Method and apparatus for encoding a motion vector

• US 6,020,933 A
• Filed: 10/27/1997
• Issued: 02/01/2000
• Est. Priority Date: 09/30/1997
• Status: Expired due to Term

• Current Assignee: Wi-LAN Inc.
• Original Assignee: Daewoo Electronics Company Limited (Dayou Holdings Co.,Ltd.)
• Inventors: Lee, Sang-Hoon
• Primary Examiner(s): BRITTON, HOWARD W
• Application Number:
  US08/958,652
• Time in Patent Office: 827 Days
• Field of Search: 348/699, 348/402, 348/407, 348/413, 348/416
• US Class Current: 348/699
• CPC Class Codes: H04N 19/51 Motion estimation or motion compensation
• CPC Class Codes: H04N 5/145 Movement estimation for video coding H04N19/51

Abstract

A current motion vector CMV of a search block is encoded based on a plurality of reference motion vectors RMV'"'"'s. A first and a second candidate predictors FCP and SCP having first components FCP_-- 1 and SCP_-- 1 and second components FCP_-- 2 and SCP_-- 2 are determined. The FCP_-- 1 and the FCP_-- 2 represent medians of the first and the second components RMV_-- 1'"'"'s and RMV_-- 2'"'"'s of the RMV'"'"'s, respectively, while the SCP_-- 1 and the SCP_-- 2 represent a RMV_-- 1 and a RMV_-- 2 which yield a least first and a second absolute differences with respect to a first and a second components CMV_-- 1 and CMV_-- 2 of the CMV, respectively. A dispersion value of the RMV'"'"'s is compared with a predetermined threshold to thereby generate a first selection signal if the dispersion value is less than the threshold and a second selection signal if otherwise. A third candidate predictor TCP having a first and a second components TCP_-- 1 and TCP_-- 2 is determined based on the number of encoded bits generated by encoding the CMV_-- 1 with respect to each of the FCP_-- 1 and the SCP_-- 1 and the number of encoded bits generated by encoding the CMV_-- 2 with respect to each of the FCP_-- 2 and the SCP_-- 2, wherein the TCP_-- 1 is one of the FCP_-- 1 and the SCP_-- 1 and the TCP_-- 2 is either the FCP_-- 2 or the SCP_-- 2. The CMV is encoded based on the FCP and the TCP in response to the first and the second selection signals, respectively, thereby generating encoded motion vector data for the CMV.

Claims

1. A method for encoding a current motion vector CMV of a search block based on a plurality of reference motion vectors RMV'"'"'s, wherein the reference motion vectors represent motion vectors of reference blocks of the search block and each motion vector includes a first and a second components, the method comprising the steps of:

(a) finding a first candidate predictor FCP having a first and a second components FCP_-- 1 and FCP_-- 2, a FCP_-- j representing a median of j_th components RMV_-- j'"'"'s of the RMV'"'"'s, j being 1 and 2;

(b) calculating an absolute difference j between a j_th component CMV_-- j of the CMV and each of RMV_-- j'"'"'s;

(c) determining a second candidate predictor SCP having a first component SCP_-- 1 and a second component SCP_-- 2, a SCP_-- j representing a RMV_-- j which yields a least absolute difference j;

(d) computing a dispersion value of the RMV'"'"'s and comparing the dispersion value with a predetermined threshold to thereby generate a first selection signal if the dispersion value is less than the threshold and a second selection signal if otherwise, the dispersion value indicating whether the RMV'"'"'s are closely related to each other or not;

(e) providing a third candidate predictor TCP having a first component TCP_-- 1 and a second component TCP_-- 2 based on the number of encoded bits generated by encoding the CMV_-- j with respect to each of the FCP_-- j and the SCP_-- j, wherein a TCP_-- j is one of the FCP_-- j and the SCP_-- j; and

(f) encoding the current motion vector CMV based on the first and the third candidate predictor FCP and TCP in response to the first and the second selection signals, respectively, thereby generating encoded motion vector data for the CMV.

2. The method according to claim 1, wherein the dispersion value is defined as:

##EQU2## wherein DIS represents the dispersion value; and

RMV(i)_-- 1 and RMV(i)_-- 2 denote a first and a second components of an i_th reference motion vector RMV(i), respectively, i being 1 to N with N being the number of the reference motion vectors.

3. The method according to claim 2, wherein the providing step (e) includes the steps of:

(e1) comparing the FCP_-- j with the SCP_-- j;

(e2) if the FCP_-- j and the SCP_-- j are identical, selecting either the FCP_-- j or the SCP_-- j as the TCP_-- j;

(e3) if the FCP_-- j and the SCP_-- j are not identical, encoding the CMV_-- j with respect to the FCP_-- j and the SCP_-- j to thereby generate first and second encoded data, respectively, and the number of bits of each of the first and the second encoded data;

(e4) comparing the number of bits of the first encoded data with the one of the second encoded data; and

(e5) choosing the FCP_-- j as the TCP_-- j if the number of encoded bits of the first encoded data is smaller than the one of the second encoded data and choosing the SCP_-- j as the TCP_-- j if otherwise.

4. The method according to claim 3, wherein the step (e3) has the steps of:

(e31) if the FCP_-- j and the SCP_-- j are not identical, calculating the difference between the CMV_-- j and the FCP_-- j;

(e32) encoding the difference in step (e31) based on a predetermined statistical coding technique and providing a statistically coded flag signal representing the FCP_-- j to thereby generate the first encoded data; and

(e33) counting the number of bits of the first encoded data.

5. The method according to claim 4, wherein the step (e3) further has the steps of:

(e34) calculating the difference between the CMV_-- j and the SCP_-- j;

(e35) encoding the difference in step (e34) based on the predetermined statistical coding technique and providing a statistically coded identification signal representing a reference block corresponding to the SCP_-- j to thereby generate the second coded data; and

(e36) counting the number of bits of the second encoded data.

6. The method according to claim 5, wherein the encoding step (f) includes the steps of:

(f1) generating, based on the predetermined statistical coding technique, a first and a second flag signals representing reference blocks corresponding to the TCP_-- 1 and the TCP_-- 2, respectively;

(f2) selecting, as an optimum predictor OP of the CMV, the FCP or the TCP in response to the first or the second selection signal, the OP having a first component OP_-- 1 and a second component OP_-- 2;

(f3) encoding the CMV with respect to the optimum predictor OP to thereby generate coded motion vector data for the CMV; and

(f4) generating, as the encoded motion vector data, the coded motion vector data in response to the first selection signal or the coded motion vector data together with the first and the second flag signals in response to the second selection signal.

7. The method according to claim 6, wherein N=3.

8. The method according to claim 7, wherein the generating step (f1) has steps of:

(f11) comparing the FCP_-- 1 and the FCP_-- 2 with the TCP_-- 1 and the TCP_-- 2, respectively;

(f12) generating the first and the second flag signals based on the comparison results obtained in step (f11).

9. The method according to claim 8, wherein the first and the second flag signals represent whether the TCP_-- 1 and the TCP_-- 2 are smaller than, equal to, or greater than the FCP_-- 1 and the FCP_-- 2, respectively.

10. The method according to claim 9, wherein the encoding step (f3) has the steps of:

(f31) calculating a first difference between the CMV_-- 1 and the OP_-- 1 and a second difference between the CMV_-- 2 and the OP_-- 2; and

(f32) encoding the first and the second differences in step (f31) based on the predetermined statistical coding technique to thereby generate the coded motion vector data.

11. The method according to claim 10, wherein the predetermined statistical coding technique is a variable length coding technique.

12. An apparatus for encoding a current motion vector CMV based on a plurality of reference motion vectors RMV'"'"'s, each of the motion vectors including a first and a second components, comprising:

means for computing a dispersion value of the first and the second components RMV_-- 1'"'"'s and RMV_-- 2'"'"'s of the RMV'"'"'s to thereby issue a first selection signal if the dispersion value is considered to be low and a second selection signal if otherwise;

means for determining a first predictor FCP having a first and a second components FP_-- 1 and FP_-- 2, wherein a FP_-- j represents a median value of RMV_-- j'"'"'s, j being 1 and 2;

means for obtaining a second predictor SP having a first and a second components SP_-- 1 and SP_-- 2, wherein a SP_-- j corresponds to one of the RMV_-- j'"'"'s which yields a minimum difference with respect to a j_th component CMV_-- j of the CMV;

means for evaluating the number of encoded bits generated when encoding the CMV_-- j with respect to each of the FP_-- j and the SP_-- j and to thereby generate a third predictor TP having a first and a second components TP_-- 1 and TP_-- 2, wherein the TP_-- j represents one of the FP_-- j and the SP_-- j, the number of encoded bits corresponding to said one of the FP_-- j and the SP_-- j being not greater than the one corresponding to the other one of the FP_-- j and the SP_-- j;

means for selecting the FP or the TP as an optimum predictor OP in response to the first or second selection signal; and

means for encoding the CMV with respect to the OP to thereby generate encoded motion vector data.

13. The apparatus according to claim 12, wherein said computing means include means for computing the dispersion value defined as:

##EQU3## wherein DIS represents the dispersion value; and

RMV(i)_-- 1 and RMV(i)_-- 2 denote RMV_-- 1 and a RMV_-- 2 of an i_th reference motion vector RMV(i), respectively, i being 1 to N with N being the number of the RMV'"'"'s.

14. The apparatus according to claim 13, wherein said evaluating means includes:

means for comparing the FP_-- j and with the SP_-- j, to thereby generate either FP_-- j or the SP_-- j as the TP_-- j if the FP_-- j is equal to the SP_-- j and generate both the FP_-- j and the SP_-- j if otherwise;

means, in response to the FP_-- j and the SP_-- j from said comparing means and the CMV_-- j, for obtaining the number of encoded bits generated when encoding the CMV_-- j with respect to each of the FP_-- j and the SP_-- j; and

means for choosing one of the FP_-- j and the SP_-- j as the TP_-- j, the number of encoded bits corresponding to the chosen one of the FP_-- j and the SP_-- j being not greater than the one corresponding to the unchosen one.

15. The apparatus according to claim 14, wherein said means for obtaining the number of encoded bits includes:

a first bit counting means, responsive to the CMV_-- j and the FP_-- j, for providing the first number of encoded bits corresponding to FP_-- j; and

a second bit counting means, responsive to the CMV_-- j, the FP_-- j and the SP_-- j, for calculating the second number of encoded bits corresponding to SP_-- j.

16. The apparatus according to claim 15, wherein said first bit counting means includes:

means for computing a first difference between the CMV_-- j and the FP_-- j;

means for encoding the first difference based on a predetermined statistical coding technique and providing a statistically coded flag signal representing the FP_-- j to thereby generate first encoded data; and

means for counting the number of bits of the first encoded data to thereby generate the first number of encoded bits.

17. The apparatus according to claim 16, wherein N=3.

18. The apparatus according to claim 17, wherein said second bits counting means includes:

means for computing a second difference between the CMV_-- j and the SP_-- j and providing an identification signal by comparing the FP_-- j with the SP_-- j to thereby generate second encoded data, the identification signal representing whether the SP_-- j is greater or smaller than the FP_-- j; and

means for counting the number of bits of the second encoded data to thereby generate the second number of encoded bits.

19. The apparatus according to claim 18, wherein said choosing means includes:

means for comparing the first number of encoded bits with the second number of encoded bits to thereby generate a switch control signal representing whether the first number of encoded bits is smaller than the second number of encoded bits or not; and

means for selecting, as the TP_-- j, the FP_-- j if the switch control signal represents that the first number of encoded bits is smaller than the second number of encoded bits, and the SP_-- j if otherwise.

20. The apparatus according to claim 19, wherein said encoded motion vector data further includes a flag signal representing whether the TP_-- j is smaller than, equal to, or greater than the FP_-- j in case the TP is selected as the OP.