4.7 Article

Dynamic Neural Network for Lossy-to-Lossless Image Coding

Journal

IEEE TRANSACTIONS ON IMAGE PROCESSING
Volume 31, Issue -, Pages 569-584

Publisher

IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TIP.2021.3132825

Keywords

Image coding; Transform coding; Discrete wavelet transforms; Computer architecture; Image reconstruction; Convolutional neural networks; Adaptation models; Lifting scheme; image compression; adaptive wavelets; optimization; neural networks; adaptive learning

Funding

  1. French National Research Agency (ANR)
  2. Institut Universitaire de France

Ask authors/readers for more resources

This paper proposes a compression scheme based on neural network learning of lifting operators, and improves the dynamic fully connected neural network model for better consideration of input images. The experimental results demonstrate the advantages of this method in lossy and lossless image compression.
Lifting-based wavelet transform has been extensively used for efficient compression of various types of visual data. Generally, the performance of such coding schemes strongly depends on the lifting operators used, namely the prediction and update filters. Unlike conventional schemes based on linear filters, we propose, in this paper, to learn these operators by exploiting neural networks. More precisely, a classical Fully Connected Neural Network (FCNN) architecture is firstly employed to perform the prediction and update. Then, we propose to improve this FCNN-based Lifting Scheme (LS) in order to better take into account the input image to be encoded. Thus, a novel dynamical FCNN model is developed, making the learning process adaptive to the input image contents for which two adaptive learning techniques are proposed. While the first one resorts to an iterative algorithm where the computation of two kinds of variables is performed in an alternating manner, the second learning method aims to learn the model parameters directly through a reformulation of the loss function. Experimental results carried out on various test images show the benefits of the proposed approaches in the context of lossy and lossless image compression.

Authors

I am an author on this paper
Click your name to claim this paper and add it to your profile.

Reviews

Primary Rating

4.7
Not enough ratings

Secondary Ratings

Novelty
-
Significance
-
Scientific rigor
-
Rate this paper

Recommended

No Data Available
No Data Available