FR2745679A1 - Digital image decoder e.g. for cable or broadcast TV - Google Patents

Digital image decoder e.g. for cable or broadcast TV Download PDF

Info

Publication number
FR2745679A1
FR2745679A1 FR9702505A FR9702505A FR2745679A1 FR 2745679 A1 FR2745679 A1 FR 2745679A1 FR 9702505 A FR9702505 A FR 9702505A FR 9702505 A FR9702505 A FR 9702505A FR 2745679 A1 FR2745679 A1 FR 2745679A1
Authority
FR
France
Prior art keywords
data
block
section
quantization
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
FR9702505A
Other languages
French (fr)
Other versions
FR2745679B1 (en
Inventor
Hideo Ohira
Kenichi Asano
Toshiaki Shimada
Kohtaro Asai
Tokumichi Murakami
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
Priority to JP4634696 priority Critical
Priority to JP20249296 priority
Priority to CA002185753A priority patent/CA2185753C/en
Priority to JP8350305A priority patent/JPH1098731A/en
Application filed by Mitsubishi Electric Corp filed Critical Mitsubishi Electric Corp
Publication of FR2745679A1 publication Critical patent/FR2745679A1/en
Application granted granted Critical
Publication of FR2745679B1 publication Critical patent/FR2745679B1/en
Anticipated expiration legal-status Critical
Application status is Expired - Fee Related legal-status Critical

Links

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/50Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
    • H04N19/503Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving temporal prediction
    • H04N19/51Motion estimation or motion compensation
    • H04N19/577Motion compensation with bidirectional frame interpolation, i.e. using B-pictures
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/102Methods 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
    • H04N19/115Selection of the code volume for a coding unit prior to coding
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/102Methods 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
    • H04N19/124Quantisation
    • H04N19/126Details of normalisation or weighting functions, e.g. normalisation matrices or variable uniform quantisers
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/134Methods 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
    • H04N19/136Incoming video signal characteristics or properties
    • H04N19/14Coding unit complexity, e.g. amount of activity or edge presence estimation
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/134Methods 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
    • H04N19/146Data rate or code amount at the encoder output
    • H04N19/15Data rate or code amount at the encoder output by monitoring actual compressed data size at the memory before deciding storage at the transmission buffer
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/134Methods 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
    • H04N19/157Assigned coding mode, i.e. the coding mode being predefined or preselected to be further used for selection of another element or parameter
    • H04N19/159Prediction type, e.g. intra-frame, inter-frame or bidirectional frame prediction
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/169Methods 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
    • H04N19/17Methods 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
    • H04N19/172Methods 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
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/169Methods 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
    • H04N19/177Methods 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]
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/42Methods 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
    • H04N19/423Methods 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
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/42Methods 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
    • H04N19/423Methods 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
    • H04N19/426Methods 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
    • H04N19/428Recompression, e.g. by spatial or temporal decimation
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/60Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding
    • H04N19/61Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding in combination with predictive coding
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/102Methods 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
    • H04N19/124Quantisation
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/134Methods 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
    • H04N19/146Data rate or code amount at the encoder output

Abstract

The decoder includes an image frame memory (103) which receives frame information from a decoder (101) and passes the information to a frame compressor (102). A frame expansion section (104) following the frame memory reads and expands stored words. The frame memory is formed from a frame prediction memory which stores predicted coded data and a frame display memory which stores coded data used for display. The compressor compresses the decoded predicted frame data before storage. The decoded display frame data are compressed before storage.

Description

 The present invention relates to the decoding of digital images and more specifically the compression of images when decoding digital images, to reduce the required storage capacity of an image frame memory, and to further reduce the alteration of the output image, which may be caused by the compression algorithm, as a result of the adaptive application of compression based on the size of the image data. Decoding digital images must be implemented in a digital image system such as digital cable television and digital broadcasting.

FIGS. 54 and 55, appended to the present application, represent the block diagram and the external memory card of an image processing device of the prior art, namely the device bearing the designation
SGS-Thomson, STi3500, described in a manual published by SGS-Thomson Microelectronics.

 In FIG. 54, the reference numeral 501 designates a microcomputer interface, the reference numeral 502 a FIFO (first-in, first-out) memory, the reference numeral 503 a start code detection unit, the numeral reference 504 an I / O (input / output) memory unit, reference numeral 505 a variable length decoder unit, reference numeral 506 a decoder unit, reference numeral 507 a processing unit display, the reference numeral 508 an external memory, the reference numeral 550 a microcomputer link interface line, the reference numeral 551 a microcomputer bus, the reference numeral 552 of the data transmission lines, the reference numeral 553 of the data lines, the reference numeral 554 an external memory bus and the reference numeral 555 an input / output line.

 In FIG. 55, the reference numeral 601 designates a bit buffer, the reference numeral 602 an on-screen display (OSD) memory, the reference numeral 603 a first predictive frame memory, the reference numeral 604 a second predictive frame memory, and the reference numeral 605 a display frame memory.

 We will now describe the operation of the device of the prior art. Encoded data accumulated in the bit buffer 601 of the external memory 508 is sent via the external memory bus 554 to the start code detection unit 504, in which the start code of the encoded data is detected. . Once the start code has been detected, the portion of the coded data following the start code is sent through the FIFO 502 to the variable length decoder unit 505, in which the coded data is subject to variable length decoding. Then the variable length decoded decoded data is processed and decoded by the decoder unit 506. A decoded picture is stored in the external memory 508 via the I / O unit of memory 504.

 External memory 508 includes the first predictive frame memory 603, the second prediction frame memory 604, and the display frame memory 605. Each of the memories 603, 604, 605 stores decoded images. The image data used to predict the other frames are recorded in the first or second predictive frame memory 603, 604. The image data used solely to control the display is written to the display frame memory 605 .

 The data written in the display frame memory 605 is then read in synchronism with signals such as horizontal / vertical synchronization signals in television pictures and is sent to the display processing unit 507 by the display. intermediate of the external memory bus 554.

 Alphanumeric character data to be displayed in the on-screen display (OSD) memory 602 of the external memory 508 can be accessed as in the display frame memory area 605, and then sent to the display processing unit 607 via the external memory bus 554. If the data in the OSD memory 602 is valid, the display processing unit 507 places the data output from the OSD memory 602 into recovering on the data read from the display frame memory 605 and delivers the overlay data to the outside.

 In this manner, the prior art system displays an image based on the display data that has been stored in the external memory 508.

 In the above-mentioned prior art digital image decoding device, the external memory 508 must store all the data required by the decoding tables. More particularly, if data, which extends over adjacent frames, are to be encoded, all data of other associated frames used to encode an image frame must be stored in external memory 508 to decode the image data. picture of this picture frame.

 Therefore, the decoding technique of the prior art requires a data storage device of enormous size to store the associated picture frames. The large capacity required for external memory 508 is a net benefit because of the large size and high cost of building such memory.

 In order to eliminate the aforementioned problems, an object of the present invention is to provide a digital image decoding device and method, which enables a reduction of the material by effectively using a memory capacity.

 Another object of the present invention is to provide a method and a device for decoding digital images having a memory having the lowest memory capacity possible, and to minimize image corruption.

 These and other objects are achieved by the present invention as will be described hereinafter in more detail.

 According to an important aspect of the present invention, a digital decoding device for decoding encoded data of an image with a given format may comprise an image frame memory having the possibility of storing the encoded data on a frame-by-frame basis. frame, a decoding section for decoding the frame-framed data and outputting decoded data, a compression section for compressing the decoded data and delivering compressed data, and an expanding section for reading and expanding the data compressed memory stored in the frame memory and deliver expanded data.

 The decoding section decodes the encoded data containing profile information of an encoding method for the encoded data. The digital image decoding device may further include a profile evaluation section for receiving the encoded data and evaluating the profile of the encoding method. The compression section, which includes a plurality of compression modes, receives the profile information and selects one of a plurality of optimum modes for the encoding method.

 The compression section may include a plurality of quantizers, each of which includes a table for a single quantization and delivers a unique quantized result of the decoded data, a selector of an optimal table for comparing the unique quantized results to select an optimal table for the quantized quantized data. decoded data, among the plurality of tables, and a selector for selecting an output signal from one of the plurality of quantizers having the optimal table selected by the selector of the optimal table.

 The digital image decoding device may further include a degree of compression evaluation section for receiving image format information to indicate the given format of the image and to evaluate a degree of compression for the compressed data. to be stored in the frame memory based on the given format of the image and the capacity of the frame memory. The compression section compresses the decoded data based on the degree of compression and sends the compressed data to the frame memory. The expansion section reads the compressed data into the frame memory and expands the compressed data based on the degree of compression.

 The compression section may be equipped with a plurality of compression modes, and select one of the plurality of modes. The selected mode produces a smaller amount of compressed data than the frame memory capacity.

 The compression section may include a quantization section for quantizing the decoded data on a block by block basis of M x N pixels to output the compressed data on a block-by-block basis. The expansion section may include an expander for suppressing the quantization of the compressed data on a block-by-block basis and outputting the expanded data according to the block-by-block basis of M x N pixels.

 The quantization section may include a plurality of quantizers, each of which has a unique quantization feature. The compression section may include a feature search section for searching a feature of the decoded data on a block-by-block basis of M x N pixels, and a quantizer selector for selecting one of the plurality of quantizers in the section. quantization based on the characteristic sought by the feature search section, and activate the selected quantizer exclusively to quantize the decoded data on a block-by-block basis of M x N pixels. The quantizer selector may include a maximum value detector for receiving the decoded data on a block-by-block basis of M x N pixels and calculating a maximum value of a difference between adjacent pixels and delivering a maximum value as a first feature a minimum value detector for receiving the decoded data on a block-by-block basis of M x N pixels and calculating a minimum value of the difference between adjacent pixels and outputting a minimum value as a second characteristic, a quantization table of characteristics for respectively quantizing the first characteristic of the maximum value and the second characteristic of the minimum value, a quantizer of characteristics for receiving and quantizing the maximum and minimum values with reference to the quantization table of characteristics, and respectively delivering maximum values and minimal quantified es. The quantizer selector may further include a selection table for selecting one of the plurality of quantizers in the quantization section based on the quantized maximum and minimum values, and a selector for selecting one of the plurality of quantizers. , optimum for the decoded data, based on the selection table.

 The expansion section may include a plurality of quantization suppressors, each of which has a unique quantization suppression characteristic corresponding to a unique quantization characteristic of the plurality of quantizers in the quantization section. The digital image decoding device may further include a control section for controlling the unique quantization characteristics of the plurality of quantizers in the compression section and the unique quantization suppression characteristics of the plurality of quantization deletions in the expansion section.

 The respective quantizers in the quantization section adaptively modify the quantization characteristic. The respective quantization suppressors in the expansion section modify the quantization characteristic in a manner corresponding to the modification of the quantization characteristic. The control section may include a quantization quantization / quantization setting section for setting the respective quantizers to modify the unique quantization characteristic and set the respective quantization suppressors to modify the quantization quantization unique feature. , an adjustment section of the selection table for setting the quantizer selector to refer to the selection table according to the setting of the unique quantization / quantization suppression characteristics, and a setting section of the characteristic quantization table to set said characteristic quantizer to refer to the feature quantization table in accordance with the setting of the unique quantization / quantization suppression features.

According to the process according to the invention, the steps indicated below are carried out. The method includes (a) decoding encoded data using inter coding
frames / intra-frames on a block-by-block basis
M x N pixels, and compress the data of M x N pixels
decoded on a block-by-block basis, by
quantification and delivery of compressed data on
a block-by-block basis, (b) storing, on a frame-by-frame basis, a frame
dictating compressed data on a block basis by
block, in a memory of predictive frames of a
frame memory, the predictive framework being used
to decode the coded data by means of coding
inter-frame / intra-frame, (c) store a frame for displaying compressed data
on a block-by-block basis in a frames memory
display of the frames memory, the display frame
chage being used to display an image, (d) expand the compressed predictive frame data,
read in the memory of predictive frames, by means
a suppression of the quantification of the data
compressed predictive framework, and submit
dilated predictive framework data at said step of
decoding, and (e) expand the compressed display data read from
the memory of display frames, by means of a
suppression of quantification of data
display frames, and deliver data
dilated display frames as data
image display.

 The method may further include the step of evaluating a degree of compression of the decoded data on a block-by-block basis, based on an image format evaluated by the encoded data in connection with a storage capacity of the frame memory and provide said compression step with the degree of compression as compression degree information.

 The method may further comprise the step of controlling an adjustment and modifying a quantization characteristic for quantization during said compressing step, and setting and modifying a quantization quantization characteristic for suppressing quantization during said quantizing steps. 'expansion.

 Another field of applicability of the present invention will appear on reading the detailed description given below. However, it will be understood that the detailed description and a specific example, although indicating preferred embodiments of the invention, are given by way of illustration only, since various changes and modifications in the scope of The invention will become apparent to those skilled in the art upon reading this detailed description.

Other features and advantages of the present invention will emerge from the description given below taken with reference to the accompanying drawings given by way of illustration and which are not limiting of the present invention and among which
FIG. 1 is a block diagram of a digital image decoding device arranged in accordance with the present invention;
FIG. 2 is a view showing different types of picture frames;
FIG. 3 represents a bit map of a frame memory;
FIG. 4 illustrates an operation of the frame memory;
FIG. 5 is a flowchart illustrating a compression process;
FIG. 6 is a flowchart illustrating another compression process;
FIG. 7 is a view illustrating a quantization operation;
FIG. 8 is a bit map of a predictive frame / display frame memory section;
Fig. 9 is a view illustrating a conversion of Harr, which is one of the compression systems;
Fig. 10 is a view showing a data area required by the expansion and a data area to be decoded;
- Figure 11 shows a block diagram of an expansion section A;
FIG. 12 is a view showing the structure of an expansion section B;
FIG. 13 is a view illustrating the process involved in the expansion section B;
Fig. 14 is a view showing various types of coded frame sequences;
FIG. 15 represents another view showing different types of coded picture frames;
FIG. 16 is a flowchart illustrating the compression operation;
FIG. 17 represents a schematic bitmap of a predictive frame memory;
Fig. 18 shows a block diagram of a digital image decoding device according to a second embodiment of the present invention;
Fig. 19 shows a flowchart of the digital image decoding device of Fig. 18;
FIG. 20 illustrates a compression method implemented in a compression section of the digital image decoding device of FIG. 18;
FIG. 21 represents another compression method implemented in the compression section of the digital image decoding device of FIG. 18;
FIG. 22 represents another compression method implemented in the compression section of the image decoding device of FIG. 18;
FIG. 23 illustrates another compression method implemented in the compression section in the digital image decoding device of FIG. 18;
Fig. 24 is a block diagram of the compression section of the digital image decoding device of Fig. 18;
FIG. 25 illustrates a compression method implemented in another compression section of a digital image decoding device according to the present invention;
FIG. 26 illustrates another compression method implemented in the other compression section of the digital image decoding device according to the present invention;
FIG. 27 illustrates another compression method implemented in the other compression section of the digital image decoding device according to the present invention;
FIG. 28 illustrates another compression method implemented in the compression section of the digital image decoding device according to the present invention;
Fig. 29 shows a block diagram of the compression section of the digital image decoding device according to the present invention;
FIG. 30 is a block diagram of another compression section of a digital image decoding device according to the present invention;
FIG. 31 represents a block diagram of another compression section of a digital image decoding device according to the present invention;
Fig. 32 shows a block diagram of a digital image decoding device according to a third embodiment of the present invention;
Fig. 33 shows a memory map of bidirectional memory of predictive frames in the digital picture decoding device of Fig. 32;
FIG. 34 represents a memory card of a forward directional frame memory in the digital image decoding device of FIG. 32;
Fig. 35 is a block diagram of a compression section of the digital image decoding device of Fig. 32;
FIG. 36 is a block diagram of a variation of the digital image decoding device of FIG. 18;
Fig. 37 is a block diagram of another variation of the digital picture decoding apparatus of Fig. 18;
Fig. 38 shows a block diagram of a digital image decoding device according to a fourth embodiment of the present invention;
Fig. 39 shows in detail a block diagram of the compression section of Fig. 38;
Fig. 40 shows in detail a block diagram of the quantization section of Fig. 39;
Fig. 41 is a diagram illustrating the quantization characteristics of the quantizer shown in Fig. 39;
FIG. 42 represents a diagram illustrating the quantization characteristic of a quantizer q2 according to the present invention;
FIG. 43 represents a diagram representing the quantization characteristic of a quantizer q15 according to the present invention;
FIG. 44 represents information, provided for each pixel, of compressed data according to the present invention;
Fig. 45 shows in detail the block diagram of the feature search section and the quantizer selection section of Fig. 39;
Fig. 46 shows a feature quantization table according to the present invention;
Fig. 47 shows a selection table of the present invention;
Fig. 48 shows in detail a block diagram of the expansion section B of Fig. 38;
Fig. 49 shows a block diagram of a digital image decoding device according to a fifth embodiment of the present invention;
Fig. 50 shows in detail the block diagrams of a control section and a compression section of Fig. 49;
Fig. 51 shows in detail a block diagram of a quantization section of Fig. 50;
Fig. 52 shows in detail the block diagram of a feature search section and quantizer selection section of Fig. 50;
Fig. 53 shows in detail a block diagram of the expansion section B of Fig. 49;
FIG. 54, which has already been mentioned, represents a block diagram of a digital image decoding device arranged in accordance with the prior art; and
FIG. 55, which has already been mentioned, represents a bit map of the memory of picture frames of the prior art.

 Reference will now be made in detail to the embodiments of the present invention, examples of which are shown in the accompanying drawings, in which the same reference numerals designate identical elements throughout the various views.

Embodiment 1.

 Fig. 1 is a block diagram of an embodiment of a digital image decoding device according to the present invention. With reference to FIG. 1, the reference numeral 101 designates a decoder for decoding coded image data, the reference numeral 102 a compression section for compressing decoded data, the reference numeral 103 a memory section of predictive frames / picture frames comprising a predictive frame memory and a display frame memory, the reference numeral 104 an expansion section A for expanding the compressed data of a predictive frame read in the frame memory (which will also be referred to as a predictive data expansion section), and the reference numeral 105 an expansion section B (which will also be referred to as a display data expansion section) to expand compressed data from a display frame and deliver expanded data to the display unit (not shown here).

 Reference numeral 150 denotes encoded data, reference numeral 151 denotes decoded data, reference numeral 152 denotes compressed data, reference numeral 153 denotes compressed data, reference numeral 154 denotes data of display (which are also hereinafter referred to as expanded display data), and reference numeral 155 designates expanded data (which will also be referred to hereinafter as dilated predictive data).

 The operation of the device shown in FIG. 1 will be described below. The decoding section 101 decodes incoming coded data 150 using the expanded data 155 as predictive data. The decoded data 151 is then compressed in lossless or lossy fashion by the compression section 102 to reduce the amount of information contained therein. In general, using lossy compression, the data can not recover to its original state after compression, while data can return to its original state through lossless compression . The compressed data 152 is used as predictive data for an image frame to be decoded later and is also written to the predictive frame / display frame memory 103 for display. The compressed data of a frame not used for the prediction is written to the display frame memory area, while the compressed data of a frame used for the prediction is written to the predictive frame memory area. Not all data is necessarily compressed, as will be described.

 The compressed data written is expanded by the expansion section B 105 for displaying images. The expanded data is read and displayed on a display unit in the frame order used by the display unit.

 On the other hand, the expansion section A 104 accesses the predictive frame / display frame memory 103.

The resulting compressed data is then expanded and sent to the decoding section 101 as expanded data 155 (predictive data) that is required by the decoding operation performed in the decoder 101.

 The predictive frame / display frame memory 103 may be structured to have less capacity than the amount of information contained in the image data to be displayed, since the predictive frame / frame memory display 103 is adapted to store the compressed data.

 Referring now to Figure 2, reference numerals 301a to 301c designate predictive frames used to decode other picture frames; and reference numerals 301a-301d designate display frames used solely to display the images.

 Referring further to Figure 3, reference numeral 310a designates a predictive frame memory for storing a first predictive frame, reference numeral 310b designates a predictive frame memory for storing a second predictive frame, and the reference 311 designates a display frame memory for storing a display frame. A sequence of coded data 150 in a coding section (not shown here) before sending them to the decoding section 101 (1) predictive frame 301a, (2) predictive frame 301b, (3) display frame 302a (predicted with predictive frames 301a and 301b), (4) display frame 302b (predicted with predictive frames 301a and 301b), (5) predictive frame 301c, (6) display frame 302c (predicted with predictive frame 301c) (7) 302d display frame (predicted with predictive frame 301c).

 The decoded data 151 and the compressed data 152 are respectively introduced into the compression section 102 and the predictive frames / display frames memory section 103, according to the same sequence as that previously indicated from (1) to (7). . Fig. 4 shows a storage sequence of the compressed data 152 whose predictive frame / display frame memory section 103 includes a display sequence.

 The predictive frames 301a, 301b, 301c are stored in the predictive frame memory area 310a, 310b and are used to display the decoding of a predictive frame, and furthermore are used to decode the display frames 302a, 3Q2b, On the other hand, the display frames 302a, 302b, 302c, 302d are stored in the display frame memory area 311 of the predictive frame / display frame memory section 103 and used only. for the display.

 The data of the display frame so any error produced by such compression is transmitted to the other picture frame. In the case of using the lossy compression system in the compression section 102, the compression is not performed for the predictive frame 301a, 301b, 301c, while the data is accumulated in the frame memory area predictive 310a, 310b. Therefore, a transmission of the error produced by the compression to other frames is prevented.

 On the other hand, when compression is to be performed in the compression section 102 by means of the lossless compression system for complete restoration of the original data by means of compression, the original precompressed data can be perfectly restored. . Therefore, compression is performed for both predictive frames 301a through 301c and for display frames 302a through 302d. This reduces the amount of information.

Fig. 5 is a flowchart showing a compression procedure. First, the step S501 determines whether the decoded picture frame delivered by the decoder 101 is a predictive frame data item or a display frame data item. If the data is a predictive frame data, it is written in the predictive frame memory area 310a, 310b of the predictive frame / display frame memory section 103 without being compressed (not
S502). On the other hand, data used for display only is written to the display frame memory area 311 of the predictive frame / display frame memory section 103 after being compressed (not
S503). This procedure is preferable when the compression does not affect the other frame and in the case where the compression section 102 uses the lossy compression system.

When the compression performed is lossless as shown in Fig. 6, both the predictive frame data area and the display frame data area are compressed. First, step S601 determines whether the decoded picture frame delivered by the decoder 101 is a predictive frame data or a display frame data. The predictive frame data is written in the predictive frame area 310a (step S602), 310b while the display frame data is written in the display frame memory area 311 (not
S603).

 Likewise, depending on the type of picture frame, there is another case, in which it is preferable that only the predictive frame data is compressed.

 Figure 7 schematically illustrates the procedure of a compression process. The data present in a block 201 of M pixels x N rows (pixels), which have been decoded by the decoder 101 of FIG. 1, are subjected to one of a plurality of conversion processes. Since each of the pixels is represented by bits present in a number equal to r, the amount of information in a block is equal to MxNxr after the data of M pixels x N lines have been subjected to a conversion process such as a discrete cosine transform (DCT) or other conversion, they contain a low frequency signal area 292 in the upper left part, an intermediate frequency signal area 293 in the central part and a high signal area. frequency 294 in the lower right.

 Fig. 8 shows a memory data map for a compressed frame in the predictive frame / display frame memory section 103. In this figure, the reference numeral 210 denotes a location, in which the information of a compressed frame is stored, and the reference numeral 211 designates a location, at which the information of the t-th block in a compressed frame is stored.

 The compression section 102 converts the block 201 of MxN pixels according to the characteristics of the image. The converted block is subdivided into the low frequency signal area 292, the intermediate frequency signal area 293 and a high frequency signal area 294. The assignment is performed such that the number of pixels in the area of the low frequency signal is equal to rl, and the assigned number of bits in the low frequency signal area is equal to sl bits / pixel, the number of pixels in the intermediate frequency signal zone is equal to r2 and the number assigned bits in the intermediate frequency signal area is equal to s2 bits / pixel, and the number of pixels in the high frequency signal area is equal to r3 and the number of pixels allocated in the low frequency signal area is equal to s3 bits / pixel, with sl> s2> s3, and r1 + r2 + r3 = MxN. The assignment of a higher number of bits in the lower frequency area is due to the fact that the signals in the lower frequency area are more strongly affecting the image. Therefore, the effect on the image can be reduced, although the amount of data can be compressed and reduced in size.

If a quantization is performed after the execution of such an assignment of the number of bits, a quantity of information S produced in the blocks
S = rlxsl + r2xs2 + r3xs3 is kept constant.

 Therefore, addressing in a block unit may be normally required and a desired compressed and accumulated frame in a memory may be read from any block. Assuming, for example, that an antenna address in a compressed frame A as shown in FIG. 8, the address of the t-th block in the compressed frame is between (A + (tl) xS) and (A + TXS-l).

If access is made to the t-th block for decoding, one can easily access any block since the location of any compressed frame in memory is known.

Figure 9 shows the case where a conversion
Harr, which is a lossy conversion type, is used as a conversion / coding algorithm. In this figure H represents a coefficient matrix for eight pixels x eight lines to be converted.

If we assume that an image of a block before it has undergone a one-dimensional Harr conversion is X and that the converted block is B, then we have
B = HX.

If the block B becomes B 'after having been quantized and compressed, a block Y obtained after the block B' has been dilated, is H-1B,
Compression and expansion can be performed using such an operation.

Such a method is lossy compression since the number of bits is reduced under the effect of quantization after the conversion. It will be understood that the present invention is not limited to the conversion of
Harr, but can be applied in a similar way to any other conversion.

 Figure 10 shows the relationship between different areas in image data of a frame. In this figure, the reference numeral 220 designates an image frame, the reference numeral 221 a predictive block containing KxL pixels required for the decoding operation, the reference numeral 222 a group of expansion blocks required by the expansion operation.

 Referring to Fig. 10 as well as Fig. 1, the decoder 101 uses the decoded prediction block 221 or KxL pixels, which is obtained from any point of the decoded picture frame 220 as data. predictive and cumulative image in the predictive frame / display frame memory section 103. On the other hand, the data in the predictive frame / display frame memory section 103 is compressed and stored in the form block units. Therefore, when the decoded prediction block 221 of KxL pixels extends between adjacent blocks, the necessary data is not obtained by expansion of a single block.

 To solve such a problem, the expansion section A 104 picks up expansion block groups 222 containing the decoded prediction block 221 from a memory section 103 of the predictive frame / display frame memory. The expansion is executed for each block. The expansion section A 104 then extracts the data from the decoded predictive block 221 which is required by the decoding section 101, this data being sent to the decoding section 101. When the expansion section A 104 extracts the compressed data of the predictive frame / display frame memory section 103, the address of the compressed data in the predictive frame / display frame memory is subject to the aforementioned addressing.

 In this manner, the data of the decoded predictive block can be obtained from any area in the stored data. By accumulating the data of the expansion block groups 222 in an expanded data block memory (not shown) in the expansion section 104, the predictive image data necessary when the decoding section 101 is to decode the block following are provided only by updating a new necessary part. In particular, the location of the decoded predictive block required by the decoding operation is predicted by using displacement vectors between the blocks and therefore most likely reused between adjacent blocks. As a result, a predefined number of expanded blocks have been stored in the expansion section A 104. When the next block requires another block, the stored data can be updated in a block unit. This improves the efficiency of the expansion operation.

It is also preferable that a memory for storing the image data in a plurality of expanded blocks has the same arrangement as in the image frames. The data can be read according to a given sequence, for example for each horizontal line, the necessary data portion being only extracted by a gate circuit. In such a case, such a memory preferably has the same structure as that of a block line memory, which will be described later.
Otherwise, the data in the necessary range can only be read in a memory in which the image data is a plurality of blocks that have been stored, the read data being then sent to the decoder 101. In other words, only the aforementioned data of KxL pixels can be read and sent sequentially to the decoder 101.

 Fig. 11 shows a timing chart for the digital image decoding device of Fig. 1.

In this figure, the reference numeral 280 denotes a block decoding time necessary for the decoding of a block by the decoder 101, the reference number designates a compression duration necessary for compressing a block by the compression section 102, and reference numeral 282 designates an expansion time necessary to expand the necessary data (KxL pixels) required by the decoder 101 in the expansion section A 104.

 The decoder 101 decodes the coded data in the block unit during the block decode time 280.

At this time, the KxL pixel data delivered by the predictive frame / display frame memory section 103 in any start position is required as the predictive data. Therefore, the expansion section A 104 extracts necessary data from the predictive frame / display frame memory section 103 in response to a request from the decoder 101, the read data being then expanded and sent to the decoder 101. The expansion time 282 is the time required to send the data to the decoder 101 from the request sent by the decoder 101 to the expansion section A 104. The decoded data 151 is transferred from the decoder 101 to the section The transferred data is then completely compressed during a time interval, during which the decoded data 151 of the next block is transferred from the decoder 101 to the compression section 102. The compressed data is then written to the memory section of the data block. predictive frameworks / display frames 103.

 In this way, the decoding operation for encoded dynamic images can be performed in real time. Even if the decoded image data is compressed and written to the frame memory to reduce the amount of information, the system can operate without problems.

 Fig. 12 shows the structure of the expansion section B 105. In this figure, the reference numeral 270 represents an expansion section, and the reference numeral 271 is a block line memory.

 The expansion section B 105 receives the data of each block read from the predictive frame / display frame memory section 103. The input block data is first expanded in the expansion section 270. Then the expanded data is stored sequentially in the block line memory 271 at a given location for each block. The block line memory 271 has sufficient capacity to accumulate all the horizontal blocks (block lines) of the picture frame 220. Assuming, for example, that the horizontal length of the picture frame 220 includes pixels present in a number T and has blocks present in a number J, the block line memory 271 has a capacity corresponding to the number of J blocks.

 The reading of the blocks is performed for each pixel along the scan lines forming the image (i.e. in the left-right direction between the blocks) rather than in a block unit as shown in FIG. 13. In other words, the data of all the pixels of a horizontal scan line are read sequentially. When the read operation is completed for a horizontal scan line, the data of all the pixels in the next horizontal scan line are read. Such a procedure is then repeated for each scan line.

 In such an arrangement, the read operation may be performed in the frame direction by accumulating the compressed data in block units, one block line at a time. This data is delivered for displaying an image. The display signal can be output for example by reading the data of a horizontal scan line in synchronism with a horizontal synchronization signal which defines a horizontal scan line for a displayed image.

 Figures 14 and 15 show two different types of coded data sequences; Fig. 16 shows a flow chart showing the operation of the compression section 102, and Fig. 17 shows a schematic bit map of a predictive frame memory which contains the compressed data.

 As shown in FIGS. 14 and 15, the coded data sequences are of a bidirectional prediction type and a forward prediction type. More particularly, the encoded data sequence in the bidirectional prediction type is adapted to decode an image using data located in both the front and back frames as predictive data. The forward-prediction type coded data sequence is adapted for decoding an image by using data located only in the forward frame as predictive data.

 As shown in Fig. 16, the type of coded data sequence is evaluated (not S1601).

It is a sequence of forward-prediction type coded data, the decoded data is written sequentially in the predictive frame memory areas 310a and 310b, without being compressed by the compression section 102 (not S1602). On the other hand, if the coded data sequence is of the bidirectional prediction type, the data is compressed in two compressed data frames, which in turn are respectively written to the predictive frame memory areas 310a and 310b (not S1603 ).

 In this manner, the data is stored as shown in Fig. 17. More particularly, the compressed data of two frames used for the prediction are stored respectively in the predictive frame areas 310a and 310b of the frame memory section. predictive / display frames 103 if the encoded data sequence is of the two-way prediction type. This is used for the decoding operation in the decoder 101. If the coded data sequence is of the forward prediction type, the decoding operation is performed using the frame data, which is stored in the decoder. the predictive frame memory area 310a, 310b.

 Since the decoding operation can be performed without compression if the coded data sequence is of the forward prediction type, the image is not corrupted by compression. If the encoded data sequence is of the bidirectional prediction type, two predictive frames can be used to predict and encode the other frames between these two frames.

This makes it possible to obtain a more efficient coding operation. If the data compressed by the compression section 102 is stored in the predictive frames / display frames section 103, a lower capacity for that memory can be maintained.

Implementation form 2
If it is executed regardless of the aspect ratio, compression of the image may result in irreparable deterioration of the image. A poorly restored image may be the result of such indiscriminate compression processing, particularly with an image that is too small to be suitable for compression in connection with the storage capacity of a frame memory.

It is assumed that an image of a format equal to 1.1 times the format of a frame memory is subjected to a compression with a degree of non-discrimination of 50 g for example.

The comparatively small image can therefore be halved, unnecessarily and inappropriately, to be stored in the comparatively large frame memory. The half-size compressed image will appear in a catastrophic damaged state when it is displayed.

 In view of the foregoing discussion, a digital image decoding device according to a second embodiment of the present invention additionally introduces an adaptive image-format responsive compression solution of the first embodiment, in accordance with FIG. minimizing image corruption associated with compression. With this solution, an image is compressed with an adaptive compression degree optimally modified for the image format in connection with the storage capacity of a frame memory. The format of the image is included in the encoded data as part of an image information. The format of the image is identified as the number of pixels per bit width per pixel in an image. In other words, the image format is defined by T pixels / line x U lines / frame x r bits / pixel.

 Fig. 18 shows a block diagram of a digital image decoding device according to this second embodiment. The digital imager decoding device of the 2axe 18 correlates with a compression degree evaluation section 106 for receiving image format information 156 and outputting a compression degree information compression 107a for compressing decoded data 151 and outputting compressed data 152, expansion section A 108 for expanding compressed predictive data, read from a predictive frame memory, and outputting expanded predictive data 155, a section of compression B 109 for expanding compressed display data read from a display frame memory and sequentially outputting expanded display data 154 in accordance with the frame display order shown in Figs. 12 and 13 The digital image decoding device further comprises functional elements equivalent to those of Figure 1, such as the 101-inch decoder. for decoding encoded data 150 with reference to the expanded predictive data 155, and the predictive frame / display frame memory 103 having predictive frame memory areas 310a, 310b and a display frame memory area 311.

 Referring to the inventive aspects of the digital image decoding apparatus of this embodiment, the decoder 101 decodes the coded data 150, which is an encoded information element of an image including the format of the image. frame-by-frame basis to provide decoded data 151 on a frame-by-frame basis. The predictive frame / display frame memory section 103 containing the frame memory, to which a predetermined memory capacity is assigned, stores the image data on a frame-by-frame basis. The compression degree evaluation section 106 receives the image format information 156 including the image format of the encoded data 150. The compression degree evaluation section 106 evaluates the degree of compression of the decoded data. 151 which can be compressed and is stored in the frame memory on the basis of the image format in connection with the storage capacity of the frame memory. The compression degree evaluation section 106 selects a compression mode from a plurality of compression modes based on the degree of compression. The image format information 156 may not be limited to the image format defined above, but may be any identifier for identifying the size of the previously defined image.

The image format information 156 may not necessarily be included in the encoded data, but be provided outside of the compression degree evaluation section 106.

 The compression section 107a compresses the decoded data 151, decoded in the decoder 101, on a block-by-block basis according to the degree of compression evaluated by the compression degree evaluation section 106, which sends compressed decoded data as compressed data 152 (which is a generalized term that includes compressed predictive data 153a and compressed display data 153b) to the frame memory 103, for storing them.

 Expansion sections A and B 108 and 109 (which can be generalized as expansion sections) read the compressed data 102 stored in the frame memory 103 and expand the compressed data 152 based on the estimated degree of compression. by the section 106 evaluation of the degree of compression.

 The frame memory may contain a predictive frame memory for storing decoded data of a predictive frame to be used as a predictive reference for decoding the coded data 150 in the decoder 101. The compression section 107a compresses the data. decoded predictive frames to be stored in the predictive frame memory. The expansion section includes the expansion section A 108 for expanding the decoded and compressed predictive frame data, the compressed predictive data 153a stored in the predictive frame memory, and sends decoded and expanded predictive frame data as that expanded predictive data 155 to the decoding section 101.

 The frame memory may also include a display frame memory for storing decoded data of a display frame to be used for display. Compression section 107a compresses the decoded display frame data to be stored in the display frame memory. The expansion section includes an expansion section B 109 for expanding the decoded display frame data or compressed display data 153b stored in the display control memory and outputting decoded display frame data. and expanded as display data 154.

 Fig. 19 is a flow chart showing an execution sequence of the digital image decoding implemented in the digital image decoding device of Fig. 18. A series of operating steps of Fig. 19 starts with the data coded 150 which are decoded in the decoder 101 in step S1, with the expanded prediction data 155 as a reference, if any. The encoded data 150 is simultaneously sent to the compression degree evaluation section 106, wherein the image format information 156 contained in the encoded data 150 is used to evaluate the degree of compression in relation to the compression capability. memory of the predictive frame memory / display frames 103 in step S3.

 With this embodiment, the degree of compression 1m is defined by (amount of precompressed data) / (amount of post-compressed data). For adaptive and optimal compression on image data different in size, n values 11 to 1n of compression degree 1m are provided as a choice when lm, l and l ~ mn (n natural number). For example, the expression TxUxr / lm <Z is obtained with an image frame having TxU pixels and r bits per pixel, and with the respective predictive frame / display frame memory areas 310a, 310b, 311 of the predictive frame memory / display frames 103 containing Z bits for the memory capacity.

Among the plurality of choices of lmS the minimum value of 1m should be defined as the optimum ratio for the image frame to be compressed.

 In step S4, the decoded data 151 delivered by the decoding section 101 is compressed in the compression section 107a on the basis of the information of the degree of compression 157 delivered by the compression degree evaluation section 106. The compression degree information 157 may be in any other form identifying or representing the rate, degree or intensity of the compression. It can be selected from values representing ranges of degrees of compression subdivided into different ranges, and identifiers or information identifying ranges of degrees of compression and the like. The degree, the value or the information representing or identifying the degree or the intensity of the compression must correspond, as well as the information of the degree of compression 157, to a compression mode implemented in the compression section 107a. described below. Then, the compressed data 152 is sent to the predictive frame / display frame memory 103 from the compression section 107a for storage. The compressed data 152 is selected in step S5 to be written to the predictive frame memory area 310a, 310b in step S6, with a predictive frame, and to be written to the display frame memory area 311 when step S18, with a display frame. The compressed data 152 stored in the display frame memory area 311 is read as compressed display data 153b when required and is expanded in the expansion section B 109 based on the information. compression degree 157 obtained in step S9. Then the expanded display frame data is sequentially read as display data 154 in accordance with the frame display command in step S10.

 The compressed data 152 stored in the predictive frame memory area 310a, 310b is read as compressed predicted data 153a from the predictive frame / display frame memory 103 when required in the decoding section 101 for the decoding the encoded data, and are expanded in the expansion section A 108 based on the information of the degree of compression 157 in step S7. When the section de dé mine the operation.

 The digital encoding device of this embodiment greatly contributes to reducing the size of the predictive frame / display frame memory section 103 by reducing the bits of image data to be stored in a frame memory contained in this section of memory. Further, the image format sensitive compression with the compression degree evaluation section 106 and the image format information 156 allow a digital image decoding device to reduce the deterioration of the image format. image, associated with compression, to a minimum value thanks to an adaptive degree of compression optimally modified according to the image format to be compressed.

Fig. 24 shows a block diagram of the compression section 107a of the digital image decoding device of Fig. 18. With the compression section 107a, adaptive image compression is obtained using a processing command adaptive unit-by-unit depending on mode, image data based on a one-dimensional pulse-code differential modulation method (lD-DPCM). Figures 20 to 23 show different types of 1D Adaptive Control
DPCM unit per unit based on the mode, implemented in the compression section 107a. With the mode-based unit-by-unit adaptive control solution, the number of pixel-1D-DPCM compression units L is changed using a mode specified by the compression degree information 157, namely, eight pixels. (L = 8) with mode 1 in FIG. 20, four pixels (L = 4) with mode 2 of FIG. 21, two pixels (L = 2) with mode 3 of FIG. 22 and one pixel (L = 1) with mode 4 of Fig. 23. In other words, the image data in a block 201 of MxN pixels (M = 8 pixels; N = 8 pixels, r = 8 bits / pixel) are subject at a fixed quantization at four bits (p = 4) unit by unit with the unit L of DPCM compression. Basically, a difference between two adjacent pixels in an adaptive DPCM unit L is quantized sequentially with the fixed quantum of 4 bits, with a head pixel remaining unquantized in the unit.

 Fig. 20 shows a 1-DPCM compression based on eight pixels, mode 1. According to this mode, seven successive 8-bit pixels, which appear following an eight-bit leading pixel, are quantized essentially with the quantum. fixed four bits. This reduces the data bits of the initial eight bits to four bits per pixel, except for the eight leading bits, and hence from 8x8 original bits to 8 + 4x7 bits every eight pixels of the L unit. This is repeated several times. times in the 8x8 block 201 (N = 8), which reduces the data bits from the 8x8x8 start bits to (8 + 4x7) x8 bits quantized per block, with the compression degree l, 78 = (8x8x8 ) / ((8 + 4x7) x8). Figure 21 illustrates a 4-pixel, mode 2, 1D-DPCM compression. In this mode, three successive 8-bit pixels, which follow the eight leading bits in the unit, are quantized in the same way, which reduces the data bits from the original 8x8 bit to 8 + 4x3 bit every four pixels of the unit. This is repeated 15 more times in the 8x8 block 201 (N = 8), which reduces the data bits to (8 + 4x3) x16 bits per block, with the degree of compression 1.6 = (8x8x8) / ( (8 + 4x3) x16). Figure 22 illustrates a two-pixel 1D-DPCM compression, mode 3. According to this mode, the data bits are reduced to 8 + 4xl bits every two pixels of the unit in the same way, with the degree of compression 1,3 = (8x8x8) / (8 + 4xl) x32. Figure 23 illustrates a pixel-based 2D-DPCM compression, mode 4. This does not lead to any data quantization or any reduction of the data bits with a compression degree of 1 = (8x8x8) / (8x8x8).

 Referring to FIG. 24, the compression section 107a comprises a subtractor 120, a quantizer 121, a quantization suppressor 122, selectors 123a and 123b, a pixel delay circuit 124 and a selection signal generator 125 .

The information 157 concerning the degree of compression corresponds to a mode implemented in the compression section 107a. The general course of a lC-DPCM compression solution implemented in the compression section 107a can be summarized as follows. With a mode specified by the compression degree information 157, the eight-bit head pixel is a given number of pixels of the decoded data 151 and is sent directly to a pixel delay circuit 124. An output signal delivered by the pixel delay circuit 124 is subtracted from the other eight bit pixels in the unit L, in the subtractor 120. The result of the subtraction or the difference is then subjected to four bit quantization. in the quantizer 121. A quantized result is output by the compression section 107a as the compressed data 152, and is simultaneously sent to a local decoding loop, in which quantized four-bit data is locally decoded in the suppressor. quantization 122 and are sent to the delay circuit 124 on a pixel.

 Specifically, with mode 1, the selection signal generator 125 generates the select signal 159 for controlling the selector 123a to select eight bits, every eight eight-bit pixels of the decoded data 151. After the leading pixel With eight unquantized bits being output, the selector 123a selects seven successive 4-bit quantized results among the other results in the eight-pixel unit from the quantizers and to be delivered. Similarly, in mode 2, the selector 123a selects eight leading bits in the L unit of the decoded data 151 to output them directly every four pixels, with the mode 3 every two pixels and with the mode 4 each time. or for each eight-bit pixel to be delivered.

 Certain variants of this embodiment may be available for the lD-DPCM compression processes. It may not be necessary to leave in the unquantified state an eight-bit head pixel in the compression unit and instead quantize it with a quantum of t bits (ter) before it is delivered. It is not necessary that the block 201 of MxN pixels be limited to 8x8 pixels, and we can have any number of pixels when we have M = L or MfN. It may not be necessary to use the horizontal solution with the 1D-DPCM compression unit L (L <N), and instead a vertical solution may be used in the decoding digital images according to this embodiment.

The control solution of the adaptive unit 1D
Mode-based DPCM, which is implemented in the compression section 107a, can be summarized as follows.

With a mode specified by the compression degree information 157, the number of pixels of the 1D-DPCM unit L is optimally varied according to the format of the image. A leading pixel in the L unit is quantized with a quantum of t bits (ter; r bits / pixels). Among the other pixels of the unit L, a difference between two adjacent pixels is sequentially quantized with a quantum of p bits (p <r). Therefore, the image data bits in a block of MxN pixels (L <M or L <N;
L is a common divisor of M and N) are adaptively reduced on the basis of the compression degree information 157 to an optimally modified degree depending on the format of the image to be compressed.

 Figures 36 and 37 show further variants of the decoding device according to this embodiment. Fig. 36 shows a predictive frame memory 103a as a replacement for the prediction frame / display frame memory 103 of Fig. 18.

This variant has a display frame of the decoded data 151 to remain uncompressed for display and therefore no display frame memory is required. Fig. 37 shows separate frame memories of a display frame memory 103b for the display and a prediction frame memory 103a for the prediction, as a replacement for the predictive frame / display frame memory. In this variant, the decoded data 151 remains in the uncompressed state for the prediction. These two variants show that the decoded image data is not necessarily compressed and then expanded for both prediction and display.

 Fig. 29 shows a block diagram of another compression section 107b of the digital image decoding device according to the present invention. The compression section 107b is the unit replacing the compression section 107a of Fig. 24 and can be formed in the digital image decoding device of Fig. 18. With the compression section 107b, compression is achieved. adaptive images using a mode-based quantization command of the image data of the lD-DPCM compression process. Figures 25 to 28 show different types of 1D-DPCM compressions implemented in the compression section 107b. With this mode-based quantization control solution, four types of quantization are provided for compressing images having different formats with a mode specified by the compression degree information 157, namely a four-bit quantization ( p = 4) with mode 1, 5-bit quantization (p = 5) with mode 2, 6-bit quantization (p = 6) with mode 3, and 7-bit quantization (p = 7) with the mode 1. In other words, the image data in the block 201 of MxN pixels (M = 8 pixels, N = 8 pixels, r = 8 bits / pixel) are subjected to a quantification based on the mode , unit by unit, with the fixed DPCM compression unit L of eight pixels (L = 8).

Basically, a difference between two adjacent pixels in the fixed DPCM compression unit L is quantized sequentially with an optimally modified adaptive quantum as a function of the image format.

 Figure 25 illustrates a 4-bit mode 1 quantization mode with eight pixels (L = 8) quantized with a 4-bit quantum (p = 4). This reduces the data limitations to (8 + 4x7) x8 bits from the original 8x8x8 bits, and in other words the image is compressed with a compression degree of 1.78 = (8x8x8) / (8+ 4x7) x8, depending on the definition of the degree of compression (precompressed data) / (post-compressed data).

Fig. 26 shows a five-bit quantization mode, mode 2, eight pixels being quantized with a quantum of five bits (p = 5). Similarly, this compresses the image with a degree of compression equal to about 1.49 = (8x8x8) / (8 + 5x7) x8. Figure 27 shows a six-bit quantization mode (mode 3), which compresses the image with a compression degree of 1.28 = (8x8x8) / (8 + 6x7) x81. Figure 28 illustrates a seven-bit mode 4 quantization mode which compresses the image with a compression degree of approximately 1.12 = (8x8x8) / (8 + 7x7) x8.

 Compression section 107b includes a plurality of quantizers 121a-121d and corresponding quantization suppressors 122a-122d, which replace quantizer 121 and quantization suppressor 122 of compression section 107a. Then, selectors 127a and 127b are provided to respectively select one of the quantized results and the quantized results when receiving a selection signal 160 produced by a selection signal generator 129 on the Compression degree information basis 157. Selectors 123c and 123d respectively replace the selectors 123a and 123b of FIG. 24.

 Referring to Fig. 29, an eight-bit leading pixel in the eight pixels of the decoded data 151 is delivered directly as compressed data 152, and simultaneously is sent to the pixel delay circuit 124. An output signal delivered by pixel-delaying circuit 124 is subtracted, in subtractor 120, from a set of seven successive eight-bit successive pixels in the decoded data unit 151. The result of subtraction or the difference delivered by the subtractor 121 can be quantized by means of four different quantization types in the quantizers. The quantized results from the four quantizers are subjected to mode-based selection upon receipt of the selection signal 160 in the selector 127a. A selected quantized result is delivered as compressed data 152 via selector 123c and is also sent to the four quantizers for different types of quantization suppression. The results, whose quantization is suppressed and which are delivered by the four quantization suppressors, are mode-selected, using the selection signal 160 in the selector 127b. A selected result, whose quantization is suppressed or which is decoded locally, is sent to the circuit 124 performing a delay on a pixel, via the selector 123d.

 Specifically, the selectors 127a and 127b select output signals from the four-bit quantizer 121a and the quantization suppressor 122a, respectively, with the mode 1, upon receiving the selection signal 160 based on the degree information. Similarly, the output signals output from the quantizer 12 il and the quantization suppressor 122b at five bits are selected with the mode 2, output signals outputted from the quantizer 121c and the quantization suppressor 122c. six bits are selected with the mode 3, and output signals provided by the quantizer 121d and the quantization suppressor 122d at seven bits are selected with the mode 4, by the selectors.

 The general solution of the lD-DPCM compression implemented in the compression section can be summarized as follows. The DPCM compression unit L is set for compressing image data in a block of MxN pixels s (L <M or LcN; L is a common divisor of M or N).

A leading pixel among a number L of pixels is quantized with an adaptive quantum of t bits (ter). With the other pixels in the L unit, a difference between two adjacent pixels is quantized sequentially with an adaptive quantum of p bits. The value of the adaptive quantum of t or p bits may be varied on the basis of the degree of compression, which optimally reduces the data bits in a MxN pixel data block, to the format of the image.

 Referring again to the compression section 107b, the selector 127a may also be disposed upstream of the quantizers, and selects a quantizer from the plurality of quantizers so as to provide an exclusive quantization result with a mode specified by the quantizer signal. 160. If the quantizers are directly connected to the corresponding quantization suppressors and the quantization suppressors are configured to trigger operation only upon receipt of a quantized result, then the selector 127b may not be necessary. .

 Figure 30 is a block diagram of another compression section 107c of a digital image decoding device. Compression section 107c is a combination of compression section 107a of FIG. 24 and compression section 107b of FIG. 29. Compression section 107c can be formed in the digital image decoding device of FIG. 18 and the like. The functional elements of the compression section 107c of Fig. 30 correspond to those designated by the same reference numeral of the compression sections 107a of Figs. 24 and 107b of Fig. 29. With this combination, the image data is compressed. through a dual mode-based control of the DPCM compression unit, and by the quantization based on the compression degree information 157.

 With reference to the compression section 107c of Fig. 30, an 8-bit head pixel in an adaptive unit LD-DPCM L, based on the mode, of pixels of the decoded data 151 is subjected to a selection, based on the mode in the selector 123a upon receiving the select signal 159 and is output as compressed data 152. The eight-bit head pixel is also sent directly to the delay circuit 124 on a pixel, through the selector 123b. An output signal delivered by the delay circuit 124 on a pixel is subtracted, in the subtractor 120, from a next set of eight-bit pixels in the unit L. The result of the subtraction or the difference is then subject to different types of quantization in the quantizers 121a to 121d. Quantified results are mode-selected in selector 127a with select signal 160. A selected quantized result is output as compressed data 152 by selector 123a with select signal 159 or is subject to to different types of quantization quantization quantization 122a to 122d, for local decoding. The results provided by the suppression of the quantization are subjected to a mode-based selection with the selection signal 160 in the selector l27b. A locally quantized or decoded selected result is mode-selected, with the select signal 159 in the selector 123b, and sent to the one-bit delay circuit 124.

 Specifically in the case of the mode-based quantization control, the selector 127a / 127b selects a quantized result / quantized quantization result delivered by quantizer 121a / quantization quencher 122a at 4 bits, with mode 1. With mode 2, an output signal output from quantizer 121b / quantizer quantizer 122b at five bits is selected, with mode 3 an output signal output from quantizer 121cZ quantization querier 122c at six bits is selected, and with the mode 4, an output signal output from quantizer 121d / quantizer quantizer 122d at seven bits is selected by selector 127a / 127b upon receipt of select signal 160.

 With the mode-based control, the selector 123a / 123b selects one or the other of an 8-bit head pixel that remains unquantized from the decoded data 151 and a quantized result / result with quantization suppression from of the selector 127a / 127b upon receipt of the selection signal 159 from the selection signal generator 125. The selection signal generator 125 selects the number of units lD-DPCM L of pixels, namely eight with the mode 1, four with mode 2, two with mode 3 or one with mode 4, with a mode based on compression degree information 157.

This leads to a uniform quantization result delivered by the compression section 107c.

 The general solution of the compression section 107c characterized by the mode-based adaptive ID-DPCM unit, and the quantization control can be summarized as follows. Quantified data in a block of MxN pixels s are compressed unit by unit with an adaptive number of the pixel unit LD-DPCM L (L <M or L <N; L is a common divisor of M or N) with a optimum mode corresponding to the size of the image based on the degree of compression information 157. A leading pixel among a given number of pixels is quantized with an adaptive quantum of t bits (t <r). With the other pixels contained in the compression unit, a difference between two adjacent pixels is quantized sequentially with an adaptive quantum of p bits. The value of t or p of the adaptive quantum can be varied on the basis of the degree of compression. This greatly helps to minimize image corruption associated with compression.

 Referring further to the mode-based quantization command with the selection signal in the compression section, the selector 127a may be placed upstream of the quantizers, this selector selecting a quantizer to act and output the quantized result optimum for the format of the image. If the respective quantizers are directly connected to the corresponding quantization suppressors and the quantization suppressors are packaged so as to start to boot only upon receipt of a quantized result, then the selector 127b may be unnecessary.

 The reference 31 represents a block diagram of another compression section 102a according to the present invention. Compression section 102a can be performed in the digital image decoding device of FIG. 1 as an alternative to compression section 102. Compression section 102a represents quantizers having quantization heads and a selection circuit the optimal table used to select an optimal quantization table among quantizer quantization tables. With the compression section 102a, the degree of compression evaluation section 106 is not necessary.

 The compression section 102a comprises a number n of quantizers 230a to 230n respectively comprising a number n of different quantization table devices, respectively delay circuits 231a to 231n, subtractors 232a to 232n, circuits 233a to 233n for forming absolute value, accumulators 234a to 234n, an optimal table selection circuit 235, and a selector 128. The optimal table selection circuit 235 compares quantized results delivered by the quantizers 230a to 230n, to select an optimal table. of quantization among the plurality of quantization tables in the quantizers. The selector 128 selects an output signal delivered by a quantizer whose quantization table has been selected by the circuit 235 for selecting the optimal table.

 The operation of the compression section 102a will now be described.

 The decoded data 151 delivered by the decoder 101 are quantized in the quantizers 230a to 230n. If the assignment of e bits is different from the number n of quantization tables, no more than 2e (n <2e) quantization tables are expected.

 Prequantized data among the decoded data 151 is sent to the quantizers 230a to 230n to be quantized. Prequantized data among the decoded data 151 is subtracted respectively from the quantized results of the quantized data 250a to 250n provided by the amplifiers, in the respective subtractors 232a to 232n. The subtraction results are processed by the absolute value formation circuits 233a to 233n and by the accumulators 234a to 234n to form absolute sums of a difference on the basis of the LD-DPCM L compression unit.

 The optimal table selection circuit 235 selects quantized data for which the summed absolute difference is minimal, based on a block. This allows the selection of an optimal quantization table based on a DPCM unit L to obtain a compression-associated impairment of the minimum image among the plurality of quantization tables.

Implementation form 3
Fig. 32 shows a block diagram of a digital image decoding device according to a third embodiment of the present invention. A profile evaluation section 110 receives the encoded data, evaluates an encoding method, and outputs profile information 158 for identifying the encoding method. A compression section 111 modifies its compression operation based on a coded method evaluated by profile evaluation section 110. Profile evaluation section 110 evaluates whether an encoded method used in the encoded data is a two-dimensional inter-frame prediction coding method or a forward-to-block prediction coding method. The two-dimensional inter-frame prediction coding method uses both previous frames and future frames for prediction, whereas the forward-to-frame prediction coding method uses only prior frames. The compression section 111 imposes a degree of compression applied to the decoded data 151 with an inter-frame bidirectional prediction coding method, higher than in the case of compressions of the decoded data 151 with an inter-frame direct coding method.

Other functional elements of FIG. 32 correspond to those of FIG. 18 having the same reference numerals.

 Figs. 33 and 34 show maps of a predictive frame memory used respectively for bidirectional prediction and direct prediction of the digital image decoding device of this embodiment.

 Fig. 35 shows in detail a block diagram of the compression section 111 of the digital image decoding device of this embodiment. A selection signal generator 126 is of a different type from that of FIG. 24.

 The operation of the compression section 111 will be described below.

 The decoding section 101 decodes the coded data 150 with reference to the expanded predictive data 155. The compression degree evaluation section 106 evaluates an optimum degree of compression in relation to the size of the predictive frame / display frame memory. 103 on the basis of the image format information 156 contained in the coded data 150. The optimum degree of compression is selected from a number n of values 11 to 1n (n: natural number, lmo men). For example, the minimum value among a plurality of values 1m is TxUxr / tm <Z is selected for an optimum degree of compression with an image frame having a size of TxU pixels and r bits per pixel, and with frame memory predictives / display frames 103 having the capacity to store Z bits per frame memory.

 Profile evaluation section 110 evaluates whether coded data 150 has been coded using inter-frame forward prediction coding using only prior frames or by using inter-frame bidirectional prediction coding using both past and future executives. Profile evaluation section 110 sends profile information 158 to compression section 111.

 The compression section 111 compresses the decoded data 151 delivered by the decoding section 101 to reduce the data bits based on the information profile 158 issued by the profile evaluation section 110, and on the basis of the compression degree information 157 issued by section 106 for evaluating the degree of compression. With the same compression method as that of the second embodiment, for example, the selection signal generator 126 positions a 1D-DPCM compression unit on the basis of the compression degree information 157 and the profile information. 158.

 Referring to the forward / bi-directional prediction schemes of Figs. 14 and 15 respectively, the bidirectional prediction uses previous frames and future frames, which requires two memory areas for storing two types of frames, while forward prediction uses only previous frames, which requires the degree of (X) format compression of the same frame in the case of bidirectional prediction.

 With predictive image data of two-way predictive predictions and having a compression rate of one-half or less, to be stored in the prediction frame / display frame memory, the predictive image data according to the prediction in the forward direction, of equivalent size, can not be subjected to any compression in relation to the storage capacity of the memory of predictive frames / display frames. The compressed data 152, which is compressed in the compression section 111, is written to the predictive frame / display frame memory 103 for use as predictive data for a frame to be decoded.

 Therefore, one of the features of the compression section of this embodiment is that data encoded only by forward prediction is compressed with a lower (or uncompressed) compression degree than the compressing the image of the same size by bidirectional prediction.

The compressed data 152 written to the frame memory is expanded in the expansion section
B 109 and are read according to the frame display order. The expansion performed in the expansion section B 109 is based on the degree of compression information 157 delivered by the compression degree evaluation section 106.

 When the expanded predictive data 155 is needed in the decoding section 101, the expansion section A 108 accesses the predictive frame / display frame memory 103 to reach the required data and expands the compressed predictive data 153a to send the predicted data. Expander data 155 at decoder 101. In the same manner as for expansion in expansion section B 109, expansion in expansion section A 108 is based on compression degree information 157 issued by section 106 of evaluation of the degree of compression.

 As a result, the predictive frame / display frame memory 103 may, due to image compression, have a smaller capacity than the initial amount of image data to be stored. With an adaptive degree of compression optimally modified according to the size of the encoded image, the image data corruption related to compression is minimized.

 Referring again to the digital image decoding device of Fig. 32, the degree of compression evaluation section 106 may not be necessary in the system. The decoded data 151 may be compressed based solely on the profile information 158 in the compression section 111 in the absence of data reception of the compression degree evaluation section 106, which also contributes to reducing the amount of compression. alteration of the image associated with compression.

Form 4
Fig. 38 shows a block diagram of a digital image decoding device according to a fourth embodiment of the present invention.

 The digital image decoding device of Fig. 38 comprises a compression section 112, an expansion section A 113 and an expansion section B 114, which distinguishes this embodiment from the first embodiment.

 Fig. 39 shows in detail a block diagram of the compression section 112.

 A quantization section 703 includes a plurality of quantizers, each of which is assigned a different quantization characteristic. A feature search section 701 receives the decoded data 151 and searches for the maximum and minimum difference values between two adjacent pixels in a block of MxN pixels of the decoded data as a given characteristic of the decoded data. Upon receiving a characteristic signal 751 for indicating the given characteristic of the maximum and minimum values delivered by the feature search section 701, a quantizer selection section 702 selects an optimum quantizer for the given characteristic of the decoded data, among the quantizers located in the quantization section 703 and delivers a selection signal 752.

 Figure 40 shows in detail a block diagram of the quantization section 703.

 The quantization section 703 comprises 16 quantizers q0 to q1S. A unique range of quantization data as indicated in a diagram of Fig. 41 is assigned to the respective quantizers. For example, a range of data of value 0 and 255 for quantization is assigned to quantizer q2. A value range of -255 and +255 is assigned to quantizer q15 for quantization.

 Fig. 42 is a diagram illustrating the quantization characteristic of quantizer q2.

The quantizer q2 quantizes the data in a range of values 0 and 255 in ten steps 0 to 9.

 Fig. 43 shows a diagram of quantization characteristic of quantizer q15.

 Quantizer q15 has its own given range of values -255 and +255 for 10-step quantization 0-9.

 As clearly shown by a comparison of the diagrams of FIGS. 42 and 43, quantizer q2 quantizes twice that performed by quantizer q15.

 Therefore a predetermined single quantization characteristic as shown in Fig. 41 is assigned to the respective quantizers q0 to q1S of Fig. 40. When the compression section 112 compresses a block of MxN pixels (8x8 pixels for example) of the decoded data 151 , the compression section 112 selects a quantizer from the plurality of quantizers present in the quantization section 703.

 Fig. 44 shows a compressed format of compressed data 152 delivered by the quantization section 703.

 The compressed data format of Fig. 44 represents compressed data for each pixel of the compressed data. The compressed data format of Figure 44 is used in common for all 16 quantizers. The data format indicates a quantizer for exclusive quantization among the plurality of quantizers contained in the quantization section 703, with y bits. Four bits are sufficient for y bits to distinguish each of the sixteen quantizers of this embodiment. The data format has a quantization index with z bits to indicate a quantized result for each pixel. With ten quantization steps of FIGS. 42 and 43, four bits are sufficient for z bits. Therefore, a set of y bits to indicate a quantizer and z bits to indicate a quantization index is provided as information, provided for each pixel, of compressed data.

 The selection of the exclusive quantizer is carried out by means of the method indicated below.

 Fig. 45 shows in detail a block diagram of the feature search section 701 and quantizer selection section 702.

 The maximum value detector 704 receives MxN pixels decoded data 151 and detects a maximum value of the difference between two adjacent pixels. A minimum value detector 705 receives MxN pixels of the decoded data 151 and detects a minimum value of the difference between two adjacent pixels. A feature quantizer 706 receives the maximum value detected in the maximum value detector 704 and the minimum value detected in the minimum value detector 705 and quantizes the maximum and minimum values respectively with respect to a feature quantization table 781.

 Fig. 46 shows a feature quantization table 781.

The table of Fig. 46 is provided for the decoded data in a data range having values -255 and +255 (with nine bits) so as to be quantized in said quantization steps. When we have
A2 <n <A3 and (-A2) <m <(-Al), n denoting a maximum value delivered by the maximum value detector 704 and m denoting a minimum value delivered by the minimum value detector 705, AD8 is allocated by as a representative maximum quantization value and 58 is assigned as the quantized maximum value 770. Similarly, AD2 is assigned as the minimum representative quantization value and S2 is assigned as the minimum quantized value 771.

 Therefore, the feature quantizer 706 quantizes the maximum and minimum values n and m with reference to the feature quantization table 781 and delivers the quantized maximum and minimum values 770 and 771, respectively, as characteristic signals 751.

 In the quantizer selection section 702, the selector 783 introduces the characteristic signals 751 and selects an optimal quantizer with reference to a selection table 782.

 Fig. 47 shows an example of the selection table 782.

The selection table 782 of Fig. 47 is arranged based on the characteristics of the respective quantizers of Fig. 41. S8 having the quantized maximum value 770 and S2 the quantized minimum value 771 for example, the quantizer q14 is selected according to the Fig. 47. With the quantizer q14, a range of data for quantization extends from the value -A3 to the value A3 in accordance with Fig. 41. With S9 of the maximum value quantized 770 and
S5 of the quantized minimum value 771, the quantizer q2 is selected. The selection table 782 indicates a quantizer, which performs optimal quantization of data identified by the maximum and minimum values of 770 and 771, among the 16 quantizers, each of which has a unique quantization characteristic. The selector 783 provides a select signal 752 for specifying a quantizer to be selected. As shown in Fig. 40, the selection signal 752 is sent to the quantization section 703, in which a quantizer selected by the selection signal is exclusively activated. Unselected quantizers do not work. In this manner, the quantization section 703 inputs the decoded data 151 and delivers the compressed data 152.

 Fig. 48 shows in detail a block diagram of the expansion section B 114.

 The expansion section B 114 includes an expansion device 270 and a line block memory 271.

The expansion device 270 is equipped with quantization suppressors r0 to rsS. The quantization suppressors r0 to r15 correspond to quantizers q0 to q1s. In other words, the quantization suppressors r0 to r15 respectively perform a quantization suppression, in a data range corresponding to that of the respective quantizers shown in FIG.

For example, with the quantization suppressor r0, which corresponds to the quantizer qO, for example the suppressor receives the data 153b, suppresses quantization of the compressed data and delivers decoded data in a range of values A0 to A3. Specifically, upon receipt of the compressed data of Fig. 44, the expander 270 activates a quantization suppressor corresponding to a quantizer specified by y bits and suppresses the quantization of the compressed data represented by the specified quantization index. by z bits.

Quantize suppressors other than the quantization suppressor corresponding to the quantizer specified by y bits do not work. After having been expanded in the expander 270, the decoded data is sent to the line block memory 271. The continuation of the digital image decoding operations is equal to that described in the first embodiment and therefore we will not repeat them here. The expansion section A 113 (not represented by a detailed figure) is equipped with the same type of expansion device as the expansion device 270 of Fig. 48. In the expansion device, a plurality of quantization suppressors decode the compressed data.

 As previously described, the image data is compressed to be stored in a frame memory, thereby reducing the size of the predictive frame / display frame memory 103 to a value less than initial amount of image data to be stored.

 The data characteristic is calculated on the basis of a compression unit and a quantization is applied to the data by an optimum quantizer for the characteristic. This provides an optimal compression of the data to be written into the predictive frame / display frame memory 103. This reduces the memory capacity of the predictive frame / display frame memory 103 to a value less than the quantity. initial image data. In addition, this minimizes image corruption associated with compression.

 In addition to reducing the memory capacity, a reduction in the size of the predictive frame / display frame memory 103 can lead to a reduction in the address space and the data bit width for the read / writing in memory. And above all, this contributes to greatly reduce the size of the digital image decoding device and also to reduce the cost of manufacture.

Embodiment 5
Fig. 49 is a block diagram of a digital image decoding device according to a fifth embodiment of the present invention.

 The digital image decoding device of Fig. 49 contains a controller 700, which distinguishes this device from the digital image decoding device of Fig. 38.

 The controller 700 controls the quantization characteristic of a compression section 112a. The controller 700 controls the quantization suppression characteristics of the expansion sections A and B 113a and 114a.

 Fig. 50 shows in detail the block diagrams of the control section 700 and the compression section 112a.

 Fig. 51 shows in detail a block diagram of a quantization section 703a.

 Fig. 52 shows in detail a block diagram of a feature search section 701a and a quantizer selector 702a.

Referring to Fig. 50, the control section 700 includes a feature quantization table setting section 784, a selection table setting section 785, and a quantization characteristic setting section 786. The feature quantization table setting section 784 selects a feature quantization table 781a in the feature search section 701a through a command line 760 as shown in FIG. of the selection table positions a selection table 782a in the quantizer selector 702a via a command line 761. The quantization characteristic setting section 786 positions a range of quantization data in the quantizers respective quantization section 703a through a control line 762. Quantizers q0 to q15 are able to modify their quantization characteristics based on a designated range of data via the line 762.

 Referring to Fig. 51, the quantization characteristic setting section 786 assigns the quantizer q0, for example, a range of data of values 0 and A3 for quantization, via the command line 762. With the quantizer q1 a range of values -A3 and 0 is assigned for quantitation.

 Fig. 53 shows in detail a block diagram of the B expansion section 114a.

 An expansion device 270a of the expansion section B 114 is equipped with a plurality of quantization suppressors. The quantization suppressors introduce a range of data for the quantization suppression respectively by means of the command line 762, in the same manner as indicated with reference to FIG. 51. In this way, the quantization suppressors are provided in FIG. correspondence with the quantifiers.

 Referring further to the fourth and fifth embodiments of the present invention, a digital image decoding device may be based on the 1D-DPCM or 2D-DPCM compression method.

 Referring again to the second and third embodiments of the present invention, the digital image decoding device may be based on a 2D-DPCM compression method or other compression methods in place of the method. lD-DPCM compression.

 Referring again to the first four embodiments of the present invention, the coded data need not be encoded by the interframe coding method, and instead they can be encoded by means of a codec. intra-frame coding method for obtaining as high a performance as possible for reducing the size of an image frame memory and reducing the image impairment associated with the compression to a minimum value.

 Having thus described several particular embodiments of the invention, various variations, modifications, modifications and improvements will be apparent to those skilled in the art.

Such variations, modifications, modifications and improvements are intended to be included within the scope of the present invention. This is why the foregoing description is given solely by way of example and is in no way limiting.

Claims (21)

  1.  A digital image decoding device for decoding encoded data of an image having a given format, characterized in that said image decoding device comprises
     an image frame memory (103) having the capability of storing the encoded data on a frame-by-frame basis;
     a decoding section (101) for decoding the frame-framed data and outputting decoded data;
     a compression section (102, 107a, 107b, 112, 112a) for compressing the decoded data and outputting compressed data; and
     an expansion section (104,105,109,114,114a) for reading and expanding the compressed data stored in the frame memory and outputting expanded data.
  2.  Digital image decoding device according to claim 1, characterized in that
     said frame memory (103) comprises a predictive frame memory (310a, 310b) for storing the encoded data of a predictive frame to be used for the predictive decoding of the encoded data in said decoding section (101), and a display frame memory (311) for storing encoded data of a display frame to be used for display;
     said compression section (102) compresses the decoded data of the predictive frame to be stored in said predictive frame memory as compressed predictive data, and compresses the decoded data of the display frame to be stored in said frame memory. display in the form of compressed display data; and
     said expansion section (104, 105) comprises
     a prediction data expansion section (104) for expanding the compressed data of the predictive frame stored in said predictive frame memory and delivering dilated predictive data, and
     a display data expansion section (105) for expanding the compressed data of the display frame stored in said display frame memory and outputting the expanded display data for display thereof.
  3.  Digital image decoding device according to claim 1, characterized in that
     said decoding section (101) decodes the decoded data sequentially on a block-by-block basis and outputs the decoded data on a block-block basis; and
     said compression section (102) compresses the decoded data on a block-by-block basis for a period less than the period used for decoding the decoded data on a block-by-block basis in said decoding section.
  4.  Digital image decoding device according to claim 1, characterized in that
     said decoding section (101) decodes the encoded data including profile information of an encoding method for the encoded data; and
     that said digital image decoding device further comprises
     a profile evaluation section (110) for receiving the encoded data and evaluating the profile of the decoding method; and
     said compression section (112), which includes a plurality of compression modes, receives the profile information and selects one of the plurality of modes, which is optimum for the encoding method.
  5.  Digital image decoding device according to claim 1, characterized in that
     said decoding section (101) decodes the coded data on a block-by-block basis and delivers a block of M pixels x N pixels x r bits of the decoded data,
     said compression section (112) compresses the decoded data on a block-by-block basis into the data compressed on a block-by-block basis by means of a conversion method, which calculates an image quality coefficient of the decoded data on a block-by-block basis, and assigns a bit length greater than the decoded data on a block-by-block basis for a coefficient that strongly influences the quality of the image, and a shorter bit length of the decoded data for a coefficient having a weaker influence, and
     said compression section (112) converts the decoded data on a block-by-block basis into a fixed bit length of the decoded data on a block-by-block basis.
  6.  Digital picture decoding device according to claim 2, characterized in that
     said compression section (112) compresses, on a block-by-block basis, the decoded data decoded on a frame-by-frame basis and delivers the compressed data on the block-by-block basis,
     said prediction data expansion section (104, 105) reads a block of compressed block-based predicted data, including at least a portion of block data of K pixels x L lines required by said decoding section, said data of the required block forming part of a predictive frame stored in said predictive frame memory, expands the read block of the compressed predictive data on a block-by-block basis, and sends the predicted data, expanded on a block-by-block basis, of the read block containing the required block data of K pixels x L lines, at said decoding section,
     said prediction data expansion section (104, 105) comprises a block memory for storing the expanded predictive data on a block-by-block basis, the block read in said prediction frame memory, and
     said block memory is updated on a block-by-block basis whenever the block data required by said decoding section is regenerated.
  7.  Digital image decoding device according to claim 2, characterized in that
     said compression section (112) compresses, on a block-by-block basis, the decoded display data on a frame-by-frame basis and delivers the compressed display data on the block-block basis,
     said display data expanding section (114) reads the compressed display data stored in said display frame memory on the block-by-block basis, expands the read data read, compressed on the block basis by block, and delivers the expanded display data on the block-by-block basis sequentially in a horizontal scanning direction,
     said display data expansion section (114) comprises
     a block memory for sequentially storing the read block of the expanded display data on a block-by-block basis, with a width and a direction corresponding to the horizontal scanning direction, and
     said display data expanding section outputs the expanded display data on a block-by-block basis, read in said display memory in response to a display scan line of the image.
  8.  Digital image decoding device according to claim 1, characterized in that said compression section (112) comprises
     a plurality of quantizers, each of which includes a table for a single quantization and delivers a unique quantized result of the decoded data,
     an optimal table selector for comparing the unique quantized results to select an optimal table for the decoded data among the plurality of tables, and
     a selector for selecting an output signal from a quantizer among the plurality of quantizers including the optimal table selected by said optimal table selector.
  9.  9. Digital image decoding device according to claim 1, characterized in that
     it further includes a compression degree evaluation section (106) for receiving image format information for indicating the given format of the image and for evaluating a degree of compression for the compressed data to be stored in said frame memory on the basis of the given format of the image and a capacity of said frame memory, and
     said compression section (112, 112a) compresses the decoded data on the basis of the degree of compression and delivers the compressed data to said frame memory, and
     said expansion section (114, 114a) reads the compressed data into said frame memory and expands the compressed data on the basis of the degree of compression.
  10.  Digital picture decoding device according to claim 9, characterized in that
     that the encoded data includes the image format information, and
     said compression degree evaluation section receives the encoded data and extracts the format information of the image from the encoded data.
  11.  Digital image decoding device according to claim 9, characterized in that
     said compression section (112, 112a) is provided with a plurality of compression modes, and
     said compression section (112, 112a) selects one of the plurality of modes, the selected mode producing a compressed amount of data less than the capacity of said frame memory.
  12.  Digital image decoding device according to claim 11, characterized in that
     said compression section (112a) compresses the decoded data on a block-by-block basis of M pixels x N pixels x r bits by means of a quantization based on a pulse-code differential modulation (LD-DPCM) method,
     said compression section (112a) quantizes a leading pixel among a given number L of pixels (L <M) with a quantum of t bits (shot) and a difference between two adjacent pixels of the other pixels among the L given pixels with a quantum of p bits (p # r), and
     said compression section (112a) modifies at least one of the values of L, p and t in an adaptive manner to provide a plurality of compression modes.
  13.  Digital image decoding device according to claim 4, characterized in that
     that said profile evaluation section (110) evaluates the coding method as a bidirectional prediction interframe coding method for predicting a previous and future frame based framework, or an interframe coding method to forward prediction based on a previous framework, and
     said compression section (112a) compresses the decoded data involving the bidirectional prediction interframe coding method and does not compress the decoded data involving the forward prediction interframe coding method.
  14.  Digital image decoding device according to claim 1, characterized in that
     said compression section (112a) includes a quantization section for quantizing the decoded data on a block-by-block basis of M x N pixels to output the compressed data on a block-by-block basis, and
     said expansion section includes an expander for suppressing the quantization of the compressed data on a block-by-block basis and outputting the expanded data on the block-by-block basis of M x N pixels.
  15.  Digital image decoding device according to claim 14, characterized in that
     said quantization section (103) comprises a plurality of quantizers, each of which has a unique quantization characteristic,
     said compression section (102a) comprises
     a feature search section (701) for searching a feature of the decoded data on a block-by-block basis, of M x N pixels, and
     a quantizer selector (702) for selecting one of a plurality of quantizers in said quantization section based on the characteristic sought by the feature search section, and enabling a quantizer selected exclusively for quantizing the data of M x N pixels, decoded on a block-by-block basis, and
     that the quantizer selector includes
     a maximum value detector (704) for receiving the M x N decoded pixel data on a block-by-block basis, and calculating a maximum value of a difference between adjacent pixels and outputting a maximum value as a first characteristic,
     a minimum value detector (705) for receiving the M x N decoded pixel data on a block-by-block basis and calculating a minimum value of the difference between adjacent pixels and outputting a minimum value as a second characteristic,
     a feature quantization table (781) for respectively quantizing the first characteristic of the maximum value and the second characteristic of the minimum value,
     a feature quantizer (781) for receiving and quantizing the maximum and minimum values with reference to the feature quantization table and respectively delivering quantized maximum and minimum values,
     a selection table (782) for selecting one of the plurality of quantizers in said quantization section on the basis of the quantized maximum and minimum values, and
     a selector (783) for selecting one of the plurality of optimum quantizers for the decoded data based on said selection table.
  16.  The digital image decoding device according to claim 15, characterized in that said expansion section (114) comprises a plurality of quantization suppressors (ro-rl5) each of which has a unique quantization suppression characteristic corresponding to a respective unique quantization characteristic of said plurality of quantizers in said quantization section.
  17.  Digital image decoding device according to claim 16, characterized in that it further comprises
     a control section (700) for controlling the unique quantization characteristics of the plurality of quantizers (qo-ql5) located in said compression section (112) and the unique quantization suppression characteristics of the plurality of quantization suppressors in said expansion section.
  18.  18. Digital image decoding device according to claim 17, characterized in that
     that respective quantizers (qo-ql5) in the quantization section adaptively modify the quantization characteristic,
     that the respective quantization suppressors (ro-rl5) in said expansion section modify the quantization quenching characteristic in a manner corresponding to the change in quantization characteristic, and
     said control section (700) comprises
     a quantization quantization / quantization feature setting section (786) for setting the respective quantizers so as to modify the unique quantization characteristic and set the respective quantization suppressors to modify the quantization quench unique characteristic,
     a selection table setting section (785) for setting the quantizer selector to refer to the selection table according to the setting of the unique quantization / quantization quantization characteristics, and
     a feature quantization table setting section (784) for setting said characteristic quantizer to refer to the feature quantization table according to the setting of the unique quantization quantization / quantization features.
  19.  19. Device for decoding digital images, characterized in that it comprises the steps of
     decoding coded data using inter-frame / intra-frame coding on a block-by-block basis
    M x N pixels,
     compressing the data of M x N decoded pixels on a block-by-block basis, quantizing and delivering compressed data on a block-by-block basis,
     storing, on a frame-by-frame basis, a predictive frame of the block-by-block compressed data in a prediction frame of a frame memory, the predictive frame being used to decode the encoded data using the inter-coding -cadres / intracadres,
     storing a frame for displaying the compressed data on a block-by-block basis in a display frame memory of the frame memory, the display frame being used to display an image,
     dilating the compressed predictive frame data, read from the predictive frame memory, by suppressing the quantization of the predictive frame compressed data, and subjecting predictive frame expanded data to said decoding step, and
     expand the compressed display data read from the display frame memory by suppressing the quantization of the compressed display frame data, and output expanded display frame data as data from the display frame memory; picture display.
  20.  20. Device according to claim 19, characterized in that it further comprises the step of
     evaluating a degree of compression of the decoded data on a block-by-block basis, based on an image format evaluated by the encoded data in connection with a storage capacity of the frame memory, and providing said compressing step with the degree of compression as compression degree information.
  21.  21. Device according to claim 20, characterized in that it comprises the step of
     controlling a setting and modifying a quantization characteristic for quantization during said compressing step, and setting and modifying a quantization quantization characteristic for suppressing quantization during said expanding steps.
FR9702505A 1996-03-04 1997-03-03 Method and device for decoding digital images Expired - Fee Related FR2745679B1 (en)

Priority Applications (4)

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

Publications (2)

Publication Number Publication Date
FR2745679A1 true FR2745679A1 (en) 1997-09-05
FR2745679B1 FR2745679B1 (en) 2003-03-14

Family

ID=27427293

Family Applications (1)

Application Number Title Priority Date Filing Date
FR9702505A Expired - Fee Related FR2745679B1 (en) 1996-03-04 1997-03-03 Method and device for decoding digital images

Country Status (4)

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

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0975178A2 (en) * 1998-06-19 2000-01-26 Nec Corporation Moving picture decoding apparatus and moving picture decoding method
EP1075146A2 (en) * 1999-07-08 2001-02-07 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

Families Citing this family (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP3680846B2 (en) 2003-05-28 2005-08-10 セイコーエプソン株式会社 Compression of video image apparatus and an imaging apparatus using the same
JP3680845B2 (en) 2003-05-28 2005-08-10 セイコーエプソン株式会社 Decompression of compressed video device and an image display apparatus using the same
JP4589709B2 (en) * 2004-12-10 2010-12-01 パナソニック株式会社 Video playback device
KR100945985B1 (en) * 2004-12-28 2010-03-09 닛본 덴끼 가부시끼가이샤 Moving picture encoding method, device 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
JP2009528751A (en) * 2006-03-03 2009-08-06 コーニンクレッカ フィリップス エレクトロニクス エヌ ヴィ Differential encoding
KR101270167B1 (en) 2006-08-17 2013-05-31 삼성전자주식회사 An image of a low complexity compression method and apparatus, an image of a low complexity restore method and apparatus
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
JP5710022B2 (en) * 2011-01-12 2015-04-30 シーメンス アクチエンゲゼルシヤフトSiemens Aktiengesellschaft Compression method and decompression method of the reference image in the video decoder

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5450130A (en) * 1994-03-30 1995-09-12 Radius Inc. Method and system for cell based image data compression
US5455629A (en) * 1991-02-27 1995-10-03 Rca Thomson Licensing Corporation Apparatus for concealing errors in a digital video processing system
EP0687111A2 (en) * 1994-06-06 1995-12-13 SICAN, GESELLSCHAFT FÜR SILIZIUM-ANWENDUNGEN UND CAD/CAT NIEDERSACHSEN mbH Method for coding and decoding a data stream
EP0707426A2 (en) * 1994-10-11 1996-04-17 Hitachi, Ltd. Digital video decoder for decoding digital high definition and/or digital standard definition television signals

Patent Citations (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
US5450130A (en) * 1994-03-30 1995-09-12 Radius Inc. Method and system for cell based image data compression
EP0687111A2 (en) * 1994-06-06 1995-12-13 SICAN, GESELLSCHAFT FÜR SILIZIUM-ANWENDUNGEN UND CAD/CAT NIEDERSACHSEN mbH Method for coding and decoding a data stream
EP0707426A2 (en) * 1994-10-11 1996-04-17 Hitachi, Ltd. Digital video decoder for decoding digital high definition and/or digital standard definition television signals

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP0975178A2 (en) * 1998-06-19 2000-01-26 Nec Corporation Moving picture decoding apparatus and moving picture decoding method
EP0975178A3 (en) * 1998-06-19 2003-08-27 Nec Corporation Moving picture decoding apparatus and moving picture decoding method
US7333542B1 (en) 1998-06-19 2008-02-19 Nec Corporation Moving picture decoding apparatus and moving picture decoding method
EP1075146A2 (en) * 1999-07-08 2001-02-07 ATI International SRL Method and apparatus for compressing and storing image data
EP1075146A3 (en) * 1999-07-08 2002-07-31 ATI International SRL Method and apparatus for compressing and storing image data
US6807311B1 (en) 1999-07-08 2004-10-19 Ati International Srl Method and apparatus for compressing and storing image data
EP2227024A1 (en) * 1999-07-08 2010-09-08 ATI Technologies ULC 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

Also Published As

Publication number Publication date
CA2185753A1 (en) 1998-03-18
CN1158876C (en) 2004-07-21
CN1166110A (en) 1997-11-26
FR2745679B1 (en) 2003-03-14
JPH1098731A (en) 1998-04-14
CA2185753C (en) 2000-09-12

Similar Documents

Publication Publication Date Title
EP1863295B1 (en) Coded block pattern encoding/decoding with spatial prediction
EP1096802B1 (en) Device for predicting and decoding image
EP1478190B1 (en) Image encoding device, image encoding method, and image encoding program
EP2268036B1 (en) Video signal decoding method
EP0849953B1 (en) System and method for performing motion compensation using a skewed tile storage format for improved efficiency
KR100430328B1 (en) Video data encoder and decoder
EP0399487B1 (en) Transformation coding device
EP1274249B1 (en) Image data encoding method and apparatus
US6330284B1 (en) Video coding and video encoding apparatus according to a reduction ratio setting information signal
US7289674B2 (en) Spatial prediction based intra coding
US6229854B1 (en) Video coding/decoding apparatus
EP0535960B1 (en) Coding signal processor
US7016547B1 (en) Adaptive entropy encoding/decoding for screen capture content
EP1156680A2 (en) Improved video coding using adaptive coding of block parameters for coded/uncoded blocks
EP0739143A2 (en) Video signal coding system and method
US9344714B2 (en) Moving picture coding method, and moving picture decoding method
EP0771120A2 (en) Video encoding and decoding apparatus
EP1345448A2 (en) Image coding apparatus and image decoding apparatus
US5384598A (en) System and method for frame differencing video compression and decompression with frame rate scalability
JP3013698B2 (en) Vector quantization coding apparatus and the decoding apparatus
US5235618A (en) Video signal coding apparatus, coding method used in the video signal coding apparatus and video signal coding transmission system having the video signal coding apparatus
US6639945B2 (en) Method and apparatus for implementing motion detection in video compression
US5815601A (en) Image encoder and image decoder
US7567719B2 (en) Method for encoding images, and an image coder
USRE40079E1 (en) Video encoding and decoding apparatus

Legal Events

Date Code Title Description
ST Notification of lapse

Effective date: 20141128