US 5,859,788 A · Modulated lapped transform method

Abstract

A fast modulated lapped transform (MLT) method and architecture for image compression and decompression systems minimizes blocking artifacts associated with JPEG based discrete cosine transform (DCT) image compression systems. The MLT method combines fast block processing capabilities of wavelet transforms and fast block processing of DCT image compression systems. The modular and pipeline MLT architecture is fast by block processing but avoids the visual blocking artifacts that can be seen in most DCT-based compression systems. Improved MLT processors are implemented by an infinite impulse response filter operating on a product of the MLT window function and the input data stream. Forward and reverse MLT processors include a new fused multiply-add logic for fast computations and localized interconnections. The MLT processors can be combined into a bank of parallel processors in a one dimensional MLT architecture, which can be used for two-dimensional image transformation. The improved MLT implementation enables a modular architecture having a reduce number of multipliers and interconnects well suited for practical VLSI implementation.

Claims20

Dependent claims

A forward transform method for forward transforming a plurality of windows of input data respectively into a plurality of blocks of forward transformed data by a forward transform defined by a window function and a modulation function, each window is segmented into a plurality of blocks of input data, each block of input data and each block of forward transformed data have a predetermined block size k, each window has a predetermined window size n and n/k is a positive integer greater than one, each window has an overlapping block of input data in common with another adjacent window of the plurality of windows, wherein the improvement comprises,

multiplying forward window values by each window of input data for providing window product data for each window,
filtering the window product data for providing a block of forward transformed data, and
repeating the multiplying and filtering steps for each successive window of the plurality of windows of input data for providing the respective plurality of blocks of forward transformed data.

An inverse transform method for inverse transforming a plurality of windows of dequantized data respectively into a plurality of blocks of decompressed data by an inverse transform defined by a window function and a modulation function of a forward transform, each window is segmented into a plurality of blocks of dequantized data, each block of dequantized data and each block of decompressed data have a predetermined block size k, each window has a predetermined window size and n/k is a positive integer greater than one, each window has an overlapping block of dequantized data in common with another adjacent window of the plurality of windows, wherein the improvement comprises,

filtering a window of dequantized data for providing a window of inverse transformed data,
multiplying inverse window values to the window of inverse transformed data for providing the respective block of decompressed data, and
repeating the filtering and multiplying steps for each successive window of the plurality of windows of dequantized data for providing the respective plurality of blocks of decompressed data.

A compression decompression method for compressing an input image into a compressed image, and for decompressing the compressed image into a decompressed image, the method comprises the steps of,

receiving the input image for forward transforming a plurality of windows of input data of the input image respectively into a plurality of blocks of forward transformed data by a forward transform defined by a window function and a modulation function, each window is segmented into a plurality of blocks of input data, each block of input data and each block of forward transformed data have a predetermined block size k, each window has a predetermined window size n and n/k is a positive integer greater than one, each window has an overlapping block of input data in common with another adjacent window of the plurality of windows,
multiplying forward window values by each window of input data for providing window product data for each window,
filtering the window product data for providing a block of forward transformed data,
repeating the multiplying and filtering steps for each successive window of the plurality of windows of input data for providing the respective plurality of blocks of forward transformed data,
quantizing the blocks of forward transform data into quantized data,
encoding the quantized data into the compressed data,
decoding the compressed data into decoded data,
dequantizing the decoded data into windows of dequantized data,
receiving the windows of dequantized data for inverse transforming the plurality of windows of dequantized data respectively into a plurality of blocks of decompressed data by an inverse transform defined by the window function and the modulation function of the forward transform, each window is segmented into a plurality of blocks of dequantized data, each block of dequantized data and each block of decompressed data have a predetermined block size k, each window has a predetermined window size n and n/k is a positive integer greater than one, each window has an overlapping block of dequantized data in common with another adjacent window of the plurality of windows,
filtering a window of dequantized data for providing a window of inverse transformed data,
multiplying inverse window values to the window of inverse transformed data for providing a respective block of decompressed data, and
repeating the filtering and multiplying steps for each successive window of the plurality of windows of dequantized data for providing the respective plurality of blocks of decompressed data of the decompressed image.

9 more elements

Dependencies are inferred from the claim text.

Modulated lapped transform method

Life of this patent

Modulated lapped transform method

Life of this patent

Abstract

Claims20