Back to results
US 5,966,470 AGranted
Coding apparatus for image compression
AssigneeMatsushita Electric Industrial Co., Ltd.
InventorsMitsuhiro Miyashita, Tatsushi Bannai
Life of this patent
Estimating the term…
Abstract
A coding apparatus for image compression which includes a transformer that transforms image data into a plurality of transform coefficients arranged in a matrix; a quantizer that quantizes each of the transform coefficients into a coefficient data of a predetermined number of bits; a data divider that divides the coefficient data into a fixed-length data part and a variable-length data part and a variable-length encoder that encodes the variable-length data part into a variable-length code.
Claims10
A coding apparatus for image compression comprising:
- a transform means for transforming image data into a plurality of transform coefficients arranged in a matrix;
- a quantization means for quantizing each of said transform coefficients into quantized coefficient data of predetermined number of x bits;
- a data division means for dividing said quantized coefficient data into a first data bit part of y bits (y<x) which is to be fixed-length encoded and a second data bit part of (x-y) bits which is to be variable-length encoded, wherein the number y is predetermined in accordance with an order (i,j) of each of said quantized coefficient data (C.sub.ij);
- and
- a variable-length coding means for encoding said second data bit part into a variable-length code.
2 more elements
A coding apparatus for image compression comprising:
- a transform means for transforming image data into a plurality of transform coefficients arranged in a matrix;
- a quantization means for quantizing said transform coefficients into a respective plurality of coefficient data each having a predetermined number of bits;
- a data division means for dividing each of said quantized coefficient data into a first data bit part, which is to be fixed-length encoded, consisting of a sign bit and lower z bits of said coefficient data (z.gtoreq.0) and a second data bit part, which is to be variable-length encoded, consisting of the remaining bits of said coefficient data;
- a data conversion means for discarding the lowest bit of said second data bit part and then inverting each bit of the result if the lowest bit of said second data bit part is a logical 1 and, if the lowest bit of said second data bit part is a logical 0, discarding the lowest bit of said second data bit part;
- and
- a variable-length coding means for encoding the converted second data bit part into a variable-length code.
3 more elements
A coding apparatus for image compression comprising:
- a transform means for transforming image data into a plurality of transform coefficients arranged in a matrix;
- a quantization means for quantizing said transform coefficients into a respective plurality of coefficient data each having a predetermined number of bits x;
- an absolute-value conversion means for converting each of a predetermined subset of said plurality of coefficient data into a sign bit and (x-1) bits respectively denoting a sign and an absolute value of a corresponding coefficient data;
- a data division means for dividing each of the converted coefficient data into a first data bit part, which is to be fixed-length encoded, consisting of the sign bit and lower w bits (w.gtoreq.0) and a second data bit part, which is to be variable-length encoded, consisting of the remaining (x-w-1) bits;
- a data conversion means for discarding the lowest bit of said second data bit part and then inverting each bit of the result if the lowest bit of said second data bit part is a logical 1 and, if the lowest bit of said second data bit part is a logical 0, discarding the lowest bit of said second data bit part;
- and
- a variable-length coding means for encoding the converted second data bit part into a variable-length code.
4 more elements
A coding apparatus for image compression comprising:
- a transform means for transforming image data into a plurality of transform coefficients arranged in a matrix;
- a quantization means for quantizing said transform coefficients into a respective plurality of coefficient data each having a predetermined number of bits x;
- a data division means for dividing each of a predetermined subset of said plurality of coefficient data into a first data bit part, which is to be fixed-length encoded, consisting of a sign bit and lower v bits (v.gtoreq.0) and a second data bit part, which is to be variable-length encoded, consisting of the remaining (x-v-1) bits;
- an absolute-value conversion means for converting said second data bit part into its absolute value of (x-v-1) bits based on the sign bit;
- a data conversion means for discarding the lowest bit of said variable-length data part and then inverting each bit of the result if the lowest bit of said second data bit part is a logical 1 and, if the lowest bit of said second data bit part is a logical 0, discarding the lowest bit of said second data bit part;
- and
- a variable-length coding means for encoding the converted second data bit part into a variable-length code.
4 more elements
Dependencies are inferred from the claim text.