Intrusion detection algorithm based on image enhanced convolutional neural network

Wang, Qian; Zhao, Wenfang; Ren, Jiadong

doi:10.3233/jifs-210863

Cited by 8 publications

(3 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The Softmax function was used rather than the Euclidean Radial Basis Function adopted in LeNet-5, because Softmax can be well applied to multi-classification problems and has better probability distribution characteristics. A Tanh activation function was chosen instead of Relu because Relu considers eigenvalues less than zero as zero, which will lose some original information of samples and increase the sparsity of the network [51]. A categorical-cross entropy loss function integrated with an adaptive moment estimation (Adam) optimizer was employed to improve training [52].…”

Section: Modeling Methodsmentioning

confidence: 99%

Identification of Geographical Origin of Chinese Chestnuts Using Hyperspectral Imaging with 1D-CNN Algorithm

Jiang

et al. 2021

Agriculture

View full text Add to dashboard Cite

The adulteration in Chinese chestnuts affects the quality, taste, and brand value. The objective of this study was to explore the feasibility of the hyperspectral imaging (HSI) technique to determine the geographical origin of Chinese chestnuts. An HSI system in spectral range of 400–1000 nm was applied to identify a total of 417 Chinese chestnuts from three different geographical origins. Principal component analysis (PCA) was preliminarily used to investigate the differences of average spectra of the samples from different geographical origins. A deep-learning-based model (1D-CNN, one-dimensional convolutional neural network) was developed first, and then the model based on full spectra and optimal wavelengths were established for various machine learning methods, including partial least squares-discriminant analysis (PLS-DA) and particle swarm optimization-support vector machine (PSO-SVM). The optimal results based on full spectra for 1D-CNN, PLS-DA, and PSO-SVM models were 97.12%, 97.12%, and 95.68%, respectively. Competitive adaptive reweighted sampling (CARS) and a successive projections algorithm (SPA) were individually utilized for wavelengths selection, and the results of simplified models generally improved. The contrasting results demonstrated that the prediction accuracies of SPA-PLS-DA and 1D-CNN both reached 97.12%, but 1D-CNN presented a higher Kappa coefficient value than SPA-PLS-DA. Meanwhile, the sensitivities and specificities of SPA-PLS-DA and 1D-CNN models were both above 90% for the samples from each geographical origin. These results indicated that both SPA-PLS-DA and 1D-CNN models combined with HSI have great potential for the geographical origin identification of Chinese chestnuts.

show abstract

Section: Modeling Methodsmentioning

confidence: 99%

Identification of Geographical Origin of Chinese Chestnuts Using Hyperspectral Imaging with 1D-CNN Algorithm

Jiang

et al. 2021

Agriculture

View full text Add to dashboard Cite

show abstract

“…Year Model Dataset Classification Performance [17] CNN KDD Binary DR=97.66 [18] CNN ISCX2012 Multiclass Acc=99.69 [19] RNN NSLKDD Multiclass Acc=83.28 [20] LSTM NSLKDD Binary Acc=98.85 [21] LSTM AWID Binary Acc =98.22 [22] BiLSTM UNSW-NB15 Binary Acc=95.71 [23] DBN NSLKDD Multiclass Acc=82.08 [24] BGRU+MLP KDD99 Multiclass Acc=99.84 [24] BGRU+MLP NSLKDD Multiclass Acc=99.24 [25] RNN NSLKDD Multiclass Acc=81.29 [26] LSTM NSLKDD Multiclass Acc=89.00 [27] LSTM NSLKDD Multiclass Acc=73.18 [27] CNN NSLKDD Multiclass Acc=80.13 [28] LSTM ISCX2012 Binary Acc =99.99 [29] LSTM, GRU Personal Dataset Multiclass Acc =96.08 [30] GRU Personal Dataset Multiclass Acc=95.60 [31] GRU Personal Dataset Multiclass F1 =80.30 [32] RNN BoT-IoT Multiclass Acc =98.20 [33] CNN NSLKDD Binary Acc =99.67 [34] CNN NSLKDD Multiclass Acc=89.41 [35] DNN KDD99 Multiclass Acc=92.90 [35] DNN KDD99 Binary Acc=93.00 [36] DNN UNSW-NB15 Binary Acc=99.19 [36] DNN UNSW-NB15 Multiclass Acc=97.04 [36] DNN CICIDS2017 Binary Acc=97.73 [37] CNN-LSTM NSLKDD Binary Acc=98.90 [38] LSTM KDD99 Binary Acc=99.43 [38] GRU KDD99 Binary Acc=99.06 [38] Bi-LSTM KDD99 Binary Acc=82.20 [39] GRU NSLKDD Binary Acc=82.87 [40] GRU Gas Pipeline Binary Acc=91.70 [16] GRU ISCX 2012 Binary Acc=99.90 [41] DNN BoT-IoT Multiclass Acc=98.09 [42] CNN-LSTM CISC-2010 Binary Acc=91.54 [42] CNN-LSTM CICISC-2017 Binary Acc=93.00 [43] CNN-LSTM CICIDS2017 Binary Acc=99.03…”

Section: Articlementioning

confidence: 99%

Design and Development of RNN Anomaly Detection Model for IoT Networks

2022

View full text Add to dashboard Cite

Cybersecurity is important today because of the increasing growth of the Internet of Things (IoT), which has resulted in a variety of attacks on computer systems and networks. As the number of various IoT devices and services grows, cyber security will become an increasingly difficult issue to manage. Malicious traffic identification using deep learning techniques has emerged as a key component of networkbased intrusion detection systems (IDS). Deep learning methods have been a research focus in network intrusion detection. A recurrent neural network is useful in a wide range of applications. This paper proposes a novel deep learning model for detecting anomalies in IoT networks using recurrent neural networks. The proposed model is implemented in IoT networks utilizing LSTM, BiLSTM, and GRU-based approaches for anomaly detection. A convolutional neural network can analyze input features without losing important information, making them particularly well suited for feature learning. In addition, we propose a hybrid deep learning model based on convolutional and recurrent neural networks. Finally, employing LSTM, BiLSTM, and GRU-based techniques, we propose a lightweight deep learning model for binary classification. The proposed deep learning models are validated using NSLKDD, BoT-IoT, IoT-NI, MQTT, MQTTset, IoT-23, and IoT-DS2 datasets. Our proposed binary and multiclass classification model achieved high accuracy, precision, recall, and F1 score compared to current deep learning implementations.

show abstract

“…Wang et al applied cloud data to texture optimization of object surfaces. The unit quaternion method was used to solve for the orientation elements of the image to achieve texture optimality and compensate for the occluded area [17]. Chambers et al used a new frame feature point matching algorithm based on FPGA technology to extract from edge feature points and then perform edge feature point matching by computing feature descriptors [18].…”

Section: Related Workmentioning

confidence: 99%

3D Animation Graphic Enhancing Process Effect Simulation Analysis

Liu

Dung

2022

Wireless Communications and Mobile Computing

View full text Add to dashboard Cite

Processing of graphics is the intermediate stage of an animated film. The adequacy of the preproduction preparation often requires the creators to repeatedly discuss a series of issues such as the story of the script, the structure of the play, the style of the art design and the setting of the scenes, the shape of the characters, and the style of the music. The study of the process of 3D animation graphics enhancing effect optimization can effectively improve the visual effect of images. The optimization of the effect of authenticity of 3D animation graphics requires the optimization of lighting of animation graphics, and the optimization of texture processing of images by extending the shadow map method to complete the optimization of the effect of authenticity of animation graphics. Image enhancement is a very important branch in the field of digital processing of graphics. It is a method of making some transformations to the information or data of an image by some strategy to make the region of interest more prominent in the image while suppressing the secondary information that does not need to be enhanced, with the ultimate goal of making the image conform to the visual response characteristics. It is in this context that this paper proposes to build a process of graphics simulation platform to simulate and analyze the effect of process of 3D animation graphics enhancing. The experimental results show that the enhancement of graphics reduces the running time by 16.5 s and 14.3 s compared to the CDD and FMM algorithms, respectively, and that the MSE and PSNR are both improved, indicating that the enhancement of graphics effectively improves the restoration effect while running quickly. Therefore, it is confirmed that the process of 3D animation graphics enhancing effect has a good balance to obtain a more desirable animation in 3D enhancement effect.

show abstract

Intrusion detection algorithm based on image enhanced convolutional neural network

Cited by 8 publications

References 19 publications

Identification of Geographical Origin of Chinese Chestnuts Using Hyperspectral Imaging with 1D-CNN Algorithm

Identification of Geographical Origin of Chinese Chestnuts Using Hyperspectral Imaging with 1D-CNN Algorithm

Design and Development of RNN Anomaly Detection Model for IoT Networks

3D Animation Graphic Enhancing Process Effect Simulation Analysis

Contact Info

Product

Resources

About