Pulse or digital communications – Bandwidth reduction or expansion – Television or motion video signal
Reexamination Certificate
2000-06-30
2003-03-25
Le, Vu (Department: 2613)
Pulse or digital communications
Bandwidth reduction or expansion
Television or motion video signal
Reexamination Certificate
active
06539060
ABSTRACT:
TECHNICAL FIELD
The present invention relates to image data processing, and more particularly, to an image data post-processing method for reducing quantization effect, and an apparatus therefor.
BACKGROUND ART
The present invention is adopted as ISO/IEC JTC1/SC29/WG11 N1902 (ISO/IEC 14496-2 Committee Draft).
Generally, picture encoding standards such as MPEG of the International Standardization Organization (ISO) and H.263 recommended by the International Telecommunication Union (ITU) adopt block-based motion estimation and discrete cosine transform (DCT) of blocks. Also, most video coding standards use an 8×8 pixel block DCT for packing information into a few transform coefficients. Such block-based DCT schemes take advantage of the local spatial correlation property of images.
However, when image data which has been coded based on blocks is restored, the restored image is considerably deteriorated, causing blocking artifacts near block boundaries, corner outliers at cross points of blocks, and ringing noise near the image edge. Because MPEG quantizes the transformed coefficient of 8×8 pixel block. In particular, when the image is highly compressed, the deterioration of the image becomes serious.
When an image is highly compressed, the block-based coding induces the well-known blocking artifacts near the block boundary, corner outliers at the cross points of the blocks, and ringing noise near the image edge.
The blocking artifacts are grid noise occurring along the block boundaries in a relatively homogeneous area. The grid noise shows traces of the block-based process at the edges between blocks when the compressed data is displayed on a screen after being restored. Thus, the edges between blocks are identified. Also, corner outliers occur at the corner points of the 8×8 block. Also, the ringing noise is a typical Gibb's phenomenon occurring by truncation when high-frequency coefficients of the DCT are quantized so as to highly compress the image. As a result, overlapping of image with a predetermined interval due to the ringing noise is noticeable.
Several methods for reducing the blocking artifacts, corner outliers and ringing noise, caused by block-based coding, have been suggested in the following articles: [1] Lee, H. C. Kim and H. W. Park, “
Blocking Effect Reduction of JPEG images by Signal Adaptive Filtering
”, in press IEEE Trans. on Image processing, 1997, [2] B. Ramanurthi and A. Gersho, “
Nonlinear Space Variant Postprocessing of Block Coded Images
”, IEEE Trans. on ASSP, Vol. 34, No. 5, pp. 1258-1267, 1986, [3] Y. Ynag, N. Galatsanos and A. Katsaggelos, “
Projection-Based Spatially Adaptive Reconstruction of Block-Transform Compressed Images
”, IEEE Trans. on Image processing, Vol. 4, No. 7, pp. 896-908, Jul. 1995, [4] Z. Xiong, M. T. Orchard, and Y. Q. Zhang, “
A Deblocking Algorithum for JPEG Compressed Images Using Overcomplete Wavelet Representations
”, IEEE Trans. Circuits Syst. Video Technol., Vol. 7, No. 2, pp. 433-437, 1997.
In the reference [1], 2-dimensional signal-adaptive filtering (SAF) for reducing quantization effect of a JPEG-decompressed image has been suggested. Also, in the reference [2], a 2-dimensional filter is used to reduce the blocking artifacts and a 1-imensional filter is used to reduce staircase noise, resulting in good effects. In the reference [3], an iterative image-recovery algorithm using the theory of projections onto convex sets (POCS) has also been proposed.
However, the main drawback of these algorithms is their computation complexity. Meanwhile, post-filtering which uses over-complete wavelet representation which reduces the complexity of computation has been suggested by the reference [4].
However, the post-filtering method is applied only to JPEG-decompressed images. Also, for low bit-rate coding, a spatio-temporal adaptive post-filtering which can be applied to a 3-dimensional subband coding has been suggested in the reference [5]: T. S. Liu and N. Jayant, “
Adaptive Postprocessing Algorithms for Low Bit Rate Video Signals
”, IEEE Trans. on Image Processing, Vol. 4, No. 7, pp. 1032-1035, Jul. 1995. However, this method also has computation complexity.
DISCLOSURE OF THE INVENTION
To solve the above problems, it is an objective of the present invention to provide an image data post-processing method for reducing quantization effects such as blocking artifacts, corner outliers and ringing noise, from an MPEG-decompressed image, which can perform low bit rate coding without complex computation, and an apparatus therefor.
According to an aspect of the present invention, there is provided an image data post-processing method for reducing quantization effect induced when image data compressed based on a block is decoded, the method comprising the steps of: (a) detecting semaphore representing whether or not post-processing is required, using distribution of inverse quantization coefficients of inverse-quantized image data and a motion vector representing the difference between the blocks of a previous video object plane (VOP) and blocks of a current VOP; and (b) filtering the decoded image data corresponding to the semaphore by a predetermined method, if it is determined by checking the detected semaphore that post-processing is required.
Preferably, the step (a) is performed on an intra-VOP in an intra-VOP mode, and on in inter-VOP in an inter-VOP mode.
Preferably, the semaphore includes a blocking semaphore representing whether or not reduction of blocking artifacts near block boundaries is required, and a ringing semaphore representing whether or not reduction of ringing noise near image edges is required.
Preferably, the blocking semaphore and the ringing semaphore of the intra-VOP are detected by investigating distribution of inverse quantization coefficients which are discrete cosine transform (DCT) coefficients after the compressed bitstream is inverse quantized.
Preferably, the blocking semaphore of the intra-VOP comprises a horizontal blocking semaphore (HBS) and a vertical blocking semaphore (VBS), assuming that the uppermost and leftmost pixel of the block, among 64 pixels constituting the 8×8 block, is pixel A, the pixel to the right of the pixel A is a pixel B, and the pixel below the pixel A is a pixel C, and the HBS and the VBS of the intra-VOP are extracted by the steps of: (a) calculating discrete cosine transform (DCT) coefficients on the inverse-quantized 8×8 block after the compressed image data is inversely quantized; (b) setting the HBS and the VBS to “1” which means that post-processing is required, if only the coefficient of the pixel A is non-zero; (c) setting the VBS to “1” which means that post-processing is required, if only the top row of the inverse-quantized 8×8 block includes non-zero coefficient pixel; and (d) setting the HBS to “1” which means that the post-process is required, if on the leftmost column of the inverse-quantized 8×8 block includes non-zero coefficient pixel.
Preferably, assuming that the uppermost and leftmost pixel of the block, among 64 pixels constituting the 8×8 block, is pixel A, the pixel to the right of the pixel A is a pixel B, and the pixel below the pixel A is a pixel C, the ringing semaphore (RS) of the intra-VOP is set to “1” which means post-processing is required, if any pixel other than the pixels A, B and C of the inverse-quantized 8×8 block has a non-zero coefficient.
Preferably, the blocking semaphore of the current inter-VOP comprises a horizontal blocking semaphore (HBS) and a vertical blocking semaphore (VBS), and assuming that a reference VOP comprises predetermined reference blocks, and the block of the reference VOP predicted by a motion vector (MVx,MVy) of a block Ac of the current inter-VOP is a motion block X, the HBS and the VBS of on the block Ac of the current inter-VOP are extracted by the steps of: checking the degree of overlapping between the motion block X and the reference blocks; performing a bi
Lee Yung-lyul
Park Hyun-Wook
Le Vu
Samsung Electronics Co,. Ltd.
LandOfFree
Image data post-processing method for reducing quantization... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Image data post-processing method for reducing quantization..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Image data post-processing method for reducing quantization... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3028050