Deep Dictionary Learning: Algorithm, Theory and Application
摘要截稿:
全文截稿: 2020-04-15
影响因子: 4.438
期刊难度:
CCF分类: C类
中科院JCR分区:
• 大类 : 计算机科学 - 2区
• 小类 : 计算机:人工智能 - 2区
Overview
Dictionary Learning (DL) is a long-standing popular topic for visual image representation due to its great success to image restoration, de-noising and classification, etc. DL aims at representing data using a linear combination of a few highly correlated atoms in a dictionary D. But how to obtain a desired dictionary from inputs still remains a challenging task to date. It is noteworthy that most existing DL algorithms represent data using a single-layer framework, so they usually fail to obtain the deep feature representations with more useful and valuable hidden information discovered. In recent years, with the fast development of deep learning and multi-layer neural networks, it will be helpful to propose deeper or multi-layer DL frameworks for representation learning. Although certain efforts have been made to incorporate the deep learning into the DL, most designs of so-called deep dictionary learning (DDL) algorithms are still less straightforward. For example, some existing algorithms feed deep features of the deep networks into DL for representation learning, or perform the DL first and then use the reconstructed data for deep learning. As such, it is now necessary to integrate the DL with deep networks, and explore the advanced algorithms, theories and optimization approaches for the deep dictionary learning.
In this special issue, we invite contributions from diverse research fields, such as deep representation learning, image processing, and computer vision, etc., developing novel algorithms from high-dimensional data.
The topics of interest include, but are not limited to:
Survey/vision/review of dictionary learning
Robust dictionary learning
Online dictionary learning
Deep/multi-layer dictionary learning
Convolutional dictionary learning
Structured dictionary learning
Bayesian dictionary learning
Coupled/semi-coupled dictionary learning
Optimization for dictionary learning/deep dictionary learning