Methods and apparatus for transform coefficient encoding and decoding
Various embodiments of the present technology may comprise methods and apparatus for transform coefficient encoding and decoding. Methods and apparatus for transform coefficient encoding and decoding may comprise an entropy encoder configured to encode each transform coefficient as a symbol, wherein the symbol comprises a context, a magnitude, and a mantissa. The context may be one of four contexts and the symbol may be encoded using a subset of Huffman codes, wherein the subset is determined based on the context.
1. An apparatus for coding transform coefficients, comprising:
a prediction module configured to generate a predicted block according to at least two blocks from a plurality of blocks;
a transform module connected to the prediction module and configured to generate a plurality of transform coefficients by transforming the predicted block using a transform algorithm;
a quantizer connected to the transform module and configured to generate a plurality of quantized coefficients by quantizing the plurality of transform coefficients; and
an entropy encoding module connected to the quantizer and configured to encode each quantized coefficient as a symbol, wherein the symbol comprises a context, a magnitude, and a mantissa.
2. The apparatus according to claim 1 , wherein the symbol further comprises a sign when the quantized coefficient is a non-zero value.
3. The apparatus according to claim 1 , wherein the transform algorithm is a fast Walsh-Hadamard transform algorithm.
4. The apparatus according to claim 1 , wherein the entropy encoding module comprises a look-up table of Huffman codes.
5. The apparatus according to claim 4 , wherein the symbol is encoded using a subset of Huffman codes from the look-up table of Huffman codes, wherein the subset of Huffman codes is selected based on the context.
6. The apparatus according to claim 1 , wherein the entropy encoding module encodes the plurality of quantized coefficients using a variable-length code table.
7. The apparatus according to claim 1 , wherein the entropy encoding module encodes the mantissa separately from the context and the magnitude.
8. The apparatus according to claim 1 , wherein the context is one of:
a first context, wherein the first context is a first coefficient from the plurality of quantized coefficients;
a second context, wherein the second context is a last coefficient from the plurality of quantized coefficients with a magnitude of zero;
a third context, wherein the third context is a last coefficient from the plurality of quantized coefficients with a magnitude of one; or
a fourth context, wherein the fourth context is a default context.
9. A method for coding transform coefficients, comprising:
generating a frame of image data;
dividing the frame of image data into a plurality of blocks;
generating a predicted block using at least two blocks from the plurality of blocks;
generating a transform coefficient by transforming the predicted block using a transform algorithm;
generating a quantized coefficient by quantizing the transform coefficient; and
within one clock cycle, encoding the quantized coefficient as a symbol, wherein the symbol comprises a context, a magnitude, and a mantissa.
10. The method according to claim 9 , wherein the symbol further comprises a sign when the quantized coefficient is a non-zero value.
11. The method according to claim 9 , wherein the context is one of:
a first context, wherein the first context is a first coefficient from the plurality of quantized coefficients in the predicted block;
a second context, wherein the second context is a last coefficient from the plurality of quantized coefficients with a magnitude of zero;
a third context, wherein the third context is a last coefficient from the plurality of quantized coefficients with a magnitude of one; or
a fourth context, wherein the fourth context is a default context.
12. The method according to claim 9 , wherein encoding the quantized coefficient comprises: encoding the quantized coefficient using a variable-length code table.
13. The method according to claim 9 , wherein encoding the quantized coefficient comprises: encoding the symbol using a subset of Huffman codes from a table of Huffman codes, wherein the subset of Huffman codes is selected based on the context.
14. The method according to claim 9 , wherein encoding the quantized coefficient comprises encoding the mantissa separately from the context and the magnitude.
15. An imaging system, comprising:
an image sensor configured to generate a frame of image data;
a block generator connected to the image sensor and configured to receive the frame of image data and divide the frame of image data into a plurality of blocks; and
an encoder comprising:
a prediction module connected to the block generator and configured to generate a predicted block according to at least two blocks from the plurality of blocks;
a transform module connected to the prediction module and configured to generate a transform coefficient by transforming the predicted block using a transform algorithm;
a quantizer connected to the transform module and configured to generate a quantized coefficient by quantizing the transform coefficient; and
an entropy encoding module connected to the quantizer and configured to encode the quantized coefficient as a symbol, wherein the symbol comprises a context, a magnitude, and a mantissa.
16. The imaging system according to claim 15 , wherein the symbol further comprises a sign when the quantized coefficient is a non-zero value.
17. The imaging system according to claim 15 , wherein the context is one of:
a first context, wherein the first context is a first coefficient from the plurality of quantized coefficients in the predicted block;
a second context, wherein the second context is a last coefficient from the plurality of quantized coefficients with a magnitude of zero;
a third context, wherein the third context is a last coefficient from the plurality of quantized coefficients with a magnitude of one; or
a fourth context, wherein the fourth context is a default context.
18. The imaging system according to claim 15 , wherein the entropy encoding module encodes the mantissa separately from the context and the magnitude.
19. The imaging system according to claim 15 , wherein the entropy encoding module comprises a look-up table of Huffman codes.
20. The imaging system according to claim 19 , wherein the symbol is encoded using a subset of Huffman codes from the look-up table of Huffman codes, wherein the subset of Huffman codes is selected based on the context.