Video decoder with bit stream based enhancements
First Claim
1. A method for enhanced video reconstruction, comprising:
- receiving encoded digital video bitstream data into a bitstream buffer, the bitstream buffer comprising a plurality of buffer portions;
receiving buffered encoded digital video bitstream data into a Bitstream Parser-Processor (BPP);
determining from the buffered encoded digital video bitstream data an initial state of a plurality of bitstream data components within the buffered encoded digital video bitstream data;
deriving a second state of the plurality of bitstream data components, the deriving comprising the BPP interacting with at least one of the plurality of buffer portions, a high-layer state storage element and at least one of a plurality of enhanced bitstream data analysis and decision elements;
receiving the second state of the plurality of bitstream data components and the buffered encoded bitstream data into a block decoder element of the plurality of enhanced bitstream data analysis and decision elements;
decoding the buffered encoded bitstream data employing the received plurality of bitstream data components;
transferring the decoded bitstream data to a frame buffer as a first video frame;
deriving a third state of the plurality of bitstream data components using the plurality of enhanced bitstream data analysis and decision elements; and
employing the third state of the plurality of bitstream data components for reconstructing the first video frame into a second video frame.
4 Assignments
0 Petitions
Accused Products
Abstract
An advanced decoder provides enhancement processing in accordance with information previously known only to the decoder and/or encoder. Preferably, advanced digital video image enhancement utilizing bitstream information is provided in an integrated manner with advanced decoding and, more preferably, provides for MPEG decoding and image enhancement. In addition to advanced edge detection and other advanced processing tools, the invention further enables the use of decoder based coding-type operations, such as reverse superresolution, as well as advanced superresolution, and 3-D display processing techniques in conjunction with video processing. The invention is therefore capable of reducing system complexity and/or bandwidth, requirements, and improving system robustness and/or the ultimate quality of the enhanced video signal, among other advantages.
-
Citations
20 Claims
-
1. A method for enhanced video reconstruction, comprising:
-
receiving encoded digital video bitstream data into a bitstream buffer, the bitstream buffer comprising a plurality of buffer portions;
receiving buffered encoded digital video bitstream data into a Bitstream Parser-Processor (BPP);
determining from the buffered encoded digital video bitstream data an initial state of a plurality of bitstream data components within the buffered encoded digital video bitstream data;
deriving a second state of the plurality of bitstream data components, the deriving comprising the BPP interacting with at least one of the plurality of buffer portions, a high-layer state storage element and at least one of a plurality of enhanced bitstream data analysis and decision elements;
receiving the second state of the plurality of bitstream data components and the buffered encoded bitstream data into a block decoder element of the plurality of enhanced bitstream data analysis and decision elements;
decoding the buffered encoded bitstream data employing the received plurality of bitstream data components;
transferring the decoded bitstream data to a frame buffer as a first video frame;
deriving a third state of the plurality of bitstream data components using the plurality of enhanced bitstream data analysis and decision elements; and
employing the third state of the plurality of bitstream data components for reconstructing the first video frame into a second video frame. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20)
using original bitstream vectors and motion mode and coefficient count data outputted from the BPP with first video frame data from the block decoder to produce temporal filter vectors and optical flow metrics; and
determining gradient measurements from temporal filter vectors and optical flow metrics.
-
-
11. The method of claim 9, where deriving third state data comprises determining motion vectors from original bitstream vectors, the motion vectors having a granularity that is finer than that of the original bitstream vectors.
-
12. The method of claim 9, where deriving third state data comprises performing motion analysis using hierarchical motion estimation.
-
13. The method of claim 1, where the plurality of enhanced bitstream data analysis and decision elements comprises a spatial analysis element, the spatial analysis element further comprising a mask generator sub-element, a gradient measurer sub-element and an edge detector sub-element, the spatial analysis element analyzing bitstream data from the BPP and first frame data from the block decoder and/or frame buffer to derive third state data.
-
14. The method of claim 13, where deriving third state data comprises:
-
using the mask generator sub-element to create occlusion labels;
using the gradient measurer sub-element to measure gradients of surrounding pixels;
using the occlusion labels and measured gradients for edge detection; and
deriving spatial-domain analysis data.
-
-
15. The method of claim 13, where deriving third state data comprises:
-
performing gradient measurements of motion vector and or spatial vector data; and
performing hierarchical motion estimation temporal processing.
-
-
16. The method of claim 1, where the plurality of enhanced bitstream data analysis and decision elements comprises a decision unit element, the decision unit element further comprising a parameter control processor sub-element, a quantization Noise estimator sub-element, a residual measurer sub-element an enhancement processor sub-element and a prior frame iteration storage sub-element, the decision unit element analyzing data received from the BPP, the block decoder and/or frame buffer, the spatial analysis element and the motion analysis element to derive data for rendering.
-
17. The method of claim 16, where deriving data for rendering comprises:
-
receiving DCT data from the BPP;
receiving vector field discontinuity data the motion analyzer element;
receiving gradient detection from the spatial analyzer element; and
performing stream vector prediction, enhanced prediction and generating final edge metrics.
-
-
18. The method of claim 16, where deriving data for rendering comprises merging an inverse point spread function with optical flow measurements.
-
19. The method of claim 1, where deriving a third state of the plurality of bitstream data components further comprises utilizing 3-D methods for warping new image areas to create interpolated video frame data.
-
20. The method of claim 1, where deriving a third state of the plurality of bitstream data components further comprises:
processing intertrame dependence between adjacent groups of pictures (GOP) data blocks, where such processing includes interframe filtering; and
applying meta data multiplexed within a plurality of adjacent GOP'"'"'s for processing each of the GOP'"'"'s.
Specification