Low-resolution video coding content extraction from high resolution
A low complexity method for directly decoding low-resolution video frames from compressed high-resolution video content that were encoded using predictive coding toques, such as the H264 video coding standard, is disclosed. The decoding resolution is inversely proportional to the decoding complexity. Thus, a smaller decoding resolution yields higher computation and power savings. Low-frequency coefficients of 2 Dimensional transformed predictions are added to low-frequency coefficients of a transformed residual error. Low-frequency coefficients of the reconstructed data are then inversely transformed, having a smaller size transform. Further computational and power savings are obtained by reconstructing only those referenced pixels necessary for accurate decoding of proceeding Intra blocks.
1. A method for extracting a low-resolution video from a high-resolution compressed video, comprising:
inputting a video encoded with predictive coding techniques;
extracting a low-resolution Intra frame from a high-resolution compressed Intra frame that has been encoded using predictive coding techniques;
extracting a low-resolution predictive/bi-predictive frame from a high-resolution compressed predictive/bi-predictive frame that has been encoded using predictive coding techniques and will not be used as reference to decode any future frame by:
generating frequency coefficients of a prediction needed, wherein the number of frequency coefficients are no more than N/2×N/2 and a fraction of the number in high resolution compressed Intra frames being sampled;
generating corresponding frequency coefficients of a residual error; reconstructing only those pixels of a particular block that will be used as reference pixels for decoding future blocks;
adding the frequency coefficients of the prediction needed to corresponding frequency coefficients of the residual error;
taking the inverse transform of the frequency coefficients that are no more than N/2×N/2 and a fraction of the number called for in high resolution compressed Intra frames;
normalizing a resulting down-sampled image, wherein any reference DC coefficients for a next frame are generated by a weighted average of four reference DC coefficients so that the DC coefficient of the reference block of the next frame is within a region covered by the four reference DC coefficients when a down-sampled frame generated is to be used as a reference frame; and
outputting reconstructed frames as thumbnails.
2. The method of claim 1 , wherein said high-resolution compressed predictive/bi-predictive frame is taken immediately before an Intra/Intra Decoder Refresh (IDR) frame or another frame which is not used as a reference frame.
3. The method of claim 1 , further comprising:
adding low frequency coefficients of 2D transformed predictions to low frequency coefficients of transformed residual error;
wherein, low frequency coefficients of reconstructed data are inverse transformed by taking an inverse transform of smaller size.
4. The method of claim 1 , further comprising: reconstructing only those reference pixels required for an accurate decoding of further Intra blocks.
5. The method of claim 1 , further comprising sub-sampling of Inter blocks in a predictive coding video bitstream, by:
using an integer portion of a motion vector to point to a reference block selected in a reference frame;
disregarding the fractional part of the motion vector;
generating frequency coefficients needed from a selected reference block, wherein, the number of these frequency coefficients are no more than N/2×N/2;
dropping a fractional part in sub-sampling to avoid the otherwise considerable computational load;
generating corresponding frequency coefficients of a residual error;
adding frequency coefficients of a prediction needed to corresponding frequency coefficients of a reconstructed residual error;
taking an inverse transform of only the few frequency coefficients needed;
normalizing a resulting down-sampled block to form a down-sampled image.