Abstract:Wavelet transform is a powerful tool for multiresolution time-frequency analysis. It has been widely adopted in many image processing tasks, such as denoising, enhancement, fusion, and especially compression. Wavelets lead to the successful image coding standard JPEG-2000. Traditionally, wavelets were designed from the signal processing theory with certain assumption on the signal, but natural images are not as ideal as assumed by the theory. How to design content-adaptive wavelets for natural images remains a difficulty. Inspired by the recent progress of convolutional neural network (CNN), we propose iWave as a framework for deriving wavelet-like transform that is more suitable for natural image compression. iWave adopts an update-first lifting scheme, where the prediction filter is a trained CNN, to achieve wavelet-like transform. The CNN can be embedded into a deep network that is analogous to an auto-encoder, which is trained end-to-end. The trained wavelet-like transform still possesses the lifting structure, which ensures perfect reconstruction, supports multiresolution analysis, and is more interpretable than the deep networks trained as "black boxes." We perform experiments to verify the generality as well as the speciality of iWave in comparison with JPEG-2000. When trained with a generic set of natural images and tested on the Kodak dataset, iWave achieves on average 4.4% and up to 14% BD-rate reductions. When trained and tested with a specific kind of textures, iWave provides as high as 27% BD-rate reduction.

9/7 Discrete Wavelet Transform with Binary Coefficients in JPEG2000

Fixed-point Implementation of LS9/7 DWT

An Improved Wavelet Coefficients Symbols Coding of Embedded Zero-tree Wavelet

A Novel 9/7 Wavelet Filter Banks for Texture Image Coding

A VLSI Architecture of DWT for JPEG2000

JPEG2000 Based High-Precision Integer Wavelet Transform

A new wavelet lifting scheme for image compression applications

Lifting Based Implementation of a Kind of DWT and Its Application to Image Coding

Extensible Jpeg2000 Image Compression Systems

DWT and EBCOT VLSI Architecture for JPEG2000

Novel Scheme for JPEG2000 Rate Control Based on Number of Coding Passes

Optimized-approximate Integer-to-integer Wavelet Transform and Its VLSI Architecture of JPEG2000 Codec

5/3 Integer Wavelet Transform for Images Based on Visual C++

A Novel Embedded Coding Algorithm Based on the Reconstructed DCT Coefficients.

Optimization Algorithm for 7/5-Tap Wavelet Filters Based on First Order Autoregression Process

General Construction of Wavelet Filters Via a Lifting Scheme and Its Application in Image Coding

Image coding using dual-tree discrete wavelet transform.

High Performance VLSI Architecture of Discrete Wavelet Transformation for JPEG2000

Iwave: CNN-Based Wavelet-Like Transform for Image Compression.

A High Performance Design of FIR Filter Based on JPEG 2000 Standard

Optimization for integer 5/3 wavelet transform Based on the DM6446 Platform