KR980006957A - Grammatical Adaptive Arithmetic Coder and / or Decoder Using Positional Probability Model and High Performance Image Coder and / or Decoder Using It - Google Patents

Abstract

The present invention relates to a grammar-based adaptive arithmetic encoder and / or decoder using a positional probability model, and a high-efficiency image encoder and / or decoder using the same. A discrete cosine transformer for performing discrete cosine transform of input data; A quantizer for quantizing a discrete cosine transform coefficient; An adaptive arithmetic coding unit for performing an adaptive arithmetic coding using a probability model based on the quantized discrete cosine transform coefficients according to the generation coefficient position and outputting the encoded bitstream; An adaptive arithmetic decoder for performing an adaptive arithmetic decoding using the probability model according to the generation coefficient position of the encoded bit stream and outputting the decoded bit stream; An inverse quantizer for inversely quantizing the decoded bit stream; And an inverse discrete cosine transformer for performing an inverse discrete cosine transform on the inversely quantized discrete cosine transform coefficients. Therefore, it is possible to perform compression closer to the individual probability of the inputted video signal, the decoder can be simplified in configuration because it does not need to have the general-purpose probability model, and the optimal probability model can be obtained by using the position- A probability model can be obtained and easily applied to H.263.

Description

Grammatical Adaptive Arithmetic Coder and / or Decoder Using Positional Probability Model and High Performance Image Coder and / or Decoder Using It

Since this is a trivial issue, I did not include the contents of the text.

FIG. 10 is a diagram illustrating an n-th coefficient process in an adaptive arithmetic coder using a predictive positional probability model according to the present invention;

FIG. 11 is a diagram illustrating a first coefficient process in an adaptive arithmetic decoder using a predictive positional probability model according to the present invention; FIG.

Claims (17)

A color inch generator for changing input discrete cosine transform coefficients to index values of last, level, and run; An encoder for generating position information using the index values output from the color inches generator and outputting position information, encoding input symbols using predetermined probability information, and supplying updated information about probability models of input symbols, ; And selecting the probability model corresponding to the calculated position from the first to fourth probability models using the position information output from the protection device, and selecting the probability information of the selected probability model from among the first to fourth probability information And a grammar-based multi-signal processor for applying the selected code to the code and updating the corresponding probability model using the updated information output to the encoder. A grammar-based adaptive arithmetic decoder for decoding an arithmetic-coded bitstream output from a grammar-based adaptive arithmetic coder as in claim 1. A discrete cosine transformer for performing discrete cosine transform on input data; A quantizer for quantizing the discrete cosine transform coefficients output from the discrete cosine transformer; And an adaptive arithmetic coder for performing an adaptive arithmetic coding using a probability model based on the quantized discrete cosine change coefficients output from the quantizer and generating an encoded bit stream. Efficient video encoder. The adaptive arithmetic coding apparatus as claimed in claim 3, wherein the adaptive arithmetic coding unit uses a distance from the current non-linear coefficient to the next non-linear coefficient or the last coefficient when encoding the quantized discrete cosine transform coefficient, . An adaptive arithmetic decoder for performing an adaptive arithmetic decoding using a probability model based on a generation coefficient position and outputting a decoded bitstream, the encoded bitstream output from the high efficiency image encoder as in claim 3; An inverse quantizer for inversely quantizing the decoded bit stream output from the adaptive arithmetic decoder; And an inverse discrete cosine transformer for inverse discrete cosine transforming the inversely quantized discrete cosine transform coefficients output from the inverse quantizer. A discrete cosine transformer for performing discrete cosine transform on input data; A quantizer for quantizing the discrete cosine transform coefficients output from the discrete cosine transformer; An adaptive arithmetic coder for performing an adaptive arithmetic coding using a probability model based on the quantized discrete cosine transform coefficients output from the quantizer, and outputting a coded bit stream; An adaptive arithmetic decoder for performing an adaptive arithmetic decoding using a probability model based on a generation coefficient position of the encoded bit stream output from the adaptive arithmetic encoder to output a decoded bit stream; Wherein the dequantization of the decoded bit stream output from the adaptive arithmetic decoder is performed by an inverse quantizer; And an inverse discrete cosine transformer for inverse discrete cosine transforming the inversely quantized discrete cosine transform coefficients output from the inverse quantizer. The method as claimed in any one of claims 3, 4, and 6, wherein the adaptive room arithmetic coding unit and the decoder use a predictive positional probability model for a position of a previous coefficient, And / or a decoder. 8. The method of claim 7, wherein, in encoding the quantized discrete cosine transform coefficient, a position in a previous non-linear coefficient is assumed to be a position of a current non-linear coefficient, and a protection is performed using a positional probability model corresponding to the position / RTI > and / or < RTI ID = 0.0 > a < / RTI > decoder. 7. The method of claim 6, wherein the adaptive arithmetic coding unit and the decoder use a distance from the current non-linear coefficient to the next non-linear coefficient or the last coefficient when encoding the quantized discrete cosine transform coefficient. Efficient Image Coder and Decoder. A discrete cosine transformer for performing discrete cosine transform of the input data; A quantizer for quantizing the discrete cosine transform coefficients output from the discrete cosine transformer; A fixed arithmetic coding unit for performing a fixed arithmetic coding using a probability model according to a generation coefficient position and outputting a coded bit stream, the quantized discrete cosine transform coefficient output from the quantizer; A fixed arithmetic decoder for performing a fixed arithmetic decoding using a probability model according to a generation coefficient position and outputting a decoded bitstream, the fixed bit stream output from the fixed arithmetic coding unit; An inverse quantizer for inversely quantizing the decoded bit stream output from the fixed arithmetic decoder; And an inverse discrete cosine transformer for inverse discrete cosine transforming the inversely quantized discrete cosine transform coefficients output from the inverse quantizer. 11. The high-efficiency image protector and decoder of claim 10, wherein the fixed arithmetic coding unit and the decoder use a predictive positional probability model for a previous coefficient position. 12. The method of claim 11, wherein, in encoding the quantized discrete cosine transform coefficient, encoding is performed using a positional probability model corresponding to a position of the current non-asymmetric coefficient, Characterized by high performance image coder and decoder. 11. The method of claim 10, wherein in the fixed arithmetic coding unit and the decoder, a distance from a current non-linear coefficient to a next non-linear coefficient or a last coefficient is used as a run in encoding the quantized discrete cosine transform coefficient. Image Coder and Decoder. A discrete cosine transformer for performing discrete cosine transform on input data; A quantizer for quantizing the discrete cosine transform coefficients output from the discrete cosine transformer; An entropy encoder for performing quantization of the quantized discrete cosine transform coefficients output from the quantizer by performing entropy encoding using a probability model based on the issue coefficient position and outputting the encoded bitstream; An entropy decoder for performing entropy decoding using the probability model based on the generation coefficient position and outputting the decoded bitstream; An inverse quantizer for inversely quantizing the decoded bit stream output from the entropy decoder; And an inverse discrete cosine transformer for transforming the inversely quantized discrete cosine transform coefficients output from the inverse quantizer. 15. The high-efficiency image encoder and decoder according to claim 14, wherein the entropy encoder and the decoder use a predictive positional probability model for a previous coefficient position. 16. The method of claim 15, wherein when the quantized discrete cosine transform coefficient coding is performed, a position in a previous non-image coefficient is assumed to be a position of a current non-image coefficient, and encoding is performed using a positional probability model corresponding to the position Characterized by high performance image coder and decoder. 15. The method of claim 14, wherein the entropy encoder and the decoder use a distance from the current non-linear coefficient to the next non-linear coefficient or the last coefficient when the quantized discrete cosine transform coefficient is encoded, Encoders and decoders.