Method and apparatus for entropy encoding/decoding

1. A method of entropy-encoding an image, the method comprising:

arithmetically encoding a binary value of a current coding symbol, by using a probability of a predefined binary value which is determined based on previously-coded symbols that are encoded before the current coding symbol; and

updating the probability of the predefined binary value by using a plurality of scaling factors, according to the binary value of the current coding symbol,

wherein, when it is assumed that the plurality of scaling factors are ai (where ai is an real number less than 1, and i is an integer equal to or greater than 1), the probability of the predefined binary value determined based on information about the previously-coded symbols is P_old, a probability of a binary value updated by using the scaling factor ai is Pi_new, and y has a value of 0 or 1 depending on whether the current coding symbol has the predefined binary value,

the updating comprises updating the probability of the predefined binary value by using a weighted average value of probabilities Pi_new of a plurality of updated binary values obtained by using Equation; Pi_new=ai*y+(1−ai)*P_old.

2. The method of claim 1 , wherein the scaling factor ai is determined by using values of an exponentiation of 2.

3. The method of claim 2 , wherein ai=1/(2^Mi) (where Mi is an integer), and the updated probability Pi_new of the binary value is obtained by using Equation; Pi_new=(y>>Mi)+P_old−(P_old>>Mi).

4. The method of claim 1 , wherein, when a probability value is initialized, the updating by using the plurality of scaling factors is performed after a number of times the probability is updated exceeds a predetermined threshold.

5. The method of claim 4 , wherein, when the probability value is initialized, if the number of times the probability is updated is equal to or less than the predetermined threshold, the probability is updated by using one scaling factor.

6. An entropy encoding apparatus for entropy-encoding an image, the entropy encoding apparatus comprising:

a binary arithmetic coder for arithmetically encoding a binary value of a current coding symbol, by using a probability of a predefined binary value which is determined based on previously-coded symbols that are encoded before the current coding symbol; and

a context modeler for updating the probability of the predefined binary value by using a plurality of scaling factors, according to the binary value of the current coding symbol,

wherein, when it is assumed that the plurality of scaling factors are ai (where ai is an real number less than 1, and i is an integer equal to or greater than 1), the probability of the predefined binary value determined based on information about the previously-coded symbols is P_old, a probability of a binary value updated by using the scaling factor ai is Pi_new, and y has a value of 0 or 1 depending on whether the current coding symbol has the predefined binary value,

the context modeler updates the probability of the predefined binary value by using a weighted average value of probabilities Pi_new of a plurality of updated binary values obtained by using Equation; Pi_new=ai*y+(1−ai)*P_old.

7. A method of entropy-decoding an image, the method comprising:

arithmetically decoding a binary value of a current coding symbol, by using a probability of a predefined binary value which is determined based on previously-coded symbols that are decoded before the current coding symbol; and

updating the probability of the predefined binary value by using a plurality of scaling factors, according to the binary value of the current coding symbol,

wherein, when it is assumed that the plurality of scaling factors are ai (where ai is an real number less than 1, and i is an integer equal to or greater than 1), the probability of the predefined binary value determined based on information about the previously-coded symbols is P_old, a probability of a binary value updated by using the scaling factor ai is Pi_new, and y has a value of 0 or 1 depending on whether the current coding symbol has the predefined binary value,

the updating comprises updating the probability of the predefined binary value by using a weighted average value of probabilities P i , new of a plurality of updated binary values obtained by using Equation; P i _ new=a i *y+(1−a i )*P_old.

8. The method of claim 7 , wherein the scaling factor ai is determined by using values of an exponentiation of 2.

9. The method of claim 8 , wherein ai=1/(2^Mi) (where Mi is an integer), and the updated probability Pi_new of the binary value is obtained by using Equation; Pi_new=(y>>Mi)+P_old−(P_old>>Mi).

10. The method of claim 7 , wherein, when a probability value is initialized, the updating by using the plurality of scaling factors is performed after a number of times the probability is updated exceeds a predetermined threshold.

11. The method of claim 10 , wherein, when the probability value is initialized, if the number of times the probability is updated is equal to or less than the predetermined threshold, the probability is updated by using one scaling factor.

12. An entropy decoding apparatus for entropy-decoding an image, the entropy decoding apparatus comprising:

a binary arithmetic decoder for arithmetically decoding a binary value of a current coding symbol, by using a probability of a predefined binary value which is determined based on previously-coded symbols that are decoded before the current coding symbol; and

a context modeler for updating the probability of the predefined binary value by using a plurality of scaling factors, according to the binary value of the current coding symbol,

wherein, when it is assumed that the plurality of scaling factors are ai (where ai is an real number less than 1, and i is an integer equal to or greater than 1), the probability of the predefined binary value determined based on information about the previously-coded symbols is P_old, a probability of a binary value updated by using the scaling factor ai is Pi_new, and y has a value of 0 or 1 depending on whether the current coding symbol has the predefined binary value,

the context modeler updates the probability of the predefined binary value by using a weighted average value of probabilities P i _ new of a plurality of updated binary values obtained by using Equation; P i _ new=a i *y+(1−a i )*P_old.