Encoder, a decoder and corresponding methods using compact MV storage
The disclosure provides a motion vector compression method, comprising: obtaining a temporal motion vector; determining a compressed motion vector using a binary representation of the temporal motion vector comprising an exponent part and/or a mantissa part, wherein the exponent part comprises N bits, the mantissa part comprises M bits, and wherein N is a non-negative integer and M is a positive integer; and performing a temporal motion vector prediction (TMVP) using the compressed motion vector.
1. A motion vector compression method, comprising:
obtaining an 18 bit temporal motion vector;
determining a compressed motion vector using a binary representation of at least one of the 18 bit temporal motion vector, a binary representation of the compressed motion vector comprising an exponent part or a mantissa part, wherein the exponent part comprises N bits, the mantissa part comprises M bits, and wherein N is a non-negative integer and M is a positive integer, wherein the representation of the compressed motion vector is 10 bits; and
performing a temporal motion vector prediction (TMVP) using the compressed motion vector.
2. The motion vector compression method of claim 1 , wherein
the exponent part corresponds to most significant bit(s) (MSB) of the binary representation and the mantissa part corresponds to least significant bit(s) (LSB) of the binary representation; or,
the exponent part corresponds to LSB of the binary representation and the mantissa part corresponds to MSB of the binary representation.
3. The motion vector compression method of claim 2 , further comprising: when the exponent part corresponds to MSB of the binary representation and the mantissa part corresponds to LSB of the binary representation, deriving a value of the compressed motion vector by:
deriving a first shift value by applying a right shift of M bit to the binary representation;
deriving last M bit of the binary representation as a first basic binary representation; and
deriving the value of the compressed motion vector by applying a left shift of the first shift value to the first basic binary representation.
4. The motion vector compression method of claim 2 , further comprising: when the exponent part corresponds to LSB of the binary representation and the mantissa part corresponds to MSB of the binary representation, deriving a value of a motion vector component by:
deriving last N bit of the binary representation as a second shift value;
deriving a second basic binary representation by applying a right shift of N bit to the binary representation; and
deriving the value of the compressed motion vector by applying a left shift of the second shift value to the second basic binary representation.
5. The motion vector compression method of claim 1 , wherein the 18 bit temporal motion vector comprises a motion vector horizontal component and a motion vector vertical component.
6. A motion vector compression method, comprising:
coding a first indicator, wherein the first indicator is used to indicate whether the 18 bit temporal motion vector is compressed according to the motion vector compression method of claim 1 .
7. The motion vector compression method according to claim 1 comprising:
determining a value of N.
8. The motion vector compression method of claim 7 , wherein determining the value of N comprises:
coding the value of N; or
setting a predetermined value as the value of N; or
deriving the value of N based on a resolution of a picture unit, wherein the picture unit comprises a picture or a tile set; or
deriving the value of N based on a size of coding tree unit (CTU) or coding unit (CU).
9. The motion vector compression method of claim 8 , wherein deriving the value of N based on the resolution of the picture unit comprises:
setting the value of N as 0, when a width of the picture unit is smaller than a first threshold and a height of the picture unit is smaller than the first threshold; or,
coding a second indicator to represent the value of N, when the width of the picture unit is smaller than a second threshold and the height of the picture unit is smaller than the second threshold; or,
coding a third indicator to represent the value of N.
10. The motion vector compression method of claim 9 , wherein the second indicator is binarized by a bit, and the third indicator is binarized by two bits.
11. The motion vector compression method of claim 6 , wherein the first indicator, a second indicator or a third indicator is included in a sequence parameter set (SPS), a picture parameter set (PPS), a slice header, or a tile group header in a bitstream.
12. A non-transitory computer-readable storage medium storing executable instructions for execution by a processing circuitry, wherein the executable instructions, when executed by the processing circuitry, configures the processing circuitry to carry out the method according to claim 1 .
13. A decoder, comprising:
circuitry configured to perform the method of claim 1 .
14. A video data decoding device, comprising:
a non-transitory memory storage, configured to store video data in a form of a bitstream; and
a video decoder, configured to perform the method according to claim 1 .
15. The motion vector compression method of claim 1 , wherein the binary representation of the compressed motion vector is less than or equal to 16 bits.