摘要:
An image compression device having a prediction unit for predicting a value of a compressing target pixel from peripheral values, and quantizing prediction error between a prediction value of the prediction unit and a value of an actual pixel in a predetermined quantizing step, comprising: a specific pixel level difference evaluation unit sequentially calculating the difference value between the level values as a sum of the prediction values corresponding to each of the compressed pixel immediately before the compressing target pixel and peripheral compressed pixels adjacent to the pixel and the quantization value of the prediction error, and evaluating the sameness of the difference value; and a quantization switching unit controlling the quantization in the quantizing step more precise than the quantizing step used when the difference value continuously indicates the same value as a result of the evaluation by the specific pixel level difference evaluation unit.
摘要:
An image compression device having a prediction unit for predicting a value of a compressing target pixel from peripheral values, and quantizing prediction error between a prediction value of the prediction unit and a value of an actual pixel in a predetermined quantizing step, comprising: a specific pixel level difference evaluation unit sequentially calculating the difference value between the level values as a sum of the prediction values corresponding to each of the compressed pixel immediately before the compressing target pixel and peripheral compressed pixels adjacent to the pixel and the quantization value of the prediction error, and evaluating the sameness of the difference value; and a quantization switching unit controlling the quantization in the quantizing step more precise than the quantizing step used when the difference value continuously indicates the same value as a result of the evaluation by the specific pixel level difference evaluation unit.
摘要:
A quantizer quantizes a “prediction error” as a difference value between a pixel X to be compressed and a prediction value X′. A variable-length encoder inputs a quantization number obtained from a quantization table of the quantizer, and outputs a variable-length code (compressed code). A prediction value candidate calculation module calculates candidates 1 through k for a prediction value on the basis of peripheral pixels A, C, and B, and transmits them to a multiplexer (MPX). A prediction value determination module determines a prediction value among the candidates 1 through k for a prediction value on the basis of an intermediate value correspondence table and an intermediate value—image format correspondence table depending on an image format, and transmits a control signal to the multiplexer (MPX). The multiplexer (MPX) outputs one of the candidates 1 through k for a prediction value as a prediction value X′.
摘要:
A quantizer quantizes a “prediction error” as a difference value between a pixel X to be compressed and a prediction value X′. A variable-length encoder inputs a quantization number obtained from a quantization table of the quantizer, and outputs a variable-length code (compressed code). A prediction value candidate calculation module calculates candidates 1 through k for a prediction value on the basis of peripheral pixels A, C, and B, and transmits them to a multiplexer (MPX). A prediction value determination module determines a prediction value among the candidates 1 through k for a prediction value on the basis of an intermediate value correspondence table and an intermediate value—image format correspondence table depending on an image format, and transmits a control signal to the multiplexer (MPX). The multiplexer (MPX) outputs one of the candidates 1 through k for a prediction value as a prediction value X′.
摘要:
A data compressing method involves the use of a data string registering unit for registering a charactor or charactor string and the context and a code information registering unit for registering coding information of the character or character string following the context, the coding information following previous a input data. The data compressing method includes a data string new registering step of, when the input character or character string and the context set is not registering in the string registering unit, register the set in the data string registering unit. The same method also includes a code information new registering step of, when the input character or character string's code information following the context is not registering in the code information registering unit, register the code information in the code information registering unit and a code outputting step of outputting the code following the code information registered in the code information registering unit and a code information updating step of updating the code information registering unit.
摘要:
The present invention relates to a method of and an apparatus for compressing/decompressing data, to improve the compression ratio of sliding-dictionary-type Ziv-Lempel codes. A sliding-dictionary-type data compression method finds a matching character string in a coding character string and in a coded character string stored in a dictionary 101. According to the length of the matching character string, the method chooses a raw data mode that codes the matching character string that must consist of a single character as it is, or a reproduction mode that codes the position and length in the dictionary 101 of the matching character string. The method allocates consecutive numbers to different codes to be prepared under the raw data mode and reproduction mode. The method relates any code prepared under one of the raw data mode and reproduction mode to a corresponding one of the consecutive numbers, codes the number according to a code book, and rearranges the code book to shorten the length of the code of the number.
摘要:
An image data compression apparatus includes: a filter coefficient calculation unit to perform a filter process to separate pixels extracted in accordance with a one-dimensional subband coding into low- and high-frequency components, and to calculate coefficient values of respective filter coefficients; prediction units to perform different predictions respectively for the calculated coefficient values; a predicted coefficient value calculation unit to calculate predicted coefficient values to be used for posterior predictions by calculating each from at least one of predicted values and from corresponding at least one of the calculated coefficient values; a memory unit to hold the predicted coefficient values calculated by the predicted coefficient value calculation unit; and a compression coding unit to calculate a difference between each predicted value and corresponding each coefficient value, and to perform coding after quantizing the difference, or to perform coding on the basis of the difference.
摘要:
When a periodic pattern is not included in image data 210, a prediction error encoding unit 204 outputs the variable-length code of a prediction error quantization value 214 generated by prediction encoding as compressed data 215. When the periodic pattern is included in the image data 210, a periodic pattern encoding unit 208 outputs only information about the periodic pattern as the compressed data 215, while the output of the prediction error encoding unit is stopped. On the decompression side, while the copy instruction of a pixel value before one cycle from the compressed data is not detected, the decompression operation in a normal prediction encoding system is executed. When the instruction is detected, the pixel value one cycle before, internally stored is copied and outputted.
摘要:
A document conversion and use system which converts a first structured document into a second structured document different in structure from the first structured document, comprising: a template storage unit storing a template indicating the structure of the second structured document; a structure specification unit for reading out the template from the template storage unit and causing a user to specify an element or an element content constituting the template; a correspondence definition specification unit for causing the user to define the correspondence definition specification indicating the correspondence between the element in the first structured document and the element in the template; and an unit for analyzing the template structure to specify the element or the element content specified by the structure specification unit, searching for the element defined by the correspondence definition starting from the element or the element content, and performing conversion processing according to the search result.
摘要:
A structured data conversion device is designed to convert a set of first structured data into a set of second structured data that is described in a language different from that of the first structured data, and comprises: an inverse conversion information creation unit extracting data as inverse conversion information from the first structured data, the data that is extracted being deleted when the first structured data is converted into the second structured data and being necessary when the second structured data is inversely converted into the first structured data; and an inverse conversion information storing unit embedding in the second structured data the inverse conversion information when the first structured data is converted into the second structured data.