Methods and apparatus for successive intra block prediction
Various embodiments of the present technology may comprise methods and apparatus for successive intra block prediction. Methods and apparatus for successive intra block prediction may comprise a matching decoder to generate data that replicates the internal state and/or the decompressed data at the decoder. The apparatus may further comprise a prediction module that utilizes the replicated data to make predictions. The apparatus may then utilize the predicted data and the original, input source data to determine a difference value and encode the difference value.
1. An apparatus for performing intra block prediction on a frame of image data, comprising:
a block generator configured to receive the frame of image data and divide the frame of image data into a plurality of blocks, wherein each block comprises a plurality of sub-blocks; and
an encoder connected to the block generator and configured to:
determine a first predicted pixel value for a first sub-block at a first location using only a previously-encoded sub-block;
replacing the first predicted value of the first sub-block with a replica value;
determine a second predicted pixel value for a second sub-block at a second location using the previously-encoded sub-block and the replica value;
wherein the first sub-block and the second sub-block are from a same block; and
compute a first difference encoded value using a first original, not-yet-encoded pixel value and the first predicted pixel value.
2. The apparatus according to claim 1 , wherein the first predicted pixel value and the first original, not-yet-encoded pixel value have a same coordinate location within the frame of image data.
3. The apparatus according to claim 1 , wherein the encoder comprises a matching decoder configured to perform dequantization and inverse transformation to generate the replica value.
4. The apparatus according to claim 1 , wherein the encoder further comprises:
a prediction module;
a transform module connected to an output terminal of the prediction module; and
a quantizer connected to an output terminal of the transform module; and
a matching decoder connected between an output terminal of the quantizer and an input terminal of the prediction module.
5. The apparatus according to claim 1 , wherein the previously-encoded sub-block is directly adjacent to the first location and the second location.
6. The apparatus according to claim 1 , wherein the encoder comprises:
a prediction module comprising:
a first input terminal to receive the plurality of blocks; and
a second input terminal;
a transform module connected to an output terminal of the prediction module, and
a quantizer connected to an output terminal of the transform module and the second input terminal of the prediction module, wherein the quantizer transmits a transform coefficient to the prediction module.
7. The apparatus according to claim 1 , wherein the previously-predicted sub-block comprises a previously-encoded value.
8. The apparatus according to claim 1 , wherein the encoder generates a compressed bitstream according to the first difference encoded value and transmits the compressed bitstream to a decoder.
9. The apparatus according to claim 1 , wherein the block generator transmits the original, not-yet encoded pixel value from an image sensor.
10. A method for performing intra block prediction, comprising:
generating a frame of image data comprising a plurality of pixel values;
generating a predicted block comprising:
generating a first predicted value for a first sub-block at a first location using only a previously predicted sub-block;
replacing the first predicted value of the first sub-block with a replica value; and
generating a second predicted value for a second sub-block at a second location using the previously-predicted sub-block and the replica value;
wherein the second sub-block is directly adjacent to the first sub-block; and
computing a difference value using the first predicted value and a first original, not-yet-encoded value from the plurality of pixel values.
11. The method according to claim 10 , wherein the first predicted pixel value and the first original, not-yet-encoded pixel value have a same coordinate location within the frame of image data.
12. The method according to claim 10 , wherein the previously-predicted sub-block comprises a previously-encoded value and is directly adjacent to the first and second locations.
13. The method according to claim 10 , further comprising:
generating a plurality of transformed coefficients by transforming the predicted block using a transform algorithm;
generating a plurality of quantized coefficient by quantizing the plurality of transformed coefficients;
encoding the plurality of quantized coefficients; and
outputting a compressed bitstream.
14. The method according to claim 10 , further comprising:
generating a dequantized value by dequantizing the plurality of quantized coefficients; and
generating an inverse transformed value by applying an inverse transform algorithm to the dequantized value.
15. A 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, wherein each block comprises a plurality of sub-blocks;
an encoder configured to:
generate a predicted block, comprising:
determining a first predicted pixel value for a first sub-block at a first location using only a first previously-predicted sub-block;
replacing the first predicted pixel value of the first sub-block with a first replica value; and
determining a second predicted pixel value for a second sub-block at a second location using the first previously-predicted sub-block and the first replica value; wherein the second sub-block is directly adjacent to the first sub-block; and
computing a difference value using the second predicted pixel value and an original, not-yet-encoded pixel value from the frame of image data; and
generate a compressed bitstream according to the predicted block;
a decoder configured to:
receive the compressed bitstream; and
reconstruct the frame of image data using the compressed bitstream, comprising:
determining a third predicted pixel value using only a second previously-predicted sub-block; and
determining a fourth predicted pixel value using the second previously-predicted sub-block and a decoded value; wherein the fourth predicted pixel value is directly adjacent to the third predicted pixel value.
16. The system according to claim 15 , wherein the encoder comprises a matching decoder configured to perform dequantization and inverse transformation to generate the first replica value.
17. The system according to claim 15 , wherein the encoder further comprises:
a prediction module;
a transform module connected to an output terminal of the prediction module; and
a quantizer connected to an output terminal of the transform module; and
a matching decoder connected between an output terminal of the quantizer and an input terminal of the prediction module.
18. The system according to claim 15 , wherein the first predicted pixel value and the first original, not-yet-encoded pixel value have a same coordinate location within the frame of image data.
19. The system according to claim 15 , wherein the first previously-predicted sub-block comprises a previously-encoded value and is directly adjacent to the first and second locations.
20. The system according to claim 15 , wherein the decoder further corrects the third predicted pixel value according to a transform coefficient.