Picture data encoding method

Details Picture data encoding method Picture data encoding method

1994-09-08

1996-11-26

Mancuso, Joseph

Image analysis

Image compression or coding

Transform coding

342420, 358432, 387246, 387250, G06K 936

Patent

active

055794130

DESCRIPTION:

BRIEF SUMMARY
The present invention relates to a method for encoding picture data, and more specifically encoding of transformed and quantized blocks of picture data, optionally groups of such blocks.
Lately digital compression techniques have been put into use in order to reduce the bit consumption in the representation of single pictures or picture sequences (live pictures). These techniques are used e.g. in picture telephony. The picture telephony encoding technique itself is standardized in the international telecommunication organization CCITT. The standardizing work was conducted by an expert group under CCITT SG XV WP XV/1, and the resulting recommendation has the number J.261. Also within the international standardizing organization ISO work has been done to provide standards for encoding both live pictures and single pictures (still pictures). The groups responsible for these activities have designations ISO/IEC JTC1/SC2/WG11 (for live pictures) and ISO/IEC JTC1/SC2/WG10 (for still pictures). The methods which are usually called MPEG and JPEG, are described in respectively ISO/IEC JTC1 Committee Draft 10918 and ISO/IEC JTC1 Committee Draft 11172.
A common feature for all these techniques is that blocks of picture values are processed. By applying a suitable transformation to the block, the result will be a block in the transform plane where the energy content is concentrated toward a block corner. Usually a discrete cosine transform (DCT) is used. All the three above mentioned methods use DCT. The transformed blocks are quantized, and the quantized values are thereafter encoded with a so-called "variable length code".
While the quantization introduces errors in the reconstructed signal (i.e. is irreversible), the variable length type encoding is a reversible process. This means that it is possible to reconstruct exactly that which has been encoded with a variable length code. The purpose of using encoding of the variable length type, is to reduce the data amount transmitted, without changing the information content. Events of small probability are encoded with many bits, while events of great probability are encoded using few bits. The so-called "Huffman encoding" is an example of encoding of the variable length type. It is a condition for improvement when the variable length coding is used, that different events occur with different probabilities. Thus, DCT transformed blocks of picture data are suitable for the variable length type encoding.
It is possible to use one or several input values to the variable length code. The case where several values are used can be called a multi-dimensional variable length code. In the coding technique which is standardized in CCITT, and in the encoding methods now presented for standardization within the ISO, two-dimensional variable length codes are used. The quantized transform coefficients in the transformed block are processed in a certain consecutive order. One input value is the level of the quantized transform coefficient, while the second value is the number of transform coefficients equal to zero since the last encoded (non-zero) coefficient. Thus, only non-zero coefficients are encoded in this case. (Encoding in this manner, i.e. only encoding values which are different from a given value, and indicating "travel length" in the block since the last encoded value, is termed "run length coding".) When the last non-zero coefficient in the block has been coded, a codeword is transmitted (the codeword is here termed EOB, "end of block") to signal that no more non-zero coefficients remain in the block. The use of EOB is suitable because it is often very many zero coefficients after the last non-zero coefficient in a block, unless the block is very small. The enclosed FIG. 1 shows the consecutive order for processing coefficients in a block, as it is conducted according to the CCITT standard algorithm for picture telephony.
Further, FIG. 2 shows a block containing quantized coefficients and the resulting events to be encoded when the CCITT standard algorithm is used. The t

REFERENCES:
patent: 4937573 (1990-06-01), Silvio et al.
patent: 5107345 (1992-04-01), Lee
patent: 5162795 (1992-11-01), Shirota

Affiliated with

Also associated with

Rating

Picture data encoding method does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Picture data encoding method, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Picture data encoding method will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-1979688