计算机与通信工程学院 School of Computer and Communication Engineering
研究生工作
当前位置: 首页 > 研究生工作 > 正文

计通学院研究生学术交流报告会(第十一场)

发布时间: 2021-04-19 15:53:01 浏览量:

   

时间:2021423 下午 230  

地点:理科楼B311  

标题:Integrating Weighted Feature Fusion and the Spatial Attention Module with Convolutional Neural Networks for Automatic Aircraft Detection from SAR Images  

汇报人:王杰岚  

摘要:  

The automatic detection of aircrafts from SAR images is widely applied in both military and civil fields, but there are still considerable challenges. To address the high variety of aircraft sizes and complex background information in SAR images, a new fast detection framework based on convolution neural networks is proposed, which achieves automatic and rapid detection of aircraft with high accuracy. First, the airport runway areas are detected to generate the airport runway mask and rectangular contour of the whole airport are generated. Then, a new deep neural network proposed in this paper, named Efficient Weighted Feature Fusion and Attention Network (EWFAN), is used to detect aircrafts. EWFAN integrates the weighted feature fusion module, the spatial attention mechanism, and the CIF loss function. EWFAN can effectively reduce the interference of negative samples and enhance feature extraction, thereby significantly improving the detection accuracy. Finally, the airport runway mask is applied to the detected results to reduce false alarms and produce the final aircraft detection results. To evaluate the performance of the proposed framework, largescale Gaofen-3 SAR images with 1 m resolution are utilized in the experiment. The detection rate and false alarm rate of our EWFAN algorithm are 95.4% and 3.3%, respectively, which outperforms Efficientdet and YOLOv4. In addition, the average test time with the proposed framework is only 15.40 s, indicating satisfying efficiency of automatic aircraft detection..  

录取期刊:remote sensing  

   

标题:HOTSPOT: A UAV-Assisted Dynamic Mobility-Aware Offloading for Mobile Edge Computing in 3D Space》  

汇报人:马银宝  

摘要:    

For massive access to the Internet of Things, edge computing servers are installed on cellular ground base stations (GBS) with fixed geographical locations, which easily suffer from traffic overload of the end-user (EU) with high density and mobility. To provide reliable and flexible offloading service, unmanned aerial vehicles (UAV) are explored to assist edge computing, which relieves the computation offloading pressure of both EUs and GBS. However, most existing UAV researches focus on trajectory design to reduce offloading delay, which ignoring the variability of user distribution and the energy limitation of UAV. This paper proposes a novel UAV-assisted edge computing framework, named as HOTSPOT, which locates the UAV in 3D space according to the time-varying hot spot of user distribution and provides the corresponding edge computing offloading assistance. By formulating the UAV positioning problem into a maximum clique problem, a light-weighted deterministic algorithm is proposed based on stochastic gradient descent to search the optimal location of UAV. With the elaborate UAV position, HOTSPOT further gives an opportunistic offloading balanced scheme to reach low latency. Simulation results show that when the GBS load is 75%, HOTSPOT reduces the average offloading delay by 33%. When the GBS load reaches 90%, the average delay reduction is up to 80%..  

录取期刊:IEEE Internet of Things Journal  

   

标题:Lightweight Feedback Convolution Neural Network for Remote Sensing Images Super-Resolution  

汇报人:吴一鸣  

摘要:    

There are lots of image data in the eld of remote sensing, most of which have low-resolution due to the limited image sensor. The super-resolution method can effectively restore the low-resolution image to the high-resolution image. However, the existing super-resolution method has both heavy computing burden and number of parameters. For saving costs, we propose the feedback ghost residual dense network (FGRDN), which considers the feedback mechanism as the framework to attain lower features through high-level rening. Further, for feature extraction, we replace the convolution of the residual dense blocks (RDBs) with ghost modules (GMs), which can remove the redundant channels and avoid the increase of parameters along with the network depth. Finally, the spatial and channel attention module (SCM) is employed in the end of the RDB to learn more useful information from features. Compared to other SOTA lightweight algorithms, our proposed algorithm can reach convergences more rapidly with fewer parameters, and the performance of the network can be markedly enhanced on the image texture and object contour reconstruction with better peak signal-to-noise ratio (PSNR) and structural similarity (SSIM).  

录取期刊:IEEE ACCESS  

   


Copyright © 2020 All Right Reserved 新利luck在线·(中国)有限公司官网 计算机与通信工程学院 版权所有

地址:新利luck在线·(中国)有限公司官网云塘校区理科楼B-404物联网实验室 电话:0731-85258462