TW492255B - Region-based information compaction as for digital images - Google Patents

Abstract

A method and apparatus for preserving the dynamic range of a relatively high dynamic range information stream, illustratively a high resolution video signal (51), subjected to a relatively low dynamic range encoding and/or transport process(es). The invention subjects the relatively high dynamic range information stream to a segmentation and remapping process (10) whereby each segment is remapped to the relatively low dynamic range appropriate to the encoding and/or transport process(es) utilized. An auxiliary information stream (54) includes segment and associated remapping information such that the initial, relatively high dynamic range information stream may be recovered in a post-encoding (i.e. decoding) or post-transport (i.e., receiving) process.

Description

492255

In this book, A% is sufficient. Please refer to Case No. 09/05 0, 3 04. At 3/30/98, Suwa visits the digital color image based on the region-based information compression part of the meter. All references are incorporated herein by reference. The present invention is generally related to a data processing system, and in particular, Mu Ming is related to a method and device for storing a relatively high dynamic range of an information signal (such as a color image mm number). The processing operation is By: Relatively low dynamic range information processing system. The invention f In some communication systems, the transmitted data is compressed so that the available bandwidth can be used more efficiently. For example, the Moving Pictures Experts Group (Moving pictures Experts Grup_MpEG) has promulgated several standards related to moving picture compression and digital data transmission systems. The first type, called MP E G-1, can be referred to the International Standards Organization (Is 0) / I E C Standard 11172 ', which is hereby incorporated by reference. The second type, called MpEG_2, can refer to the International Standards Organization (I SO) / IEC standard 1 3 8 1 8 and is incorporated herein by reference. A compressed digital video system is described in the Digital Television Standards Document A / 53 of the Advanced Television Systems Co. (ATSC), which is incorporated herein by reference. The above reference standards describe data processing and manipulation techniques that are well-suited for video, audio, and other digital transmission systems that use fixed or variable length 'especially' the above reference standards and other " similar to MPEG '. Intra-frame coding techniques (such as run-length coding, Huffman coding, etc.) and inter-frame coding techniques (forward or backward predictive coding)

page

492255

/ l · (2) (forward an (J backward predictive coding), mobile compensation ^ Γ π ο Li on compensa t i on), etc.) to compress the image information. Especially, taking color image processing system as an example, MPEG and MPEG-like image processing systems are characterized by prediction-based image frame compression coding, with or without frame-intra-frame and / or inter-frame motion compensation. coding. τ Regarding gangster images (or single image frames), several well-known standards are used for effective compression of image information. For example, the Joint Photographic Experts Group (JPEG) has issued several standards related to still image compression, the most famous being js 10 918-1 (1 丁 11-1'1 \ 81 ) Standard, this; ^ Jin; ^ Tian Ji: ^ 千 付 千 This is the first of many collections of standards for still image compression. $ = ... processing and digital image processing, information like pixel acceptance and color depth of image points are encoded with 0 to ^ one-bit integers. For example, the room usually uses a bit image point; shadow; TV work news, which will produce a video of poor households and TV studios located between 0 and 1023, and t, 10 Kinetic energy ~ Sequoia value. Only for the use of electricity, the video information of the above-mentioned 夂 mr heart stem is usually the traditional system). Usually only a dynamic range of 8 bits is used. use. The two quasi-communications are one of the users of information, such as movies, television, or other information provided to a terminal. The scope of the limit is limited, and the source will be deteriorated due to the information coding network. The way for users to provide such information. Therefore, a method and device are provided to ensure the security of the system. It is based on relatively low dynamic range technology, private video, and other relatively high dynamic range

492255 厶, the invention is called ^ ___ m 肜 Λ's Robbery Fan® is 冇 its necessity. g ^ Preservation, and to provide such dynamic range techniques efficiently (e.g., the mpegs referenced above are rooted in these relatively low motion necessary " cup standards and techniques) at an appropriate ratio are the current stream of the present invention. Relatively low motion versus relative: each segmented transmission operation is connected with an encoding device to accompany, a $ solution μ save a relatively high dynamic range resolution image signal, and code and / or transmit you Sound ^ I is used for m assignment. The invention is a needle cutter. J and remapping operations, which are suitable for use in (A8 π to get more) encoding and / or the affiliate stream contains the segmented area and the initial, relatively high dynamic range information stream can be transmitted (equal to receiving) and returned in the operation. The invention is used to encode an information frame. The information frame is divided into a number of information regions, including a method and dynamic range. For example, the dynamic range (many) compilation dynamic range information flow region is remapped lower dynamic range. (Equal to decoding) or a method based on this explicit method includes the following steps: at least one information region includes at least one information parameter, and the information parameter is accompanied by a plurality of regions of the dynamic range of the education parameter defined by the upper and lower limits. Intra (i * egi〇n) value 'determines the individual maximum and minimum values of at least one information parameter for each area of at least one information area; for each area of at least one information area, the maximum value based on individual definitions And re-mapping each intra-region value based on the minimum value and the intra-region value based on at least one information parameter to encode each information region.

V. Description of the Invention (4) Summary The present invention is described in Shu Eight] content may be readily understood by considering n $ Bu η hand, 'field description and the related drawings are not described round 1 j, a distributed information system; Summary of SLV-He? Flowchart of stream encoding method and decoding method; numerous areas; segmentation into circles using ~ point coordinate technology 3B description-the image has been divided into macro block areas; A non-linear figure 4B describes a legend of a machine with a number of 4Λ; a legend; and a flowchart of a, non-linear coding function of a code function 5 describes a coding and decoding method. ▲ Set one of the South-order function block flows For the sake of understanding, the same reference value P 彳 Quan Ding is the same component between the diagrams. May have been adopted to indicate that j_ sheep refers to that 2 :: those skilled in the art will clearly understand the relatively high dynamic range of any information processing system in its coding), and subsequent processing For example, 2 pairs of low dynamic range processing (for example, the original high dynamic range information or _ = 乂) will produce ideal conditions. Although the identification of the present invention t I also can be used by the U member.. For example, MPEG and other views: processing Soil: The scope of moving image processing. Those who know the art will recognize that 'wood inventions such as m: (; similar to the second can be immediately applied to single or static image processing (for example, application 5 ^, for image processing ) ° More generally, the content of the present invention can be used ^ ^ — or multiple information parameters, together with a relatively high dynamic range of dynamic range M machine form. The present invention provides reductions for use in processing or transmission, for example. R1 1's ability of 并且 'and then re-recovered the dynamic range. Fix 1 time _,., 7 pieces · The information distribution system 1 00, which uses a pre-processing range enhancement: "... 10-bit dynamic range information The stream is encoded to produce one percent and two meters "Stream" and use an 8-bit encoding process (for example) MPEG > Bay I snail code total encoding information to produce another, 8-bit complete encoding information stream, the 8-bit aw " The enhanced information stream is transmitted to, for example, a receiver, which is located at the receiver ^ 々 々 u: outside the 8-bit coded information stream to perform a decoding process, for example, a decoding program of the type ^ to generate an 8-bit Meta-decoded information stream. One-time wide = rationale, using this range to enhance the information stream to enhance the 8-bit decoding ^ ^ two moving range, so that the original 10-bit dynamic range can be realized. The system 100 of FIG. 1 includes: an information coding section (1 0-3 0) is suitable for use, for example, an information provider is used by a television studio; an information policy, section (35), For example, a communication channel such as a land-based broadcast channel, and a tribute decoding section (40-60) are suitable for, for example, an information user with an appropriate decoding device. The information coding section contains An area mapping and proportional unit 丨 〇, used for self-information sources like- Source (not shown), receives _ relatively high dynamic range information signal S1, for example, a 10-bit dynamic range video signal. ^ Area mapping and scaling unit 10 cuts each by a 10-bit video signal S 1 It constitutes 492255, and / u / Ι recognizes u / j, which is a typical part of a sub-circle image area, a sub-frame area, or a sub-field area. The operation of the area mapping and scale unit 10 will be performed in accordance with Figure 2 below. ° Simply put, 'Each area of the many areas will be processed: identify' ^ an example and & One of the image points used = degree level (ΥΜΛχ) and-minimum brightness level (YMIN). Pressing =: in each area will change the original 10-bit dynamic range (equal to 0 to 1023) == Or reduce (equivalent to remapping) to the relative area Ϊ = 立;,) and the maximum brightness level (,) ^ 喟 ~ :: dynamic range, used to generate -8 bit basis 4 in the output part Like k number S3. The maximum value associated with each area = ^ The information identified from the area is connected with the-output section, the situation mapping of the area where the mapping area ID signal is possible, 8-bit -H i mapping and The proportional unit 10 receives the 8 bits again: ΐ ΐ see! Or video) signal S3, the encoder 15 video m base frequency sees the video signal to encode to produce a compressed;: ", like an MPEG-type video basic stream.

The video encoder 20, like an M source (not shown), receives a base m: device, and the self-view signal S2- is generally the same as the base bandwidth = phase; the base frequency, the base 20, the baseband bandwidth audio signal Perform :; = V ⑽ ^ ^^ ^ MPEG „^ ^ 4; // ^ ^ ^ ^ ^ m 4I this flow path, I must pay attention

苐 Page 10 492255 to ‘Sound’ Λ code; ^ and other audio functions to be introduced later are not absolutely necessary for the implementation of the present invention. A service multiplex E25, which maps the area ID signal S4, the basic flow path S5, and the audio basic flow path S 1 6 into individual different length or fixed length packaging structures. 'Generally referred to as packaged basic stream (packetizede 1 ementary streams) , PES), the packaged basic flow path (pES) is combined to form a multiplexed packaged basic flow path (PES) S6, and the packaged basic flow path (peS) structure provides, for example, video, audio, and other information Decoding and representation identification and synchronization functions. A transmission encoder 3 〇 encapsulates the multiplexed encapsulated elementary stream (PES) S 6 encapsulated elementary stream (p £ s) into a fixed-length transmission packet in a predictable manner to generate a transmission stream 7. I should note that the mapped area ID signal S4 can be connected to an end user (such as a decoder) in many ways in the case of different communication standards. ~ As described in the MPEG and ATSC standards, the combination of the user ’s private breeding list and private feed or information descriptor with the mapping area 丨 D signal S4 ’can place the entire message at the designated location. The inventor of this case also noticed the use of such materials and other MPEG, ATSC, DVB and similar private, user, or satellite transmission formats. For example, because the mapped area ID signal S4 includes information related to the coded area information located in the basic flow path §5, in a specific embodiment of the present invention, the mapped area ID signal is in the form of private data. It is included in the multiplex basic flow path. The mapped area ID signal S4 can communicate with an ancillary data stream, an MpEG type data stream, or a user's private data or information stream. Private data can include one and one special Packet identification (PID) -related data streams, privacy or encryption

492255 * # «/]%« ;! (S) user data, while the interpolation is like another path containing the cluster path S5) {, r,] such as the combination of the #basic flow WW private fork section. Two =: ί: ΐZhao implements a shot, the transmission encoder enhances the stream in the shape of '...' state. Combining values with ... Invented another packet identification ⑺⑸ Second, the completed encoded information stream is incorporated into a package stream (packet two transmission encoding s ream). In addition, the transmission encoder includes two parts in a header before a package stream. The information dispersion section includes a communication network 35, such as optical fiber, radio, or other public or private data transmission. Receiving the f-thru project of the 7 μ communication network ^^ ^ ,,., Σ ,, m „, (i? Ij Λΐρ: 〇 ^ a non-synchronous material mode (ATM) format) generated by the information coding section m is replaced by 4 Hai;:;: =) Packets are received to transport stream II-Multiplexed into the area Saki S-,-Video elementary stream === Fundamental = 492255

The path S10 corresponds to the area ID signal S4, the basic flow path S5, and the sound basic flow path S16, respectively.泫 泫 Video basic stream s 丨 2 Decodes in a known way by a video decoder 5 5 to generate a 8-bit base bandwidth corresponding to the remapped 8-bit base bandwidth video signal s3 Video signal § 1 3. The basic audio stream sj 〇 is decoded by an audio decoder 50 in a known manner to generate a baseband audio output signal S11 corresponding to the baseband audio signal S2, and an audio processor (not shown) ) For further processing.

A reverse area mapping and scaling unit 60 receives an 8-bit baseband video ^ 虎 su and a mapping area ID signal S14. The reverse area mapping and scaling remaps the 8-bit baseband frequency early on the basis of the area. Wide video signal si3, an area for generating a 10-bit video signal S15 corresponding to the original 10-bit dynamic range video signal §1, the generated 10-bit video signal and a video processor (not (Shown) for further processing. The operation of the reverse area mapping and proportional unit 60 will be described in more detail with reference to FIG. 2 below. In short, the reverse region mapping and scaling unit 60 self-maps the area ID signal S14 to obtain the previously identified accompanying image, frame,

t = = (field region) the maximum brightness level (γ_) and the minimum brightness level 7 (ιμιν), and receive any necessary identification information to obtain a special area within the 8-bit 7L baseband video signal S i 3 The associated maximum and minimum values are linked. The brightness information accompanying each area is then scaled up or down (equal to remapping) to the original 8-bit dynamic range that is limited by the minimum brightness level (ymin) and maximum brightness level (YMAX) that accompany the area. 10-bit (equal to 0- 丨 〇2 3) dynamic range to produce 10-bit

Page 13 492255 V. Description of the invention (ίο) Meta video signal S 1 5. Those skilled in the art will recognize that other high dynamic range parameters accompanying an information signal (such as color primitives, high dynamic range audio information, etc.) can also be effectively processed using the device and method of the present invention. As described previously, the mapped area ID signal S4 can communicate with the end user through, for example, a variety of methods in the form of different communication standards. Therefore, in a specific embodiment of the present invention, the mapped area ID 丨The D signal S4 can be returned from a private data section of one of the side transmission channels. In another specific embodiment of the present invention, the mapped area ID signal is accompanied by a packet identification (PI) value and is used for reply. In another specific embodiment of the present invention, the coded information stream is returned from an encapsulation flow path accompanied by a predetermined packet identification (/ if) value. At week, the mapping information area d letter ί ^ 1 A flowchart of the encoding method and decoding method of the Besson flow path with f 2 spoons as the front end of the encapsulation flow path with a predetermined packet identification (PID) value. When the high dynamic range information flow is 200 and is received at step 210 = proportional and proportional unit 10, the method is entered each time the method is received. Method 20 proceeds to step 215, where the standard frame is at 205. ° The logical frame of the cut is based on the image. Based on the image, the frame is added to the area. The standard of the box-shaped 205 includes: implant, and and move: once the field, fragment Macro blocks, blocks and image points are fixed or changed in the coordinate area of the video information stream m, etc. 'The size of the area. In the moon, any typical area contains a macro block

V. Description of the invention (11) Proceed to step 220, after castration into the area (step 215), the method 200 with the largest and smallest values will be at the place of ^-or the number of the ones we are concerned about, the most important example, and the number one. Mother: Regional decision. Video-to-Video: Bayson signal is a motion vector and so on. S—bone parameter (Υ), color discrimination parameter (U, V), method 20 0 and then advance; ^ The parameters within the two_ will be remapped to two "25, where each: the image point system of the domain -Range of parameter values ... Limited by individual maximum and minimum parameter values. Lo Fen Chui Zhu N. That is, if the parameter of an image point is a brightness parameter, all of them in a particular γ θ ^ n JU, ^ ^ & domain The brightness parameter will be remapped to a range determined by the maximum brightness value and the minimum shell value determined by the previous step 22 0 in the special region. :: The above steps for the segmentation of the ear frame wood, the maximum and minimum parameters are determined. 4 The mapping includes the necessary steps to generate an information flow and an enhanced flow path with dynamic range. To be clear, the purpose is to follow up at a relatively low dynamic range. The reduction of the target range used in the information flow caused by step f (for example, step 23 below) can be greatly modified by a second and subsequent processing steps (for example, steps 240-245 below). This concept Important to understanding the invention 0 After remapping all parameters to one or more regions (step 225), method 20 proceeds to step 230 where the information in the region is encoded to produce an encoded Information flow. Taking a video information stream as an example, the encoding may include a reference to the above-mentioned MPEG-type encoding standard. The method 20 then proceeds to step 2 35, where the information stream is encoded, the maximum of each area accompanying the encoded information stream, and Minimal data and enough information to accompany each area

Page 15 492255

492255 V. Description of the invention (13) Blocks and macro blocks can also be associated with movement parameters or other parameters. A parameter associated with an image point, block, or macro block is only defined if the information of the dynamic range of the parameter is correct, and the dynamic range of certain parameters such as the brightness of the image point is saved and such as movement The preservation of the dynamic range of other parameters is more important. Therefore, due to the limitation of the dynamic range, some parameters may be attenuated or acceptable. Other parameters should be preserved with as high accuracy as possible. . Taking the brightness parameter as an example, in an image containing a very bright area (equivalent to a high degree of sadness) and a `` very dark area (equivalent to a low brightness value), the dynamic range of ^ image shell information can be completely Utilizing, that is, accompanying, the brightness parameter value of the image point in v may be between (in a 10-bit dynamic range v ^ black) and 0 2 3 (white). Therefore, if the representative image brightness ii: state range ', for example, a 10-bit studio image, exceeds the dynamic range of image processing using: where: for example, an 8-bit MPEG-type encoding is = j, it will be calculated by mistake. Negative effects are generated on the processed image, and by dividing the image into smaller regions, the probability that the full alpha region dynamic range of luminance information will be used in a region will be reduced. The 2 domains can be selected according to any internal frame selection criteria. For example, in the case of Shi Xunxun's poor news frame, the appropriate criteria include scan lines, areas defined by == point markers, blocks, macro blocks, and cut planes. Wait, two = g, the smaller the selected area, the greater the chance of saving the sadness of the number of information units. 4 motions, describing an image 300 has been divided into areas 30 1 -307 using an image coordinate technology. In the present invention, an image using area according to FIG. 3A is used.

Page 17 Λ. · The name of the call is called __. For example, the area 仏 ^. In the example, the identification index of the position of the area contains the coordinates of the image points that describe the two = smart or edge. r (,-(; · ");! · core >% like Μ 〇 has been cut into columns (Ri — rn) and rows

The area is much smaller than the area taken by a macro block. Because it is defined in Figure 3B, the image formation area is defined in the solid 3A area. Therefore, there is a greater chance that the area can be used with less division and moving range. In the present invention, an image according to FIG. 3B contains a macro address that is defined by the number of rows (number of rows) and the number of rows defined by the column (such as the embodiment of the skeleton embodiment, the identification of the location index package); . _ Self-circle image (equivalent to self-quoting attack morning ^ # method includes the beginning of the number of macro blocks represented by a macro gap value < the number of macro blocks of a macro block) a simple linear reconstruction: : Don't Every: Area (equivalent to a macro block). -Target dynamic range image 々 :::-from-original dynamic range representation, and in which TD occupies redundancy or color parameters, γ ^ "TP == head image point · f) P π can be determined by mathematical formula 1 And 0R = p 家 ·· 』, 〇 = original shadow silly couple in the studio. Yu-since -1. Bit Λ target # β $ lyrics $) remap to —δ 位元 ^ # 疋 视频Point (like the processing system for X-members) 愔 Where knows the image f «

Female, 0 yh, brother, math formula 1 is pretty sloppy 4 (like using MPEG in an 8-bit tl image point transcript into math + π ^ ^ occupies remapping back to 10 bits & + formula 2. Similarly, Eight 1 variables and mathematical formula 3. Five 2 Cantonese image fir L, 4 We should note that in the case of # 豕 点, the mathematical value of 1— 円 or the result represents the removal of ++ ~~ bottom function operator Integer point. TP = L〇p * (TR / OR) + 0.5 ″ (eq. 1) (called 2) (eq · 3)

TP = L〇p ★ (256/1024) + 0.5 ″ TP = Lop * (1024/256) + 0.5J 492255 5 Description of the invention (15)

Using mathematical formula 2, a 0P equal to 525 will produce a secondary equal to 丨 ", using mathematical formula 3, a equal to 131 〇p will generate a secondary equal to 524. We have found that the dynamic range from 10 to 10 bits The process of linear mapping to 8-bit dynamic range and mapping back to 0-bit dynamic range will result in the loss of information due to calculation errors. W. The above mathematical formula 丨 -3 illustrates mathematically based on the current remapping The calculation error 1 in 2 2 is comparative. In contrast, the following remapping mathematical formula is suitable for the area mapping and scaling unit 丨 0 and the reverse mapping mapping and scaling unit 6 of Figure i.

In a specific embodiment of the linear remapping function of the wind invention, a typical linear remapping function like Shujin # 4 is used, where TP = target = point, OP = original image point; TR = Heading range; Μχ = maximum number of parameters and i ^ and 1 ^ 1. Minimum parameter values. Taking a 10-bit system with a regional minimum of 400 and a regional maximum of 600, as an example, mathematical formula 4 becomes mathematical

Tp = I (OP-MIN) * (TR / (MAX-MIN)) + 〇.5j (eq. 4)

Tp = L (OP-400) * (TR / (600-400)) + 0.5j (eq. 5)

As far as the contents of the distance are not deducted, as long as the difference between the maximum and minimum parameter values, ^ over the range ratio defined by the original dynamic range and the target dynamic range ^ The function of Math. 4 will be able to original image points The relative high bit of the parameter is ^ i. That is, taking a 10-bit original dynamic range and an 8-target dynamic range as examples, the ratio is 丨 〇2 3: 2 5 5 (equal to 4: 丨),

Page 19 492255, invention description (16), the difference between the maximum and minimum values must not be greater than one quarter of the original dynamic range. Therefore, a gate level for each area's dynamic range will be constructed to determine whether the complete, original parameter dynamic range will be preserved by the present invention. Because in the mathematical formula 5, the difference between the maximum value (^ QQ) and the maximum value of j (4 0 0) is less than one-fourth of the dynamic range of 10 bits— (2 5 6), a complete 10 bit Meta dynamic range will be saved. We must note that Mathematical Formulae 4 and 5 should not be considered to limit the scope of the present invention in any case, but Mathematical Formulae 4 and 5 are just one of the many linear functions that represent a suitable use of the present invention. The present invention can also be performed using non-linear functions (such as gamma correction and compancling functions), and the present invention can use a combination of linear and non-linear functions to optimize data compression. The selected linear and / or non-linear function will change depending on the following factors: the type of information flow to be processed, the distribution of the shape in the information unit of the information flow, and the dynamic range allowed for an application. Numerical values, encoding and / or decoding processing restrictions on assets, and other criteria. / 、 In order to help confirm that the difference between the maximum and minimum values continues to decrease: ", the area of the area is necessary. However, the extra maximum and minimum resources that Yu Sheng must be identified and dealt with will now be burdened. Prove as non-dominant. The Yuan i method effectively provides the subsequent selection of the selected information in an information shelf. One: The preservation of the complete dynamic range of the number, which is considered to be I. The early yuan is used to apply to Taigongri *. The extension is used in the present invention, for example, because the minimum and maximum image point values of each area of an image are generated to discuss. Understand the sentence, μ; J Block 玍 is also responsible for, and will now be misunderstood by ’such as the communication network 3 5 in Figure 1,

Page 20 492255 V. Description of the invention (17) The number of bits lost will be discussed. Consider a 10-bit dynamic range of the brightness parameter of the video information stream processing based on the 8-bit dynamic range. Assume, for example, that a 16 × 1 6 block of an 8-bit image point (gray scale) A small area has been selected. The 16 x 16 square of the 8-bit image point is represented by 256 * 8 bits = 2048 bits. If a minimum and maximum two 10-bit values are added to this square, the bit The number will increase from 20 to 2 0 8 bits, or equal to 1%. In this case, the bit resolution of the image point will never be worse than 8 bits, and may be the same as 10 bits. This is one of the four improvement factors for the brightness density resolution. Consider a case of a 10-bit digital video stream based on a well-known 4: 4: 4 format. In this example, the luminance (γ) and color difference (U, v) signals all have a 10-bit dynamic range. Once again, a small area such as a 16 × 16 block of an 8-bit image point (gray scale) has been selected. The 8-bit image point is composed of 256 * 8 * 3 bits = 6144. Bit. Also in this example, the block adds six 10-bit values, that is, the minimum and maximum values of each brightness value (γ) and color difference :, 〇i, V), then the number of bits increases to "by The bit increase Λ is almost 1%. In this case, the per-brightness value ⑺ (此,) will never be worse than 8 bits, and may be the same as 10 bits. One of the four improvement factors of resolution. Present = f: r Example 'If all image points are represented by ⑺ bits: 1 will be 256 * 1 0 = 256 ° bits, in other words.,:, The 10-bit table will be more illegal than the above area: ... this method improves the dynamic range by 2 "without increasing the number of bits at the same time. Even more: clever improvement ’and Ya ^ 乂; ^ 兄’ use a lot of mistakes

P.21 / i, "I / quote8") encoding and decoding of liveness, microtuning standard, m: G and J 丨 WG type dead image R to MPEG and MPEG type compression, Sichuan, this method Variations can be realized '-technology long known to him) to provide a 10-bit (or, Γ, 1, sub 8-bit integrated circuit group cost savings of the above specific two of the present invention :: : "; Wai. The desired result is ϋ, and the information processed can be achieved in some compressive methods. For example, 纟, it is best to use a non-linear method to say gamma correction to compare δίι signal is processed by non-linearity (such as the non-linearity of the pipeline. Based on: for example, image gamma correction and other functions in the television, I < non-linear mapping method can be used and linear and non-linear methods can be combined Let = the dynamic range of the fundamental signal, which is suitable for using non-linearity in the mapping function because the original range and the target range are different. Another assumption is that the compression method cannot be improved), far (even using the above-mentioned brightness, non- Linear mapping is when using ^ 's correctness impairment, in this example range) is in the range Some points in the original shirt image point value (equal to the dynamic 4B described below, where the knife 11-the situation will be accompanied by Figure 4A and the news 祜 切 Αβ ^ # 乂 low position where the range (such as 0-131) of the second In order for the ratio to be in a higher bit range (eg, 132 on the side of 3; = not; the legend code 4 of the sex coding function, the legend includes the Yuanwei 420 humanity range of 41〇Α and 25 5bit-target dynamic two: Dagger One / number_, 44〇Α has -1024-bit dynamic range is 42: new mapping to 255-bit dynamic range space. The first part uses a substantially linear conversion function, while the second part 440Α uses

苐 22 492255 Λ · minutes, 丨 / UV, called (---------: 縻% η for j, the formula, that is, within the range of 0-m within a primitive circle image u marks Qin η, and the range of 132 to 1023 in the original circle image is reduced to the range of 3 ～ 2 5 5 in the circle image. W 4 1 > 彳 4 One of the encoding functions in Α is the non-linear decoding factor of the & in the decoder to retrieve the information encoded according to the remapping function described in Figure 4 Signal, 哕: encoder will use-described in circle 4B with remapping with conversion function: circle 5 description of encoding and decoding method and device according to the present invention

Hierarchical function square circle, the sun and the moon say ‘the encoding and decoding method and process. It includes a field-type imager 5 3 0 ′. The imager responds to information received from, for example,: point, point source 510. The function mapper remaps the information according to the different function standard provided by the source 520 = S1 to generate a remapping information stream S3 and an associated mapping information stream to re-block shots tfl capture and code-4 ( M is accompanied, and the 3 _ line is encoded to generate a coded information stream s5. The 10 yuan reverse editing and the negative signal stream S 3 and Ying Cheng Zi * ^ y _ port A 7 ′ will be transmitted individually to a Decoding ram 33 ° and a reverse 映 mapper 56.

The decoder 550 inputs and compiles the information of the stone horse & performs decoding by taking one of the information streams corresponding to the re-ejected information stream. 、、 Reverse: The function J 玦 成器 5 6 0 performs a reverse function mapping on the message stream according to the transmission mapping information stream, a # ^ ψ ^ ^ ^ ψ -ΓΤ ^ ^ "Heyuan" Beishe ... Ying Zhibeiji ~, I must note that the reverse function

The 492255 (2D) generating S 5f; 0 may include linear and / or non-linear changes to suit some special applications (such as gamma correction, etc.). The operator should note that the function mapper 5 3 0 and the reverse function mapper 5 6 0 can be physically mapped to the area describing the circle 1 and the proportional unit 10 and the reverse area are mapped. The operation is performed in the same manner as in the proportional unit 60.

In a specific embodiment of the present invention, the remapping function is constituted by, for example, a function mapper 530 or a region mapping and a scale unit 10 according to an arbitrary function. An arbitrary function that remaps parameters such as image point brightness or color parameters from the original dynamic range to a target dynamic range can be represented by Mathematical Formula 6, where D 2 is the target image point; 0P = original image point; TR = target dynamic Range; 0R = original dynamic range; MAX = maximum value; M IN = minimum value; and F = arbitrary function. TP = F (〇P, MAX, M1N, TR) (called 6)

I must note that the function F can take many forms and can be used in many ways. For example, the function F can be used: 1) like the simple linear function described above in relation to Figures 1 and 2; 2) — The 仂 口 玛 correction function changes the brightness level of the input video, which corresponds to the brightness response level of the display device; 3) an arbitrary polynomial; or 4) a function table (that is, a function uses only a comparison table As described, each input bit should go to an address in the table to retrieve the content stored in it). For a remapping using a fixed gamma correction function, the following forms of functions can be used:

Page 24 492255 V. Inventive call (eq. 7) τι) = | ^ [(ΟΡ-ΜΙΝ ^ ΤΚΛΜΑχ-ΜΙΝΓΙ + Ο 』″ Take remapping using a parabolic (XHX) polynomial fragment as an example, a JV列 The function in the form of column F can be used, assuming that the polynomial fragment is less than 0 and not greater than the target dynamic range: < TP = L [(OP-MIN) 2 + (OP-MIN) * (eq. 8)

TR / [(MAX-MIN) 2 + (MAX-M1N)] + 0.5J Take a remapping using a table function as an example. The table contains an indexable array, where the index value is the original dynamic range. And the values in the table are included in the target dynamic range β, which allows any mapping between the two ranges. Unless 'like gamma correction, the mapping operation is only: Yi Er T' "does not want to make " unmapped, clear situation occurs), and then located in the inverse function table or reverse mapping of Xieyi 33 and The proportion unit 60 will revert to the original information value. Every two people should / the main core arrives, the dynamic range enhancement flow and the area identification flow path are used in the same way to describe carrying additional or other information to enable Yu back to) according to the present invention-part of the dynamic range of the information flow being processed. Only S ^ 3 different embodiments of the content described in the present invention have been detailed, 'Tian: and Describe &' It is known to those skilled in the art that other specific embodiments can be constructed immediately based on these contents.

苐 Page 25, 492255 Case No. 89107083, H 3 amendment. Year /] u Correction diagram, simple description, component symbol description, 10 area mapping and scale unit, 15 video editing stone horse, 20 audio stone horse, 25 months Device 30 transmission stone horse 35 communication network 40 transmission calcite horse 45 servo demultiplexer 50 audio decoder 55 video calcite horse 60 reverse area mapping and scale unit 100 system 300 image 510 image point source 520 Function Standard Source 530 Function Mapper 540 Encoder 550 Decoder 560 Inverse Function Mapper

O: \ 63 \ 63722-910401.ptc Page 26

Claims (1)

492255 Amendment No. 89107083 6. Scope of Patent Application 1. A method for encoding an information frame, including the following steps: Cut (2 1 5) the information frame into a plurality of information areas, at least one of the information areas Containing at least one information parameter, the parameter has a number of internal values associated with it that are defined by the upper and lower limits of the dynamic range of one of the information parameters; for each of at least one of the information areas (2 2 0) the Individual maximum and minimum values of at least one information parameter; remapping (2 2 5) internal values of individual multiple regions of the at least one information parameter for each of the at least one information area and the individual maximum and minimum values determined according to the information; And encode each information area. 2. The method according to item 1 of the patent application range, wherein the encoding and determining steps respectively generate an encoded information stream (5 5) and a related dynamic range enhancement stream (5 4). The encoding method further includes the following steps: Transmission encoding (2 3 5) is performed on the encoded information stream and the related dynamic range enhancement stream to form a transmission stream for subsequent transmissions. 3. The encoding method according to item 2 of the patent application range, wherein the transmission encoding step includes the following steps: including a private data portion of the transmission stream and the dynamic range enhancement stream 4. The encoding method according to item 2 of the patent application range, wherein the transmission encoding includes the following steps: linking the encoded information stream and the related dynamic range enhancement flow path to the phase O: \ 63 \ 63722-910401.ptc Page 27 492255 _Case No. 89107083 彳 月 # Yuet__ Sixth, the scope of patent application Corresponding packet identification (PI) value. 5. The encoding method according to item 2 of the patent application scope, wherein the transmission encoding includes the following steps: incorporating the encoded information stream into a packet flow path; and including the related dynamic range enhancement flow path into the packet flow path. The front end. 6. The coding method according to item 1 of the scope of patent application, wherein: the information frame includes an image frame, and at least the information parameter includes at least a brightness parameter and a color parameter; and the encoding and determination steps individually generate an encoded Image stream and a related dynamic range enhancement flow path. 7. A method for decoding a coded information frame represented by a plurality of coded information regions located in a coded information stream, wherein at least one of the plurality of coded information regions includes at least one information parameter, and The parameter has a plurality of internal values of the remapped regions associated with the parameter. The method includes the following steps: decoding each region of the plurality of encoded information regions (240) to form a corresponding plurality of decoded information regions, The decoded information region represents a decoded information frame; from a dynamic range enhancement flow path related to the encoded information stream, each of the information parameters having at least a plurality of internal values of the remapped regions associated with the decoded information stream is retrieved (2 4 5) the corresponding maximum and minimum values; and according to the corresponding maximum and minimum values, for each of at least one information area having at least one of the values in the plurality of remapped areas associated with it O: \ 63 \ 63722-910401.ptc Page 28 492255 _Case No. 89107083 This year, more than this month Amendment_ VI. Scope of patent application One less information parameter for reverse remapping (2 4 5). 8. The decoding method according to item 7 of the patent application scope, wherein the encoded information stream and the dynamic range enhancement stream related to the encoded information stream include a part of a transport stream, and the decoding method further includes the following steps: Demultiplexing (40) the transport stream to recover the encoded information stream and the dynamic range enhancement stream. 9. The decoding method according to item 8 of the patent application scope, wherein the demultiplexing step includes the following steps: extracting the dynamic range enhanced stream from a private data portion of the transport stream. 10. The decoding method according to item 8 of the scope of patent application, wherein the demultiplexing $ step includes the following steps: Retrieving the encoded information stream and the related dynamic range enhancement according to the related packet identification (PID) value #. Flow path. 1 1. The decoding method according to item 8 of the patent application scope, wherein the demultiplexing step includes the following steps:. The code is retrieved from a packet flow path related to a preset packet identification (p ID) value Information flow; and-extracting the relevant dynamic range enhancement flow from the front end of the packet flow path related to a preset packet identification (PID) value. 12. The decoding method according to item 9 of the scope of patent application, wherein the decoding information frame includes an image frame and the at least one information parameter includes at least one brightness parameter and a color parameter. 1 3 · — a type of coded information in a 4 stream O: \ 63 \ 63722-910401.ptc Page 29 492255 Case No. 89107083 Amendment 6. Device for decoding the coded information frame indicated by the patent application area, wherein at least one of the plurality of coded information areas contains at least one piece of information Parameter, the parameter has a plurality of internal values of remapping regions related to it, the device includes: a decoder (5 5), which decodes each region of the plurality of encoded information regions to form a corresponding plurality of decoded information regions , The decoded information area represents a decoded information frame; and a reverse mapping and scaling unit (60) for enhancing a flow path from a dynamic range related to the encoded information flow, each of which has at least a correlation with it Mining the corresponding maximum and minimum values of the information parameter of the multiple internal values of the remapping area; According to the corresponding maximum and minimum values, at least one of the information parameters of each at least one information area having at least one internal value associated with a plurality of remapping areas is reverse mapped and the * scale unit reverse remapped. 14. The device according to item 13 of the scope of patent application, wherein the encoded information stream and the dynamic range enhanced stream related to the encoded information stream include relevant parts of a transport stream, and the device further includes: The transport stream transmits a decoder (40) and a demultiplexer (45) in response to the encoded information stream and one of the dynamic range enhancement streams. O: \ 63 \ 63722-910401.ptc Page 30