Graph Neural Network-Based Attribute Auxiliary Structured Grouping for Person Re-Identification by Geyu Tang, Xingyu Gao, Zhenyu Chen, Huicai Zhong

“…Different from the existing clustering-based approaches that only utilize the similarity in feature space, we learn the feature representation from the similarities in both attribute space and feature space by graph learning on the pedestrian attribute graph. …”
Get full text

The Application of Imperialist Competitive Algorithm in Determining the Optimal Parameters of Empirical Area Reduction Method to Predict the Sedimentation Process in Dez Dam by Arash Azari, Sadegh Soori Hossein Bonakdari, Hossein Bonakdari

“…Then, by introducing new reservoir hydrography data in the optimal model, the sedimentation trend was predicted in the feature years (1410 and 1420). The results showed that this method was more consistent with the actual volume of the reservoir at different levels of the dams rather than the methods of Borland and Miller and Lara. …”
Get full text

ROLLING BEARING FAULT DIAGNOSIS BASED ON LMD AND ICA by CHEN ChongYang, XIONG BangShu, HUANG JianPing, MO Yan, LI XinMin

“…For the problem of Local Mean Decomposition（ LMD） was easily affected by noise interference when in the extraction of fault features,a rolling bearing fault diagnosis method which based on LMD and Independent Component Analysis（ ICA） was proposed. …”
Get full text

A Deep Neural Network-Based Approach to Media Hotspot Discovery by Pan Luo

“…First, the text data features are extracted based on the graphical convolutional neural network, and the temporal correlation of numerical information is modeled using gated recurrent units, and the numerical feature vectors are fused with the text feature vectors. …”
Get full text

A lightweight deep-learning model for parasite egg detection in microscopy images by Wenbin Xu, Qiang Zhai, Jizhong Liu, Xingyu Xu, Jing Hua

“…First, the neck of the YOLOv5n is modified to from a feature pyramid network (FPN) to an asymptotic feature pyramid network (AFPN) structure. …”
Get full text

Mapping Invasive <italic>Spartina alterniflora</italic> Using Phenological Information and Red-Edge Bands of Sentinel-2 Time-Series Data by Yiwei Ma, Li Zhuo, Jingjing Cao

“…Since the phenological information and red-edge spectral differences have been considered as informative features for identifying <italic>S. alterniflora</italic>, current studies mainly used them separately as classification features and seldom considered the differences of red-edge information at different phenological periods. …”
Get full text

Internet service providers as subjects of prevention of sexual crime on the Internet by A. K. Zharova

“…However, not all articles establishing criminal liability for violation of the sexual inviolability of children contain a qualifying feature - this is the use of the Network in the implementation of such activities. …”
Get full text

Position-Aware Graph Neural Network for Few-Shot SAR Target Classification by Jia Zheng, Ming Li, Peng Zhang, Yan Wu, Hongmeng Chen

“…Then, the self-attention network is brought in to capture the spatial dependence of any two positions in the feature maps, and the cross-attention network calculates the cross-attention between support and query feature maps to handle the problem of unseen categories. …”
Get full text

MFI-Net: A multi-resolution fusion input network for retinal vessel segmentation. by Yun Jiang, Chao Wu, Ge Wang, Hui-Xia Yao, Wen-Huan Liu

“…In this paper, we propose the MFI-Net (Multi-resolution fusion input network) network model to alleviate the above problem to a certain extent. The multi-resolution input module in MFI-Net avoids the loss of coarse-grained feature information in the shallow layer by extracting local and global feature information in different resolutions. …”
Get full text

ILR-Net: Low-light image enhancement network based on the combination of iterative learning mechanism and Retinex theory. by Mohan Yin, Jianbai Yang

“…Additionally, an efficient feature extraction module-global feature attention is designed to address the problem of feature loss. …”
Get full text

Gradient pooling distillation network for lightweight single image super-resolution reconstruction by Zhiyong Hong, GuanJie Liang, Liping Xiong

“…In the GPDN we leverage multi-level stacked feature distillation hybrid units to capture multi-scale feature representations, which are subsequently synthesized for dynamic feature space optimization. …”
Get full text

Regularization for Deep Imbalanced Regression Based on Quantitative Relationship by Heng Zhao, Jiehao Chen, Xianghua Fu

“…The Quantitative Relationship (QuanRel) regularizer is proposed to mitigate the problem of under-representation of features. The effect of frequent samples on the rare samples in the feature space is mitigated by the QuanRel. …”
Get full text

Penerapan Metode K-Medoids untuk Pengelompokan Mahasiswa Berpotensi Drop Out by Syamsul Bahri, Dwi Marisa Midyanti

“…Data mining analyzes data that already exists in the database to solve problems. One of the analyzes carried out is the clustering method. …”
Get full text

Gear Fault Diagnosis Based on EMD Decomposition and Levy-SSA-BP Neural Network by Xu Jingwen, Yang Ping, Yin Xiaojun

“…Secondly, calculating the correlation coefficient of each IMF with the original signal, and the feature extractions of each component are carried out to form a feature matrix. …”
Get full text

False traffic information detection based on weak classifiers integration in vehicular ad hoc networks by Xiang-wen LIU, Ya-li SHI, ENGXia F

“…Vehicles report traffic information mutually by self-organized manner in vehicular ad hoc networks (VANET),and the message need to be identified in the open network environment.However,it is very difficult for fast moving ve-hicles to detect a lot of traffic alert information in a short time.To solve this problem,a false traffic message detection method was presented based on weak classifiers integration.Firstly,the effective features of traffic alert information was extended and segmentation rules were designed to divide the information feature set into multiple feature subsets,then the corresponding weak classifiers were used to process feature subsets respectively according to the different character-istics of the subsets' features.Simulation experiments and performance analysis show that the selected weak classifiers integration method reduces the detection time,and because of the application of combined features,the detection rate is better than the test of using only some of the characteristics.…”
Get full text

Network traffic anomaly detection method based on multi-scale characteristic by Xueyuan DUAN, Yu FU, Kun WANG, Taotao LIU, Bin LI

“…Aiming at the problem that most of the traditional network traffic anomaly detection methods only pay attention to the fine-grained features of traffic data, and make insufficient use of multi-scale feature information, which may lead to low accuracy of anomaly detection results, a network traffic anomaly detection method based on multi-scale features was proposed.The original traffic was divided into sub-sequences with multiple observation spans by using multiple sliding windows of different scales, and the multi-level sequences of each sub-sequence were reconstructed by wavelet transform technology.Multi-level reconstructed sequences were generated by Chain SAE through feature space mapping, and a preliminary judgment of abnormality was made by the classifiers of each level according to the errors of the reconstructed sequences.The weighted voting strategy was adopted to summarize the preliminary judgment results of each level to form the final result judgment.Experimental results show that the proposed method can effectively mine the multi-scale feature information of network traffic, and the detection performance of abnormal traffic is obviously improved compared with traditional methods.…”
Get full text

Research on the detection of obstacles in front of unmanned vehicles in opencast mines based on binocular vision by Shunling RUAN, Huiguo ZHANG, Qinghua GU, Caiwu LU, Di LIU, Jing MAO

“…The Feffol network model proposed in this paper selects Efficient-v2 as the backbone network structure for feature extraction in the feature extraction stage, selects the Ebifpn feature pyramid module based on the SppCSP structure with SppCSP structure to improve the feature sensing field while enhancing the feature information of different sizes, uses the Focal Loss and CIoU Loss loss functions to balance positive and negative samples, and solve the problem of method failure when there is no intersection between the prediction frame and the real detection frame. …”
Get full text

Research on real-time monitoring method of mine personnel protective equipment with improved YOLOv8 by Lei ZHANG, Zhipeng SUN, Hongjing TAO, Shangkai HAO, Qianru YAN, Xiwei LI

“…In the process of multi-scale feature fusion, the less efficient feature transmission nodes are deleted. …”
Get full text

Spectrum sensing method based on residual dense network by Jianxin GAI, Xianfeng XUE, Ruixiang NAN, Jingyi WU

“…Aiming at the problem that the traditional spectrum sensing method based on convolutional neural network（CNN） did not make full use of image feature and the ability of extracting the image feature was limited by the shallow network structure， a spectrum sensing method based on the residual dense network （ResDenNet） was proposed.By adding dense connections in the traditional neural network， the information reuse of the image feature was achieved.Meanwhile， shortcut connections were employed at both ends of the dense unit to implement deeper network training.The spectrum sensing problem was transformed into the image binary classification problem.Firstly， the received signals were integrated into a matrix， which was normalized and transformed by gray level.The obtained gray level images were used as the input of the network.Then， the network was trained through dense learning and residual learning.Finally， the online data was input into the ResDenNet and spectrum sensing was implemented based on image classification.The numerical experiments show that the proposed method is superior to the traditional ones in terms of performance.When the SNR is as low as -19 dB， the detection probability of the proposed method is still high up to 0.96 with a low false alarm probability of 0.1， while a better generalization ability is displayed.…”
Get full text

A New Approach to ORB Acceleration Using a Modern Low-Power Microcontroller by Jorge Aráez, Santiago Real, Alvaro Araujo

“…This problem has commonly been solved by delegating this task to hardware-accelerated solutions like FPGAs or ASICs. …”
Get full text