade information of corresponding image element block
CN1489391A (en) Space predicting method and apparatus for video encoding
CN1270541C (en) Decoder and method thereof, coding device and method thereof, image processing system and method thereof
CN1397137A (en) Data processing device
CN1652608A (en) Data processing device and method of same, and encoding device and decoding device
CN1240225C (en) Picture coding device and picture coding method
CN1115054C (en) Picture signal coding device, picture signal coding method, picture signal decoding device, picture signal decoding method, and recording medium
CN1163076C (en) Picture coding device, picture coding method, picture decoding device, picture decoding method, and recording medium
CN1154235C (en) Image coding and decoding using mapping coefficients corresponding to class information of pixel blocks

Legal Events

Date Code Title Description
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C06 Publication
PB01 Publication
C14 Grant of patent or utility model
GR01 Patent grant
C17 Cessation of patent right
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20040721

Termination date: 20140304