Conveyor-Belt Detection of Conditional Deep Convolutional Generative Adversarial Network

In underground mining,the belt is a critical component,as its state directly affects the safe and stable operation of the conveyor.Most of the existing non-contact detection methods based on machine vision can only detect a single type of damage and they require pre-processing operations.This tends to cause a large amount of calculation and low detection precision.To solve these problems,in the work described in this paper a belt tear detection method based on a multi-class conditional deep convolutional generative adversarial network(CDCGAN)was designed.In the traditional DCGAN,the image generated by the generator has a certain degree of randomness.Here,a small number of labeled belt images are taken as conditions and added them to the generator and discriminator,so the generator can generate images with the characteristics of belt damage under the aforementioned conditions.Moreover,because the discriminator cannot identify multiple types of damage,the multi-class softmax function is used as the output function of the discriminator to output a vector of class probabilities,and it can accurately classify cracks,scratches,and tears.To avoid the features learned incompletely,skiplayer connection is adopted in the generator and discriminator.This not only can minimize the loss of features,but also improves the convergence speed.Compared with other algorithms,experimental results show that the loss value of the generator and discriminator is the least.Moreover,its convergence speed is faster,and the mean average precision of the proposed algorithm is up to 96.2%,which is at least 6%higher than that of other algorithms.

RETRACTED:<i>Realization of Virtual Human Face Based on Deep Convolutional Generative Adversarial Networks</i>

Short Retraction Notice The authors claim that this paper needs modifications. This article has been retracted to straighten the academic record. In making this decision the Editorial Board follows COPE's Retraction Guidelines. The aim is to promote the circulation of scientific research by offering an ideal research publication platform with due consideration of internationally accepted standards on publication ethics. The Editorial Board would like to extend its sincere apologies for any inconvenience this retraction may have caused. Editor guiding this retraction: Prof. Baozong Yuan(EiC of JSIP) The full retraction notice in PDF is preceding the original paper, which is marked "RETRACTED".

Robust Image Watermarking Based on Generative Adversarial Network

Digital watermark embeds information bits into digital cover such as images and videos to prove the creator’s ownership of his work.In this paper,we propose a robust image watermark algorithm based on a generative adversarial network.This model includes two modules,generator and adversary.Generator is mainly used to generate images embedded with watermark,and decode the image damaged by noise to obtain the watermark.Adversary is used to discriminate whether the image is embedded with watermark and damage the image by noise.Based on the model Hidden(hiding data with deep networks),we add a high-pass filter in front of the discriminator,making the watermark tend to be embedded in the mid-frequency region of the image.Since the human visual system pays more attention to the central area of the image,we give a higher weight to the image center region,and a lower weight to the edge region when calculating the loss between cover and embedded image.The watermarked image obtained by this scheme has a better visual performance.Experimental results show that the proposed architecture is more robust against noise interference compared with the state-of-art schemes.

Exploration of the Relation between Input Noise and Generated Image in Generative Adversarial Networks

In this paper,we propose a hybrid model aiming to map the input noise vector to the label of the generated image by the generative adversarial network(GAN).This model mainly consists of a pre-trained deep convolution generative adversarial network(DCGAN)and a classifier.By using the model,we visualize the distribution of two-dimensional input noise,leading to a specific type of the generated image after each training epoch of GAN.The visualization reveals the distribution feature of the input noise vector and the performance of the generator.With this feature,we try to build a guided generator(GG)with the ability to produce a fake image we need.Two methods are proposed to build GG.One is the most significant noise(MSN)method,and the other utilizes labeled noise.The MSN method can generate images precisely but with less variations.In contrast,the labeled noise method has more variations but is slightly less stable.Finally,we propose a criterion to measure the performance of the generator,which can be used as a loss function to effectively train the network.

Deep convolutional adversarial graph autoencoder using positive pointwise mutual information for graph embedding

Graph embedding aims to map the high-dimensional nodes to a low-dimensional space and learns the graph relationship from its latent representations.Most existing graph embedding methods focus on the topological structure of graph data,but ignore the semantic information of graph data,which results in the unsatisfied performance in practical applications.To overcome the problem,this paper proposes a novel deep convolutional adversarial graph autoencoder(GAE)model.To embed the semantic information between nodes in the graph data,the random walk strategy is first used to construct the positive pointwise mutual information(PPMI)matrix,then,graph convolutional net-work(GCN)is employed to encode the PPMI matrix and node content into the latent representation.Finally,the learned latent representation is used to reconstruct the topological structure of the graph data by decoder.Furthermore,the deep convolutional adversarial training algorithm is introduced to make the learned latent representation conform to the prior distribution better.The state-of-the-art experimental results on the graph data validate the effectiveness of the proposed model in the link prediction,node clustering and graph visualization tasks for three standard datasets,Cora,Citeseer and Pubmed.

Adversarial Training-Aided Time-Varying Channel Prediction for TDD/FDD Systems

In this paper, a time-varying channel prediction method based on conditional generative adversarial network(CPcGAN) is proposed for time division duplexing/frequency division duplexing(TDD/FDD) systems. CPc GAN utilizes a discriminator to calculate the divergence between the predicted downlink channel state information(CSI) and the real sample distributions under a conditional constraint that is previous uplink CSI. The generator of CPcGAN learns the function relationship between the conditional constraint and the predicted downlink CSI and reduces the divergence between predicted CSI and real CSI.The capability of CPcGAN fitting data distribution can capture the time-varying and multipath characteristics of the channel well. Considering the propagation characteristics of real channel, we further develop a channel prediction error indicator to determine whether the generator reaches the best state. Simulations show that the CPcGAN can obtain higher prediction accuracy and lower system bit error rate than the existing methods under the same user speeds.

Multi-Generator Discriminator Network Using Texture-Edge Information

In the proposed paper,a parallel structure type Generative Adversarial Network(GAN)using edge and texture information is proposed.In the existing GAN-based model,many learning iterations had to be given to obtaining an output that was somewhat close to the original data,and noise and distortion occurred in the output image even when learning was performed.To solve this problem,the proposed model consists of two generators and three discriminators to propose a network in the form of a parallel structure.In the network,each edge information and texture information were received as inputs,learning was performed,and each character was combined and outputted through the Combine Discriminator.Through this,edge information and distortion of the output image were improved even with fewer iterations than DCGAN,which is the existing GAN-based model.As a result of learning on the network of the proposed model,a clear image with improved contour and distortion of objects in the image was output from about 50,000 iterations.

Toward Improved Accuracy in Quasi-Static Elastography Using Deep Learning

Elastography is a non-invasive medical imaging technique to map the spatial variation of elastic properties of soft tissues.The quality of reconstruction results in elastography is highly sensitive to the noise induced by imaging measurements and processing.To address this issue,we propose a deep learning(DL)model based on conditional Generative Adversarial Networks(cGANs)to improve the quality of nonhomogeneous shear modulus reconstruction.To train this model,we generated a synthetic displacement field with finite element simulation under known nonhomogeneous shear modulus distribution.Both the simulated and experimental displacement fields are used to validate the proposed method.The reconstructed results demonstrate that the DL model with synthetic training data is able to improve the quality of the reconstruction compared with the well-established optimization method.Moreover,we emphasize that our DL model is only trained on synthetic data.This might provide a way to alleviate the challenge of obtaining clinical or experimental data in elastography.Overall,this work addresses several fatal issues in applying the DL technique into elastography,and the proposed method has shown great potential in improving the accuracy of the disease diagnosis in clinical medicine.

Enhancing Pneumonia Detection in Pediatric Chest X-Rays Using CGAN-Augmented Datasets and Lightweight Deep Transfer Learning Models

Pneumonia ranks as a leading cause of mortality, particularly in children aged five and under. Detecting this disease typically requires radiologists to examine chest X-rays and report their findings to physicians, a task susceptible to human error. The application of Deep Transfer Learning (DTL) for the identification of pneumonia through chest X-rays is hindered by a shortage of available images, which has led to less than optimal DTL performance and issues with overfitting. Overfitting is characterized by a model’s learning that is too closely fitted to the training data, reducing its effectiveness on unseen data. The problem of overfitting is especially prevalent in medical image processing due to the high costs and extensive time required for image annotation, as well as the challenge of collecting substantial datasets that also respect patient privacy concerning infectious diseases such as pneumonia. To mitigate these challenges, this paper introduces the use of conditional generative adversarial networks (CGAN) to enrich the pneumonia dataset with 2690 synthesized X-ray images of the minority class, aiming to even out the dataset distribution for improved diagnostic performance. Subsequently, we applied four modified lightweight deep transfer learning models such as Xception, MobileNetV2, MobileNet, and EfficientNetB0. These models have been fine-tuned and evaluated, demonstrating remarkable detection accuracies of 99.26%, 98.23%, 97.06%, and 94.55%, respectively, across fifty epochs. The experimental results validate that the models we have proposed achieve high detection accuracy rates, with the best model reaching up to 99.26% effectiveness, outperforming other models in the diagnosis of pneumonia from X-ray images.

Research on Image Recognition Using Deep Learning Techniques

This study delves into the applications,challenges,and future directions of deep learning techniques in the field of image recognition.Deep learning,particularly Convolutional Neural Networks(CNNs),Recurrent Neural Networks(RNNs),and Generative Adversarial Networks(GANs),has become key to enhancing the precision and efficiency of image recognition.These models are capable of processing complex visual data,facilitating efficient feature extraction and image classification.However,acquiring and annotating high-quality,diverse datasets,addressing imbalances in datasets,and model training and optimization remain significant challenges in this domain.The paper proposes strategies for improving data augmentation,optimizing model architectures,and employing automated model optimization tools to address these challenges,while also emphasizing the importance of considering ethical issues in technological advancements.As technology continues to evolve,the application of deep learning in image recognition will further demonstrate its potent capability to solve complex problems,driving society towards more inclusive and diverse development.

High-speed multimode fiber imaging system based on conditional generative adversarial network

The multimode fiber(MMF)has great potential to transmit high-resolution images with less invasive methods in endoscopy due to its large number of spatial modes and small core diameter.However,spatial modes crosstalk will inevitably occur in MMFs,which makes the received images become speckles.A conditional generative adversarial network(GAN)composed of a generator and a discriminator was utilized to reconstruct the received speckles.We conduct an MMF imaging experimental system of transmitting over 1 m MMF with a 50μm core.Compared with the conventional method of U-net,this conditional GAN could reconstruct images with fewer training datasets to achieve the same performance and shows higher feature extraction capability.

Deep-learning-based methods for super-resolution fluorescence microscopy

The algorithm used for reconstruction or resolution enhancement is one of the factors affectingthe quality of super-resolution images obtained by fluorescence microscopy.Deep-learning-basedalgorithms have achieved stateof-the-art performance in super-resolution fluorescence micros-copy and are becoming increasingly attractive.We firstly introduce commonly-used deep learningmodels,and then review the latest applications in terms of the net work architectures,the trainingdata and the loss functions.Additionally,we discuss the challenges and limits when using deeplearning to analyze the fluorescence microscopic data,and suggest ways to improve the reliability and robustness of deep learning applications.

基于格拉姆角场与深度卷积生成对抗网络的行星齿轮箱故障诊断

针对行星齿轮箱故障诊断中样本分布不均衡所引起的模型泛化能力差及诊断精度低等问题,采用格拉姆角场图像编码技术和深度卷积生成对抗网络相结合进行数据增强,融合AlexNet卷积神经网络进行故障诊断。将采集到的一维振动信号转化为格拉姆角场图,按比例划分训练集与测试集,将训练集样本与随机向量输入到深度卷积生成对抗网络模型中,交替训练生成器与判别器,达到纳什平衡,生成与原始样本类似的生成样本,从而实现故障样本的增广。用原始样本与生成的增广样本训练卷积神经网络分类模型,完成行星齿轮箱的故障识别。实验结果表明,所提方法能够有效提升样本不均衡条件下的行星齿轮箱故障诊断精度,使之达到99.15%,且能使收敛速度更快。

摩尔纹图案自动去除技术综述

如今,数码相机和智能手机在人们的生活中扮演着越来越重要的角色,已经成为人们感知世界、记录信息和沟通交流的主要工具.当使用这些设备拍摄电子屏幕时,显示设备和摄像头传感器网格之间往往会发生混叠,通常导致图片中存在不规则分布的摩尔纹干扰图案,从而严重影响了拍摄图像的视觉质量效果.因此,摩尔纹图案去除方法研究对于拍摄图像的后期处理具有重要意义.为此,详细梳理了摩尔纹去除研究的发展脉络,并根据不同的适用场景和技术实现将现有方法分为2类:基于先验知识的摩尔纹去除方法和基于深度学习的摩尔纹去除方法.鉴于深度学习网络中训练数据集的收集和对齐方式不同,该类方法又分为基于卷积神经网络(CNN)的摩尔纹去除方法和基于生成式对抗网络(GAN)的摩尔纹去除方法.在此基础上,选择相同的公开数据集,对主流的深度学习方法进行算法实现和性能对比分析,并分别总结了各类方法的优缺点.最后,对未来的研究方向进行展望.

基于深度学习的指纹室内定位对抗样本攻击研究

基于深度学习的指纹室内定位系统因其能够有效抽取接收信号强度(RSS)指纹数据的深层特征而大幅提高了室内定位性能,但该类方法需要大量多样化的RSS指纹数据训练模型,并且目前对其安全漏洞也缺乏充分的研究,这些安全漏洞源于无线Wi-Fi媒体的开放性和分类器的固有缺陷(如易遭受对抗性攻击等)。为此,对基于深度学习的RSS指纹室内定位系统的对抗性攻击进行研究,提出一种基于Wi-Fi指纹室内定位的对抗样本攻击框架,并利用该框架研究对抗攻击对基于深度学习的RSS指纹室内定位模型性能的影响。该框架包含离线训练和在线定位两个阶段。在离线训练阶段,设计适用于增广Wi-Fi RSS指纹数据的条件生成对抗网络(CGAN)来生成大量多样化的RSS指纹数据训练高鲁棒的室内定位深度学习模型;在线定位阶段,构造最强的一阶攻击策略来生成针对Wi-Fi RSS指纹室内定位系统的有效RSS对抗样本,研究对抗攻击对不同室内定位深度学习模型性能的影响。实验结果显示:在公开UJIIndoorLoc数据集上,由所提框架生成的RSS指纹对抗样本对卷积神经网络(CNN)、深度神经网络(DNN)、多层感知机(MLP)、pixeldp_CNN指纹室内定位模型的攻击成功率分别达到94.1%、63.75%、43.45%、72.5%;对由CGAN增广数据训练的上述4种指纹室内定位模型的攻击成功率分别达到84.95%、44.8%、15.7%、11.5%。由此表明,现有的基于深度学习的指纹室内定位模型易遭受对抗样本攻击的影响,由真实数据和增广数据混合训练的室内定位模型在面临对抗样本攻击时具有更好的鲁棒性。

复杂退化模型下图像超分辨率算法综述

图像的超分辨率(super-resolution,SR)一直以来是计算机视觉(computer vision,CV)领域的一项热门的研究方向,它旨在从单张或多张低分辨率图像中通过一系列的图像处理和深度学习技术,重建带有丰富边缘纹理等细节特征的高分辨率图像。自从深度卷积神经网络应用于图像超分辨率算法后,其性能相较于传统的基于重构和基于样例的SR算法有了非常大的提升。然而,目前的SR算法在实际场景应用、算法性能、模型质量评估标准等方面仍然需要改良和优化。因此,为推进图像超分辨率技术的发展,总结并分析了基于深度学习的SR算法。首先,将目前主流的SR算法分为基于卷积神经网络、基于生成对抗网络、基于Transformer这三类;其次,详细评述了每一类算法的网络结构、算法优缺点、算法特色及适用场景等;然后,对常见的超分辨率数据集及各种评价指标进行阐述,重点比较了不同SR算法在各类数据集上的性能;最后,总结了图像超分辨率目前研究所面临的问题并探讨了图像超分辨率的未来研究方向。

基于时序生成对抗网络的居民用户非侵入式负荷分解

现有的非侵入式负荷分解算法往往需要大量电器设备级的负荷数据才能保证分解精度,但由于用户对隐私性的考虑以及安装成本过高等问题,很难获取这些数据。因此,构建一种能深度挖掘电力负荷数据时序特性和电器相关性的时序生成对抗网络。利用降维网络对所有电器有功功率序列组成的高维向量进行降维以降低计算的复杂度,通过复原网络将结果还原为高维向量。基于电器运行状态和深度学习的非侵入式分解方法,运用卷积神经网络-双向门控循环单元构建状态复杂电器的负荷分解回归模型,对状态简单电器利用深度神经网络构建负荷识别分类模型。通过对比其他数据生成方法,以及改变典型公开数据集中生成数据比例所得的负荷分解结果验证了所提方法的有效性。

边缘对抗结合层次门控卷积的人脸修复研究

针对目前人脸缺损图像修复存在边缘模糊和失真等问题,提出了一种两阶段层次门控卷积网络(Hierarchical gated convolutional network,HGCN),并将其与边缘对抗网络相结合用于人脸图像修复。首先,采用边缘对抗网络生成边缘图像。其次,将边缘图、掩模和缺损图像作为输入,训练HGCN的GAN模型以修复缺损人脸图像。HGCN网络采用门控卷积取代传统卷积,并引入了扩张卷积,网络的主体由粗修复模块和精修复模块组成。在粗修复模块中,编码器和解码器网络结构用于粗修复;在精细修复模块中,引入注意力机制来增强特征提取能力,进一步细化修复结果。实验使用Celeba-HQ数据集和NVIDIA不规则掩码数据集作为训练数据集,采用门控卷积网络和注意力机制网络作为实验对比模型,PSNR、 SSIM和MAE作为实验评估指标。实验结果表明,对于缺损区域小于20%的人脸图像,所提出的网络在上述三个指标上优于两种比较网络,而对于缺失区域大于20%的图像,所提出的网络与两种比较方法性能指标接近。在视觉效果方面,所提出的方法在细节上也优于两种对比网络。因而,所提出的网络可以明显提高图像修复效果,尤其是对图像细节的修复效果。

变工况下动态卷积域对抗图神经网络故障诊断

针对基于无监督域自适应故障诊断方法忽略了域间数据结构信息、传统域对齐平均最大差异法全局泛化能力差等问题,本文提出一种基于无监督域自适应理论的动态卷积域对抗图神经网络故障诊断模型,分别通过对数据的类别标签、域标签和数据结构信息进行建模。通过分类器和域鉴别器分别建模类别标签和域标签,通过图神经网络将数据结构信息嵌入到实例图节点中,利用高斯Wasserstein距离来度量不同领域的实例图之间的差异。本文对比了不同工况下共14种迁移任务在各模型下故障识别的准确率。实验结果表明:基于动态卷积的域对抗图神经网络模型在变工况下的故障诊断效果均优于其他对比模型,且模型性能稳定。

融合迁移学习与CGAN的风电集群功率超短期预测

针对可再生能源不确定性导致电力系统消纳能力不足的问题,提出一种基于条件生成对抗网络与迁移学习融合的风电集群功率超短期预测方法。首先,分析了风电集群功率预测样本模式的不均衡性以及导致的神经网络预测误差偏移现象;其次,构建了条件生成对抗网络修复不均衡问题;最后,采用迁移学习结合时间卷积网络构建了风电集群功率超短期预测模型。测试结果表明,所提方法能够显著提高风电集群功率超短期预测精度。