公开(公告)号：US6049630A

公开(公告)日：2000-04-11

申请号：US962927

申请日：1997-10-27

申请人： Zhengrong Wang ,  Paul Steven Houle

发明人： Zhengrong Wang ,  Paul Steven Houle

IPC分类号： G06T9/00 ,  H03M7/30 ,  H03M7/40 ,  H03M7/46 ,  H04N1/413 ,  H04N7/26 ,  H04N7/30

CPC分类号： H03M7/46 ,  H04N19/117 ,  H04N19/126 ,  H04N19/14 ,  H04N19/152 ,  H04N19/172 ,  H04N19/176 ,  H04N19/186 ,  H04N19/192 ,  H04N19/60 ,  H04N19/61 ,  H04N19/63 ,  H04N19/10 ,  H04N19/115 ,  H04N19/13 ,  H04N19/146

摘要： A method and apparatus for adaptive bit allocation and hybrid lossless entropy encoding in a lossy compression system. The invention includes three components: (1) a transform stage to decorrelate image data into a baseband and multiple subbands, (2) a quantization stage to quantize the resulting transform coefficients, and (3) a lossless entropy coder stage to encode the quantized indexes. In the preferred embodiment, the transform stage uses a wavelet transform algorithm. The quantization stage adaptively estimates values for parameters defining an approximation between quantization size and the logarithm of quantization error, and recursively calculates the optimal quantization size for each band to achieve a desired bit rate. The baseband and subbands are transformed into quantization matrices using the corresponding quantization sizes. The lossless entropy coder stage uses the observation that the entropy property of run lengths of zero index values in the subband quantization matrices is different from the entropy property of non-zero indices. Each quantization matrix is parsed so that each non-zero index is extracted into a separate stream, and the remaining position information is parsed into an odd stream of run length values for "0" and an even stream of run length values for "1". These three streams are Huffman coded separately in conventional fashion. This hybrid algorithm gives an approximately 10% percent improvement over conventional run length and Huffman coding for similar images. The overall compression algorithm gives about 2.about.6 dB improvement in terms of peak signal-to-noise ratio (PSNR) over JPEG algorithms for similar images at similar bit rates.

摘要翻译： 一种用于有损压缩系统中的自适应比特分配和混合无损熵编码的方法和装置。 本发明包括三个部分：（1）将图像数据解相关到基带和多个子带中的变换阶段，（2）量化级量化所得到的变换系数，以及（3）无损熵编码器级对量化索引进行编码 。 在优选实施例中，变换阶段使用小波变换算法。 量化级自适应地估计用于定义量化尺寸与量化误差的对数之间的近似的参数的值，并递归地计算每个频带的最佳量化尺寸以获得期望的比特率。 使用相应的量化尺寸将基带和子带变换成量化矩阵。 无损熵编码器级使用子带量化矩阵中的零索引值的游程长度的熵特性与非零指数的熵特性不同的观察。 每个量化矩阵被解析为使得每个非零索引被提取到单独的流中，并且剩余位置信息被解析为“0”的游程长度值的奇数流和“1”的游程长度值的偶数流 。 这三个流以常规方式分别进行霍夫曼编码。 这种混合算法比常规运行长度提供了大约10％的改进，并且对于类似的图像给出了霍夫曼编码。 对于相似比特率的JPEG图像，整体压缩算法在峰值信噪比（PSNR）方面提供了大约2dB的改进。

公开(公告)号：US6018594A

公开(公告)日：2000-01-25

申请号：US119739

申请日：1998-07-21

申请人： Shinichi Kikuchi ,  Tetsuya Kitamura ,  Hideki Mimura ,  Kazuhiko Taira

发明人： Shinichi Kikuchi ,  Tetsuya Kitamura ,  Hideki Mimura ,  Kazuhiko Taira

IPC分类号： H04N1/64 ,  F42B10/64 ,  G06T9/00 ,  G11B20/00 ,  G11B20/12 ,  H03M7/46 ,  H03M13/51 ,  H04N1/41 ,  H04N5/85 ,  H04N7/32 ,  H04N7/52 ,  H04N9/804 ,  H04N9/806 ,  H04N9/82 ,  H04N21/236 ,  H04N21/434 ,  G06K9/36

CPC分类号： H04N9/8042 ,  G11B20/00007 ,  G11B20/1217 ,  H03M7/46 ,  H04N19/93 ,  H04N21/236 ,  H04N21/434 ,  H04N5/85 ,  H04N9/8063 ,  H04N9/8227

摘要： A bit train of a plurality of continuous pixels is compressed according to a new run-length compression scheme. In this run-length compression scheme, the run information of one unit of compression includes run-length information indicating the continuous number of the same pixel data blocks, or the number of pixels followed, and pixel data having a two-bit configuration for discriminating three or more colors of the pixels.

公开(公告)号：US6009202A

公开(公告)日：1999-12-28

申请号：US118975

申请日：1998-07-20

申请人： Shinichi Kikuchi ,  Tetsuya Kitamura ,  Hideki Mimura ,  Kazuhiko Taira

发明人： Shinichi Kikuchi ,  Tetsuya Kitamura ,  Hideki Mimura ,  Kazuhiko Taira

IPC分类号： G06T9/00 ,  G11B20/00 ,  G11B20/12 ,  H03M7/46 ,  H04N1/41 ,  H04N5/85 ,  H04N7/32 ,  H04N9/804 ,  H04N9/806 ,  H04N9/82 ,  H04N21/236 ,  H04N21/434 ,  G06K9/30

CPC分类号： H04N9/8042 ,  G11B20/00007 ,  H03M7/46 ,  H04N19/93 ,  H04N21/236 ,  H04N21/434 ,  H04N5/85 ,  H04N9/8063 ,  H04N9/8227

摘要： A bit train of a plurality of continuous pixels is compressed according to a new run-length compression scheme. In this run-length compression scheme, the run information of one unit of compression includes run-length information indicating the continuous number of the same pixel data blocks, or the number of pixels followed, and pixel data having a two-bit configuration for discriminating three or more colors of the pixels.

公开(公告)号：US5999111A

公开(公告)日：1999-12-07

申请号：US563018

申请日：1995-11-27

申请人： Juha Park ,  Byeungwoo Jeon ,  Jechang Jeong

发明人： Juha Park ,  Byeungwoo Jeon ,  Jechang Jeong

IPC分类号： H04N7/32 ,  H03M7/46 ,  H03M7/40

CPC分类号： H03M7/46

摘要： A variable length encoding/decoding apparatus using symbol-codeword re-association of a variable length code table includes a re-associator for storing symbol-address assignment information, address-codeword assignment information and address-counter-value assignment information, altering the symbol-address assignment information and the address-counter value assignment information according to the varied counter value due to the symbol identifying and outputting a variable length code table in which a symbol-codeword is re-associated to the encoder/decoder apparatus, and a controller for producing the initialized counter values based on probabilities of symbol occurrences belonging to the pre-defined variable length code table and the information stored in the re-associator based on the pre-defined updating unit, and storing the symbol-codeword re-associated variable length code table every predetermined updating unit in the encoder/decoder. A data compression efficiency can be improved even though global statistics corresponding to the pre-defined variable length code table and local statistics of the symbols/codewords which are actually variable-length-encoded/decoded do not match each other. Since it is not necessary to transmit the symbol-codeword re-association information at the encoding end to a decoding end, threre is no additional data to be transmitted by employing the present invention.

摘要翻译： 使用可变长度码表的符号 - 码字重新关联的可变长度编码/解码装置包括用于存储符号地址分配信息，地址码字分配信息和地址 - 计数器值分配信息的重新关联器，改变符号 由于符号识别而根据变化的计数器值地址分配信息和地址 - 计数器值分配信息，并输出其中符号码字被重新关联到编码器/解码器装置的可变长度码表，以及控制器 用于基于预定义的可变长度代码表的符号出现的概率和基于预定义的更新单元存储在重新关联器中的信息来产生初始化的计数器值，以及存储符号码字重新相关变量 编码器/解码器中的每个预定更新单元的长度代码表。 即使与预定义的可变长度代码表相对应的全局统计信息和实际上可变长度编码/解码的符号/码字的本地统计信息彼此不匹配，也可以提高数据压缩效率。 由于不需要将编码结束处的符号 - 码字重新关联信息发送到解码端，所以采用本发明不需要发送附加数据。

公开(公告)号：US5991340A

公开(公告)日：1999-11-23

申请号：US880410

申请日：1997-06-23

申请人： Masakazu Isomura

发明人： Masakazu Isomura

IPC分类号： G06T9/00 ,  H03M7/46 ,  H04B14/04

CPC分类号： G06T9/004 ,  H03M7/46

摘要： A data encoding method having a prediction setting process that designates a numerically superior symbol as either a "0" or a "1", and the other one is designated as a numerically inferior symbol. Then a binary bit string composed of "0"s and "1"s is input, and the numerically superior symbol is predicted to continuously repeat for n symbols, where n is set as a prediction bit number. A prediction result output process outputs a prediction correct signal that is either a "0" or a "1" when a prediction is correct for the observed series. The process then moves to an operation to perform encoding of a bit series containing the next n symbols, or otherwise outputs the other signal (i.e., that is not used to represent that the prediction is correct) as a prediction failure signal. A similar prediction setting process and prediction result output process is recursively repeated by setting a new prediction bit number that is smaller than the above prediction bit number by n symbols when the prediction fails for a predetermined number of times. Further, the same prediction setting process and prediction result output process are recursively repeated using a prediction bit number changed to a new prediction bit number increased by n symbols when the prediction is correct for a predetermined number of times. Similar algorithms are applied to perform decoding.

摘要翻译： 一种数据编码方法，具有将数字上乘的符号指定为“0”或“1”的预测设定处理，另一个被指定为数字下差的符号。 然后，输入由“0”和“1”组成的二进制位串，并且对n个符号预测数字上级符号连续重复，其中n被设置为预测位数。 预测结果输出处理输出对于观察到的序列的预测正确的“0”或“1”的预测校正信号。 然后，该处理移动到执行包含下一个n个符号的比特序列的编码，或以其他方式输出另一个信号（即，不用于表示预测是正确的）作为预测失败信号。 当预测失败预定次数时，通过设置小于上述预测比特数的新的预测比特数乘以n个符号来递归地重复类似的预测设置处理和预测结果输出处理。 此外，当预测正确预定次数时，使用改变为增加了n个符号的新预测比特数的预测比特数，递归地重复相同的预测设置处理和预测结果输出处理。 应用类似的算法进行解码。

公开(公告)号：US5835033A

公开(公告)日：1998-11-10

申请号：US553914

申请日：1995-11-06

申请人： Yoshinobu Mita

发明人： Yoshinobu Mita

IPC分类号： H03M7/46

CPC分类号： H03M7/46

摘要： A decoding apparatus includes a first decoder for performing decoding by dividing first coded data consisting of a bit sequence into divided data each having a bit length of 1 bit or more, and a second decoder which performs decoding on the basis of the "1" or "0" run length and the remaining bit string of second coded data consisting of a bit sequence, and shares at least a circuit portion with the first decoder.

摘要翻译： 解码装置包括：第一解码器，用于通过将由比特序列组成的第一编码数据分成具有1比特或更大的比特长度的分割数据来执行解码;以及第二解码器，其基于“1”或 “0”行程长度和由比特序列组成的第二编码数据的剩余比特串，并且与第一解码器共享至少一个电路部分。

公开(公告)号：US5818873A

公开(公告)日：1998-10-06

申请号：US924293

申请日：1992-08-03

申请人： Robert Lyle Wall ,  Kel D. Winters

发明人： Robert Lyle Wall ,  Kel D. Winters

IPC分类号： G06F5/00 ,  G06T9/00 ,  H03M7/30 ,  H03M7/46 ,  H04B14/04 ,  H04B1/66

CPC分类号： G06T9/005 ,  H03M7/3088

摘要： A single clock cycle adaptive data compressor/decompressor with a string reversal mechanism is described which can perform data compression and decompression at the rate of one uncompressed symbol per clock cycle. The LZW data compression algorithm has been improved for use in this data compressor. The compressor builds a string table as the data is received. Each string within the table is made up of the address within the table of the longest previously seen matching string and the one character that makes this string different. This data compressor/decompressor utilizes a content addressable memory to store the string table. This content addressable memory allows the compressor to store the current symbol string in a table while that same string table is simultaneously searched for the current string. During decompression the characters within a symbol string are output in reverse of the order in which they were input. In order for the data to be restored to its original format the characters within a string must be reversed before they are output. Two dual-ported random-access memories are used as circular queues to perform this string reversal. These dual-ported random-access memories have the capability to output the characters within a string from the string reversal mechanism in the order that they were input. The first dual-ported random access memory is used to store the strings of symbols to be reversed and the second dual-ported random access memory is used to store the addresses of the first and last symbol of each string that is stored in the first dual-ported random-access memory.

摘要翻译： 描述了具有字符串反转机制的单个时钟周期自适应数据压缩器/解压缩器，其可以以每个时钟周期的一个未压缩符号的速率执行数据压缩和解压缩。 LZW数据压缩算法已被改进，用于该数据压缩器。 压缩器在接收到数据时构建一个字符串表。 表中的每个字符串都由最早以前看到的匹配字符串的表中的地址和使此字符串不同的一个字符组成。 该数据压缩器/解压缩器利用内容可寻址存储器来存储字符串表。 该内容可寻址存储器允许压缩器将当前符号串存储在表中，同时搜索当前字符串。 在解压缩期间，符号串中的字符按照输入顺序的顺序输出。 为了使数据恢复到其原始格式，字符串中的字符在输出之前必须相反。 两个双端口随机存取存储器用作循环队列来执行此字符串反转。 这些双端口随机存取存储器具有从字符串反转机制输出串中的字符的能力，以便输入它们的顺序。 第一双端口随机存取存储器用于存储要反转的符号串，并且第二双端口随机存取存储器用于存储存储在第一双端口随机存取存储器的第一双端口随机存取存储器中的每个字符串的第一个和最后一个符号的地址。 - 运行的随机存取存储器。

公开(公告)号：US5731811A

公开(公告)日：1998-03-24

申请号：US847836

申请日：1997-04-28

申请人： Alphonsus A. J. De Lange ,  Gerard D. La Hei

发明人： Alphonsus A. J. De Lange ,  Gerard D. La Hei

IPC分类号： G06F12/00 ,  G06T1/60 ,  G06T3/00 ,  G06T9/00 ,  G09G5/00 ,  G09G5/12 ,  G09G5/14 ,  G09G5/39 ,  G09G5/42 ,  H03M7/46 ,  H04N1/387 ,  H04N5/073 ,  H04N5/262

CPC分类号： G09G5/39 ,  G09G5/14 ,  G09G2340/125 ,  G09G2360/123

摘要： An image processing system combines a multitude of image signals to create a compound image, The system has a plurality of memory modules operative to store the image signals as pixels for the compound image as respective segments of consecutive pixels. A specific row of consecutive pixels of the compound image is formed by consecutive arrangement of the respective segments. This permits the use of a simple addressing protocol, and of simple page-mode access DRAMs.

摘要翻译： 图像处理系统组合多个图像信号以创建复合图像。该系统具有多个存储器模块，其操作以将图像信号作为复合图像的像素存储为连续像素的相应片段。 复合图像的特定行的连续像素由相应的段的连续排列形成。 这允许使用简单的寻址协议和简单的页面模式存取DRAM。

公开(公告)号：US5710919A

公开(公告)日：1998-01-20

申请号：US536872

申请日：1995-09-29

申请人： Peter D. Rail

发明人： Peter D. Rail

IPC分类号： G06F17/30 ,  G06T9/00 ,  H03M7/30 ,  H03M7/46

CPC分类号： G06F17/30595 ,  G06T9/005 ,  H03M7/30 ,  H03M7/46 ,  Y10S707/99942

摘要： A system (10) for compressing and decompressing records includes a decompressed file (12) and a compressed file (14). A first record is stored in a first buffer (18) and a second record is stored in a second buffer (20). A vertical compression technique compares the first record and the second record, on a field-by-field basis, and replaces the field of the second record with a vertical repeat indicator if the field of the second record matches the corresponding field of the first record. A horizontal compression technique further compresses the second record by replacing adjacent matching characters with a horizontal repeat indicator.

摘要翻译： 用于压缩和解压缩记录的系统（10）包括解压文件（12）和压缩文件（14）。 第一记录存储在第一缓冲器（18）中，第二记录被存储在第二缓冲器（20）中。 垂直压缩技术在逐场的基础上比较第一记录和第二记录，并且如果第二记录的字段与第一记录的相应字段匹配，则用垂直重复指示符替换第二记录的字段 。 水平压缩技术通过用水平重复指示符替换相邻匹配字符来进一步压缩第二记录。

公开(公告)号：US5701125A

公开(公告)日：1997-12-23

申请号：US261225

申请日：1994-06-15

申请人： Gary J. Berlin

发明人： Gary J. Berlin

IPC分类号： G06T9/00 ,  H03M7/30 ,  H03M7/46

CPC分类号： G06T9/005 ,  H03M7/3086 ,  H03M7/46

摘要： A method used preferably with LZSS-based compression methods for compressing a stream of digital data. The method uses a run-length encoding scheme especially suited for data strings of identical data bytes having large run-lengths, such as data representing scanned images. The method reads an input data stream to determine the length of the data strings. Longer data strings are then encoded in one of two ways depending on the length of the string. For data strings having run-lengths less than 18 bytes, a cleared offset and the actual run-length are written to an output buffer and then a run byte is written to the output buffer. For data strings of 18 bytes or longer, a set offset and an encoded run-length are written to the output buffer and then a run byte is written to the output buffer. The encoded run-length is written in two parts obtained by dividing the run length by a factor of 255. The first of two parts of the encoded run-length is the quotient; the second part is the remainder. Data bytes that are not part of data strings of sufficient length are written directly to the output buffer.

摘要翻译： 优选使用基于LZSS的压缩方法来压缩数字数据流的方法。 该方法使用特别适用于具有较大游程长度的相同数据字节的数据串（例如表示扫描图像的数据）的游程长度编码方案。 该方法读取输入数据流以确定数据串的长度。 然后根据字符串的长度，以两种方式之一对较长的数据串进行编码。 对于长度小于18字节的数据串，清零偏移量和实际运行长度被写入输出缓冲区，然后运行字节被写入输出缓冲区。 对于18字节或更长的数据串，设置的偏移量和编码的游程长度被写入输出缓冲区，然后将一个运行字节写入输出缓冲区。 编码的游程长度被写入通过将游程长度除以系数255获得的两个部分。编码游程长度的两个部分中的第一部分是商; 第二部分是剩余部分。 不是足够长度的数据串的一部分的数据字节直接写入输出缓冲区。