Peng Liu

Dr. Peng Liu

AIRCAS
orcid

Research Focus & Metrics

Academic Career Length

20Years (2007-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
31 January 2022 in ACM Computing Surveys
Verified
Journal Article
Open access
Peng LiuPeng Liu
Peng Liu
LWLizhe Wang
Lizhe Wang
RRRajiv Ranjan
Rajiv Ranjan
GHGuojin He
Guojin He
LZLei Zhao
Lei Zhao
Which samples should be labelled in a large dataset is one of the most important problems for the training of deep learning. So far, a variety of active sample selection strategies related to deep learning have been proposed in the literature. We defined them as Active Deep Learning (ADL) only if their predictor or selector is a deep model, where the basic learner is called the predictor and the labeling schemes are called the selector. In this survey, we categorize ADL into model-driven ADL and data-driven ADL by whether its selector is model driven or data driven. We also introduce the different characteristics of the two major types of ADL, respectively. We summarized three fundamental factors in the designation of a selector. We pointed out that, with the development of deep learning, the selector in ADL also is experiencing the stage from model driven to data driven. The advantages and disadvantages between data-driven ADL and model-driven ADL are thoroughly analyzed. Furthermore, different sub-classes of data-drive or model-driven ADL are also summarized and discussed emphatically. Finally, we survey the trend of ADL from model driven to data driven.

1 January 2021 in Atmospheric Pollution Research
Verified
Journal Article
Closed access
LZLuo Zhang
Luo Zhang
Peng LiuPeng Liu
Peng Liu
LZLei Zhao
Lei Zhao
GWGuizhou Wang
Guizhou Wang
WZWangfeng Zhang
Wangfeng Zhang
JLJianbo Liu
Jianbo Liu

17 June 2015 in Frontiers in Environmental Science
Verified
Journal Article
Open access
PDF file
Peng LiuPeng Liu
Peng Liu
We have entered an era of big data. It is popular to refer to the three Vs when characterizing big data: remarkable growths in the volume, velocity and variety of data. However, this statement is too general. Remote-sensing big data has several concrete and special characteristics: multi-source, multi-scale, high-dimensional, dynamic-state, isomer, and non-linear characteristics. This survey explains these characteristics in detail. Furthermore, according to whether the characteristics are closely related to the instruments or methods of data acquisition, we points out that the dynamic-state, multi-scale and non-linear characteristics are intrinsic characteristics of remote-sensing big data while the multi-source, high-dimensional and isomer characteristics are extrinsic characteristics of remote- sensing big data. In addition, we briefly review promising techniques and applications of remote-sensing big data.

5 July 2016 in Soft Computing
Verified
Journal Article
Closed access
PDF file
Lizhe WangLizhe Wang
Lizhe Wang
JZJiabin Zhang
Jiabin Zhang
Peng LiuPeng Liu
Peng Liu
KCKim-Kwang Raymond Choo
Kim-Kwang Raymond Choo
FHFang Huang
Fang Huang

23 May 2022 in IEEE Geoscience and Remote Sensing Magazine
Verified
Journal Article
Closed access
Peng LiuPeng Liu
Peng Liu
JLJun Li
Jun Li
LWLizhe Wang
Lizhe Wang
GHGuojin He
Guojin He
In the past decades, remote sensing (RS) data fusion has always been an active research community. A large number of algorithms and models have been developed. Generative adversarial networks (GANs), as an important branch of deep learning, show promising performances in a variety of RS image fusions. This review provides an introduction to GANs for RS data fusion. We briefly review the frequently used architecture and characteristics of GANs in data fusion and comprehensively discuss how to use GANs to realize fusion for homogeneous RS, heterogeneous RS, and RS and ground observation (GO) data. We also analyze some typical applications with GAN-based RS image fusion. This review provides insight into how to make GANs adapt to different types of fusion tasks and summarizes the advantages and disadvantages of GAN-based RS data fusion. Finally, we discuss promising future research directions and make a prediction on their trends.

24 May 2022 in IEEE Transactions on Geoscience and Remote Sensing
Verified
Journal Article
Closed access
BSBingze Song
Bingze Song
Peng LiuPeng Liu
Peng Liu
Jun LiJun Li
Jun Li
Lizhe WangLizhe Wang
Lizhe Wang
LZLuo Zhang
Luo Zhang
JLJianbo Liu
Jianbo Liu
Due to the limitation of technology and budget, it is often difficult for sensors of a single remote sensing satellite to have both high-temporal and high-spatial (HTHS) resolution at the same time. In this article, we proposed a new multilevel feature fusion with generative adversarial network (MLFF-GAN) for generating fusion HTHS images. The MLFF-GAN mainly uses U-net-like architecture, and its generator is composed of three stages: feature extraction, feature fusion, and image reconstruction. In the feature extraction and reconstruction stage, the generator employs the encoding and decoding structure to extract three groups of multilevel features (MLFs), which can cope with the huge difference of resolution between high-resolution images and low-resolution images. In the feature fusion stage, adaptive instance normalization (AdaIN) block is designed to learn the global distribution relationship between multitemporal images, and an attention module (AM) is used to learn the local information weights for the change of small areas. The proposed MLFF-GAN was tested on two Landsat and Moderate-Resolution Imaging Spectroradiometer (MODIS) datasets. Some state-of-the-art algorithms are comprehensively compared with MLFF-GAN. We also carried on the ablation experiment to test the effectiveness of different submodules in the MLFF-GAN. The experiment results and ablation analysis show the better performances of the proposed method when compared with other methods. The code is available at https://github.com/songbingze/MLFF-GAN .