Lei Jiang

Lei Jiang

Auto-generated profile, based on publicly available information

Lei Jiang

Information and Navigation College, Air Force Engineering University, Xi’an, China

Research Focus & Metrics

Academic Career Length

14Years (2013-Present)

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

Representative Publications

View All Publications
29 July 2020 in IEEE Access
Verified
Journal Article
Open access
YSyunhao shi
yunhao shi
HXHua Xu
Hua Xu
LJLei Jiang
Lei Jiang
YLYinghui Liu
Yinghui Liu
In modulation classification domain, handcrafted feature based method can fit well from a few labeled samples, while deep learning based method require a large amount of samples to achieve the superior classification performance. In order to improve the modulation classification accuracy under the constraint of limited labeled samples, this paper proposes a few-shot modulation classification method based on feature dimension reduction and pseudo-label training (FDRPLT), which combines handcrafted feature based method with deep learning based method. First, an optimal low-dimensional feature subset is created by the combination of the handcrafted features and autoencoder-extracted features post-processed by a feature selection algorithm. Then, a fully connected network (FCN), trained on a small number of labeled signals, is designed for the automated annotating, where unlabeled samples can be annotated and used for the later convolution neural network (CNN) training. The simulation results show that the classification accuracy of eight kinds of modulation types can reach to 98.3% when the SNR is 20dB.

1 July 2013
Verified
Conference Paper
Closed access
SXSiyuan Xiao
Siyuan Xiao
BCBangyu Cai
Bangyu Cai
LJLei Jiang
Lei Jiang
YWYiweng Wang
Yiweng Wang
WCWeidong Chen
Weidong Chen
XZXiaoxiang Zheng
Xiaoxiang Zheng
Event-related potentials (ERP)-based image triage (or search) in the context of Rapid Serial Visual Presentation (RSVP) exploits difference in the human brain response to target and distracted stimuli in the form of an image. So far, most paradigms focus on image triage (or search) among rough object categories. In this paper, we explored the possibility and effectiveness of target detection among finer categories like different animals. We analyzed on the difference of ERP components in two image search tasks, a simple-recognition task in which all images of a target are the same and a discriminative-recognition task in which all images are randomly different but belong to the same target category (the same kind of animal). We observed that the P3 amplitude reduced and the P3 latency delayed on the discriminative-recognition condition due to the increased difficulty of identifying different images belonging to the same category. But the average area under ROC curve reached 0.82 which indicated that rapid target detection among finer categories by single-trial ERP were feasible with trivial contribution of N1 and stable contribution of N2 and P3.