Quantization method for use in image compression

Claim 2

A quantization apparatus for use in image compression, the apparatus comprising: means for receiving input pixels of N by N pixel blocks, means for transforming said pixel blocks to a frequency domain format resulting in a plurality of frequency transform coefficients including a DC coefficient and a plurality of AC coefficients, means for quantizing said frequency coefficients to form multi-bit integers, each integer including a most significant bit and including at least three decimal bits representative of a positive or negative fractional value, means for examining the first decimal bit and the second or third decimal bits of said integers, means for examining the most significant bits of said integers, means for adding a logical one to each of said integers only if said first decimal bit and said second or third decimal bits are a logical one and said most significant bit is a logical zero, and means for run length encoding of said integers.

Claim 3

The apparatus as in claim 1 wherein said means for quantizing include hardware means.

Claim 4

The apparatus as in claim 1 wherein said means for quantizing include software means.

Claim 5

The apparatus as in claim 1 wherein said fractional value is between the range of 0.625 and 0.750.

Claim 6

The apparatus as in claim 1 including means for adding a logical one to said integer if said fractional value is approximately 0.625.

Claim 7

The apparatus as in claim 4 wherein said pixel blocks include center pixels and corner pixels and wherein said means for quantizing include means for quantizing said center pixels.

Claim 8

The apparatus as in claim 6 including means for quantizing said corner pixels.

Claim 9

A quantization apparatus for use in image compression, the apparatus comprising: means for receiving input pixels of N by N pixel blocks; means for transforming said pixel blocks to a frequency domain format resulting in a plurality of frequency transform coefficients including a DC coefficient and a plurality of AC coefficients; means for quantizing said frequency coefficients by division by a quantization factor to form multi-bit numbers, each multi-bit number having a sign, an integer portion and positive fractional portion, the value of the multi-bit number being equal to said sign times the sum of said integer portion and said fractional portion; means for mapping each of said multi-bit numbers to a mapped integer, wherein said mapped integer is equal to zero if the corresponding multi-bit number is between a positive fractional quantity and a negative fractional quantity, wherein said positive fractional quantity is greater than 0.5 and said negative fractional quantity is less than -0.5, thereby increasing the number of multi-bit numbers mapped to zero; and means for run length compression encoding of said integers.

Claim 10

A quantization apparatus for use in image compression, the apparatus comprising: means for receiving input pixels of N by N pixel blocks; means for transforming said pixel blocks to a frequency domain format resulting in a plurality of frequency transform coefficients including a DC coefficient and a plurality of AC coefficients; means for quantizing said frequency coefficients by division by a quantization factor to form multi-bit numbers, each multi-bit number having a sign, an integer portion and positive fractional portion, the value of the multi-bit number being equal to said sign times the sum of said integer portion and said fractional portion; means for mapping each of said multi-bit numbers to a mapped integer, wherein said mapped integer is equal to zero if the corresponding multi-bit number is between a positive fractional quantity and a negative fractional quantity, wherein said positive fractional quantity is greater than 0.5 and said negative fractional quantity is less than -0.5, thereby increasing the number of multi-bit numbers mapped to zero; and means for run length compression encoding of said integers.

Claim 11

The quantization apparatus of claim 8 wherein said fractional portion is represented by at least two bits.

Claim 12

The quantization apparatus of claim 9 wherein said fractional portion is represented by at least three bits.

Claim 13

The quantization apparatus of claim 9 wherein said positive fractional quantity is approximately 0.750 and said negative fractional quantity is approximately -0.750.

Claim 14

The quantization apparatus of claim 10 wherein said positive fractional quantity is approximately 0.625 and said negative fractional quantity is approximately -0.625.

Claim 15

The quantization apparatus of claim 8 wherein said positive fractional quantity is equal to the negative of said negative fractional quantity.

Claim 16

The quantization apparatus of claim 13 further including means for mapping each of said multi-bit numbers having an integer portion of value less than a cutoff integer to a mapped integer, wherein said mapped integer has the sign and value of the sign and integer portion of the corresponding multi-bit number, respectively, if the fractional portion of the corresponding multi-bit number is between zero and said positive fractional quantity, and wherein said mapped integer has the sign of the corresponding multi-bit number and an integer value one greater than the integer portion of the corresponding multi-bit number if the fractional portion of the corresponding multi-bit number is between one and said positive fractional quantity.