基于深度学习的标准金属量器液位图像检测方法

传统基于灰度梯度分割的液位识别方法容易受到光照、清晰度等因素的影响,鲁棒性较低。为了解决这一问题,本文提出采用深度学习的液位图像检测算法;针对量器玻璃管液位线特点裁剪网络压缩卷积层,加快提取速度;使用K-means聚类设计先验框,增强尺度适应性。实验结果表明,基于YOLOv4的改进模型在液位线动态识别中,平均准确率mAP达到98.63%,帧处理速度达到了40fps。     

卷烟包装外观缺陷数据集构建及深度学习检测技术研究

目的 为了提升烟包缺陷检测的准确率,构建卷烟包装外观缺陷识别基准数据集,并开展主流深度学习模型在卷烟包装外观缺陷智能检测中的应用研究。方法 首先,从生产运行中的ZB45型细支烟硬盒包装机组采集缺陷图像,经过人工审核与筛选后获取典型的缺陷数据。然后,根据缺陷的特征与成因,将缺陷数据划分为23个类别,并逐一进行目标检测框标注。最终,形成了包含13 000余张缺陷图像的卷烟包装外观缺陷识别基准数据集,并针对烟包缺陷识别、缺陷分类、目标检测、模型迁移4项任务开展实验。结果 结果表明,数据集能够满足高准确率深度学习模型的训练需求;通过模型迁移,能够利用该数据集大幅提高不同牌号卷烟的缺陷检测效果;DenseNet模型在烟包缺陷识别与缺陷分类任务上表现较好,准确率分别达到93.70%和95.43%,YOLOv5模型在缺陷目标检测任务上mAP@0.5值达到了96.61%。结论 该数据集能够作为烟包缺陷检测领域的基准数据集,研究成果将进一步支撑卷烟包装领域的数据应用与数字化转型。     

Optimized Deep Learning Model for Fire Semantic Segmentation

Recent convolutional neural networks (CNNs) based deep learning has significantly promoted fire detection. Existing fire detection methods can efficiently recognize and locate the fire. However, the accurate flame boundary and shape information is hard to obtain by them, which makes it difficult to conduct automated fire region analysis, prediction, and early warning. To this end, we propose a fire semantic segmentation method based on Global Position Guidance (GPG) and Multi-path explicit Edge information Interaction (MEI). Specifically, to solve the problem of local segmentation errors in low-level feature space, a top-down global position guidance module is used to restrain the offset of low-level features. Besides, an MEI module is proposed to explicitly extract and utilize the edge information to refine the coarse fire segmentation results. We compare the proposed method with existing advanced semantic segmentation and salient object detection methods. Experimental results demonstrate that the proposed method achieves 94.1%, 93.6%, 94.6%, 95.3%, and 95.9% Intersection over Union (IoU) on five test sets respectively which outperforms the suboptimal method by a large margin. In addition, in terms of accuracy, our approach also achieves the best score.     

Enhance Intrusion Detection in Computer Networks Based on Deep Extreme Learning Machine

Networks provide a significant function in everyday life, and cybersecurity therefore developed a critical field of study. The Intrusion detection system (IDS) becoming an essential information protection strategy that tracks the situation of the software and hardware operating on the network. Notwithstanding advancements of growth, current intrusion detection systems also experience dif- ficulties in enhancing detection precision, growing false alarm levels and identifying suspicious activities. In order to address above mentioned issues, several researchers concentrated on designing intrusion detection systems that rely on machine learning approaches. Machine learning models will accurately identify the underlying variations among regular information and irregular information with incredible efficiency. Artificial intelligence, particularly machine learning methods can be used to develop an intelligent intrusion detection framework. There in this article in order to achieve this objective, we propose an intrusion detection system focused on a Deep extreme learning machine (DELM) which first establishes the assessment of safety features that lead to their prominence and then constructs an adaptive intrusion detection system focusing on the important features. In the moment, we researched the viability of our suggested DELMbased intrusion detection system by conducting dataset assessments and evaluating the performance factors to validate the system reliability. The experimental results illustrate that the suggested framework outclasses traditional algorithms. In fact, the suggested framework is not only of interest to scientific research but also of functional importance.     

基于深度学习耦合稀疏语义度量的商标检索算法

目的针对当前商标图像检索中的语义鸿沟问题,提出一种深度学习耦合稀疏语义度量的商标图像检索方案,有效抑制噪声干扰,降低冗余特征维数。方法首先,根据由卷积与池化组成的无监督学习机制,对输入商标图像进行多层特征提取,输出一维特征向量。随后,通过L2-支持向量机(L2-SVM)进行分类,利用特征向量进行训练,获得多级联特征。然后,根据商标图像的多级联特征和用户标签信息的异构数据结构,设计一种稀疏语义度量方法进行相似检索,减少语义鸿沟。此外,引入一种混合范数作为相似度量的稀疏约束,以抑制原始输入空间中的冗余特征维数和噪声,优化检索结果。结果实验表明,与当前流行的商标检索方案相比,所提算法具有更高的检索精度,其输出的结果中仅有1幅无关图像。结论该方案具有较高的检索精度和较强的鲁棒性,在商标检测、商标保护等方面中具有良好的应用价值。     

Deep Learning Enabled Object Detection and Tracking Model for Big Data Environment

Recently, big data becomes evitable due to massive increase in the generation of data in real time application. Presently, object detection and tracking applications becomes popular among research communities and finds useful in different applications namely vehicle navigation, augmented reality, surveillance, etc. This paper introduces an effective deep learning based object tracker using Automated Image Annotation with Inception v2 based Faster RCNN (AIA-IFRCNN) model in big data environment. The AIA-IFRCNN model annotates the images by Discriminative Correlation Filter (DCF) with Channel and Spatial Reliability tracker (CSR), named DCF-CSRT model. The AIA-IFRCNN technique employs Faster RCNN for object detection and tracking, which comprises region proposal network (RPN) and Fast R-CNN. In addition, inception v2 model is applied as a shared convolution neural network (CNN) to generate the feature map. Lastly, softmax layer is applied to perform classification task. The effectiveness of the AIA-IFRCNN method undergoes experimentation against a benchmark dataset and the results are assessed under diverse aspects with maximum detection accuracy of 97.77%.     

Research on Privacy Disclosure Detection Method in Social Networks Based on Multi-Dimensional Deep Learning

In order to effectively detect the privacy that may be leaked through socialnetworks and avoid unnecessary harm to users, this paper takes microblog as the researchobject to study the detection of privacy disclosure in social networks. First, we performfast privacy leak detection on the currently published text based on the fastText model. Inthe case that the text to be published contains certain private information, we fullyconsider the aggregation effect of the private information leaked by different channels,and establish a convolution neural network model based on multi-dimensional features(MF-CNN) to detect privacy disclosure comprehensively and accurately. Theexperimental results show that the proposed method has a higher accuracy of privacydisclosure detection and can meet the real-time requirements of detection.     

Optimal Deep Reinforcement Learning for Intrusion Detection in UAVs

In recent years, progressive developments have been observed in recent technologies and the production cost has been continuously decreasing. In such scenario, Internet of Things (IoT) network which is comprised of a set of Unmanned Aerial Vehicles (UAV), has received more attention from civilian to military applications. But network security poses a serious challenge to UAV networks whereas the intrusion detection system (IDS) is found to be an effective process to secure the UAV networks. Classical IDSs are not adequate to handle the latest computer networks that possess maximum bandwidth and data traffic. In order to improve the detection performance and reduce the false alarms generated by IDS, several researchers have employed Machine Learning (ML) and Deep Learning (DL) algorithms to address the intrusion detection problem. In this view, the current research article presents a deep reinforcement learning technique, optimized by Black Widow Optimization (DRL-BWO) algorithm, for UAV networks. In addition, DRL involves an improved reinforcement learning-based Deep Belief Network (DBN) for intrusion detection. For parameter optimization of DRL technique, BWO algorithm is applied. It helps in improving the intrusion detection performance of UAV networks. An extensive set of experimental analysis was performed to highlight the supremacy of the proposed model. From the simulation values, it is evident that the proposed method is appropriate as it attained high precision, recall, F-measure, and accuracy values such as 0.985, 0.993, 0.988, and 0.989 respectively.     

基于深度学习的易拉罐缺陷检测技术

目的 现有的易拉罐缺陷检测系统在高速生产线中存在错检率和漏检率高,检测精度相对较低等问题,为了提高易拉罐缺陷识别的准确性,使易拉罐生产线实现进一步自动化、智能化,基于深度学习技术和迁移学习技术,提出一种适用于易拉罐制造的在线检测的算法。方法 利用深度卷积网络提取易拉罐缺陷特征,通过优化卷积核,减短易拉罐缺陷检测的时间。针对国内外数据集缺乏食品包装制造的缺陷图像,构建易拉罐缺陷数据集,结合预训练网络,通过调整VGG16提升对易拉罐缺陷的识别准确率。结果 对易拉罐数据集在卷积神经网络、迁移学习和调整后的预训练网络进行了易拉罐缺陷检测的性能对比,验证了基于深度学习的易拉罐缺陷检测技术在学习率为0.0005,训练10个迭代后可达到较好的识别效果,最终二分类缺陷识别率为99.7%,算法耗时119 ms。结论 相较于现有的易拉罐检测算法,文中提出的基于深度学习的易拉罐检测算法的识别性能更优,智能化程度更高。同时,该研究有助于制罐企业利用深度学习等AI技术促进智能化生产,减少人力成本,符合国家制造业产业升级的策略,具有一定的实际意义。     

Deep Learning Approach for Cosmetic Product Detection and Classification

As the amount of online video content is increasing, consumers are becoming increasingly interested in various product names appearing in videos, particularly in cosmetic-product names in videos related to fashion, beauty, and style. Thus, the identification of such products by using image recognition technology may aid in the identification of current commercial trends. In this paper, we propose a two-stage deep-learning detection and classification method for cosmetic products. Specifically, variants of the YOLO network are used for detection, where the bounding box for each given input product is predicted and subsequently cropped for classification. We use four state-of-the-art classification networks, namely ResNet, InceptionResNetV2, DenseNet, and EfficientNet, and compare their performance. Furthermore, we employ dilated convolution in these networks to obtain better feature representations and improve performance. Extensive experiments demonstrate that YOLOv3 and its tiny version achieve higher speed and accuracy. Moreover, the dilated networks marginally outperform the base models, or achieve similar performance in the worst case. We conclude that the proposed method can effectively detect and classify cosmetic products.     

Deep Learning and Machine Learning for Early Detection of Stroke and Haemorrhage

Stroke and cerebral haemorrhage are the second leading causes of death in the world after ischaemic heart disease. In this work, a dataset containing medical, physiological and environmental tests for stroke was used to evaluate the efficacy of machine learning, deep learning and a hybrid technique between deep learning and machine learning on the Magnetic Resonance Imaging (MRI) dataset for cerebral haemorrhage. In the first dataset (medical records), two features, namely, diabetes and obesity, were created on the basis of the values of the corresponding features. The t-Distributed Stochastic Neighbour Embedding algorithm was applied to represent the high-dimensional dataset in a low-dimensional data space. Meanwhile,the Recursive Feature Elimination algorithm (RFE) was applied to rank the features according to priority and their correlation to the target feature and to remove the unimportant features. The features are fed into the various classification algorithms, namely, Support Vector Machine (SVM), K Nearest Neighbours (KNN), Decision Tree, Random Forest, and Multilayer Perceptron. All algorithms achieved superior results. The Random Forest algorithm achieved the best performance amongst the algorithms; it reached an overall accuracy of 99%. This algorithm classified stroke cases with Precision, Recall and F1 score of 98%, 100% and 99%, respectively. In the second dataset, the MRI image dataset was evaluated by using the AlexNet model and AlexNet + SVM hybrid technique. The hybrid model AlexNet + SVM performed is better than the AlexNet model; it reached accuracy, sensitivity, specificity and Area Under the Curve (AUC) of 99.9%, 100%, 99.80% and 99.86%, respectively.     

Multimodal Fused Deep Learning Networks for Domain Specific Image Similarity Search

The exponential increase in data over the past few years, particularly in images, has led to more complex content since visual representation became the new norm. E-commerce and similar platforms maintain large image catalogues of their products. In image databases, searching and retrieving similar images is still a challenge, even though several image retrieval techniques have been proposed over the decade. Most of these techniques work well when querying general image databases. However, they often fail in domain-specific image databases, especially for datasets with low intraclass variance. This paper proposes a domain-specific image similarity search engine based on a fused deep learning network. The network is comprised of an improved object localization module, a classification module to narrow down search options and finally a feature extraction and similarity calculation module. The network features both an offline stage for indexing the dataset and an online stage for querying. The dataset used to evaluate the performance of the proposed network is a custom domain-specific dataset related to cosmetics packaging gathered from various online platforms. The proposed method addresses the intraclass variance problem with more precise object localization and the introduction of top result reranking based on object contours. Finally, quantitative and qualitative experiment results are presented, showing improved image similarity search performance.     

Efficient Deep Learning Modalities for Object Detection from Infrared Images

For military warfare purposes, it is necessary to identify the type of a certain weapon through video stream tracking based on infrared (IR) video frames. Computer vision is a visual search trend that is used to identify objects in images or video frames. For military applications, drones take a main role in surveillance tasks, but they cannot be confident for long-time missions. So, there is a need for such a system, which provides a continuous surveillance task to support the drone mission. Such a system can be called a Hybrid Surveillance System (HSS). This system is based on a distributed network of wireless sensors for continuous surveillance. In addition, it includes one or more drones to make short-time missions, if the sensors detect a suspicious event. This paper presents a digital solution to identify certain types of concealed weapons in surveillance applications based on Convolutional Neural Networks (CNNs) and Convolutional Long Short-Term Memory (ConvLSTM). Based on initial results, the importance of video frame enhancement is obvious to improve the visibility of objects in video streams. The accuracy of the proposed methods reach 99%, which reflects the effectiveness of the presented solution. In addition, the experimental results prove that the proposed methods provide superior performance compared to traditional ones.     

3D Head Pose Estimation through Facial Features and Deep Convolutional Neural Networks

Face image analysis is one among several important cues in computer vision. Over the last five decades, methods for face analysis have received immense attention due to large scale applications in various face analysis tasks. Face parsing strongly benefits various human face image analysis tasks inducing face pose estimation. In this paper we propose a 3D head pose estimation framework developed through a prior end to end deep face parsing model. We have developed an end to end face parts segmentation framework through deep convolutional neural networks (DCNNs). For training a deep face parts parsing model, we label face images for seven different classes, including eyes, brows, nose, hair, mouth, skin, and back. We extract features from gray scale images by using DCNNs. We train a classifier using the extracted features. We use the probabilistic classification method to produce gray scale images in the form of probability maps for each dense semantic class. We use a next stage of DCNNs and extract features from grayscale images created as probability maps during the segmentation phase. We assess the performance of our newly proposed model on four standard head pose datasets, including Pointing’04, Annotated Facial Landmarks in the Wild (AFLW), Boston University (BU), and ICT-3DHP, obtaining superior results as compared to previous results.     

Deep Transfer Learning Based Detection and Classification of Citrus Plant Diseases

Citrus fruit crops are among the world’s most important agricultural products, but pests and diseases impact their cultivation, resulting in yield and quality losses. Computer vision and machine learning have been widely used to detect and classify plant diseases over the last decade, allowing for early disease detection and improving agricultural production. This paper presented an automatic system for the early detection and classification of citrus plant diseases based on a deep learning (DL) model, which improved accuracy while decreasing computational complexity. The most recent transfer learning-based models were applied to the Citrus Plant Dataset to improve classification accuracy. Using transfer learning, this study successfully proposed a Convolutional Neural Network (CNN)-based pre-trained model (EfficientNetB3, ResNet50, MobiNetV2, and InceptionV3) for the identification and categorization of citrus plant diseases. To evaluate the architecture’s performance, this study discovered that transferring an EfficientNetb3 model resulted in the highest training, validating, and testing accuracies, which were 99.43%, 99.48%, and 99.58%, respectively. In identifying and categorizing citrus plant diseases, the proposed CNN model outperforms other cutting-edge CNN model architectures developed previously in the literature.     

基于YOLOv3的医药玻璃瓶缺陷检测方法

目的为了准确而快速地自动检测医药玻璃瓶的外观缺陷。方法基于YOLOv3算法,建立深度卷积神经网络学习检测模型,通过将神经网络结构浅层和深层特征向量连接并标准化,提取玻璃瓶多尺度特征信息。采用K-means聚类方式获得锚点框初始大小,提高模型对边界框预测的准确性;利用多尺度训练策略,增强模型对不同尺寸图像的鲁棒性。结果实验结果表明,提出的医药玻璃瓶缺陷检测方法能够准确检测识别玻璃瓶上的管端残损、气线、气泡、划伤、污渍和结石等缺陷种类。与主流的目标检测方法相比,提出的方法在处理速度和准确度上都有提高,缺陷目标检测精确率达到96.23%,召回率为93.82%,平均精度为89.35%。结论该方法已经在国内几家大型医药玻璃包装生产公司成功应用,显著提高了医药玻璃包装产品的质量和合格率,降低了人工成本。     

Enhanced Deep Autoencoder Based Feature Representation Learning for Intelligent Intrusion Detection System

In the era of Big data, learning discriminant feature representation from network traffic is identified has as an invariably essential task for improving the detection ability of an intrusion detection system (IDS). Owing to the lack of accurately labeled network traffic data, many unsupervised feature representation learning models have been proposed with state-of-the-art performance. Yet, these models fail to consider the classification error while learning the feature representation. Intuitively, the learnt feature representation may degrade the performance of the classification task. For the first time in the field of intrusion detection, this paper proposes an unsupervised IDS model leveraging the benefits of deep autoencoder (DAE) for learning the robust feature representation and one-class support vector machine (OCSVM) for finding the more compact decision hyperplane for intrusion detection. Specially, the proposed model defines a new unified objective function to minimize the reconstruction and classification error simultaneously. This unique contribution not only enables the model to support joint learning for feature representation and classifier training but also guides to learn the robust feature representation which can improve the discrimination ability of the classifier for intrusion detection. Three set of evaluation experiments are conducted to demonstrate the potential of the proposed model. First, the ablation evaluation on benchmark dataset, NSL-KDD validates the design decision of the proposed model. Next, the performance evaluation on recent intrusion dataset, UNSW-NB15 signifies the stable performance of the proposed model. Finally, the comparative evaluation verifies the efficacy of the proposed model against recently published state-of-the-art methods.     

Text Detection and Recognition for Natural Scene Images Using Deep Convolutional Neural Networks

Words are the most indispensable information in human life. It is very important to analyze and understand the meaning of words. Compared with the general visual elements, the text conveys rich and high-level moral information, which enables the computer to better understand the semantic content of the text. With the rapid development of computer technology, great achievements have been made in text information detection and recognition. However, when dealing with text characters in natural scene images, there are still some limitations in the detection and recognition of natural scene images. Because natural scene image has more interference and complexity than text, these factors make the detection and recognition of natural scene image text face many challenges. To solve this problem, a new text detection and recognition method based on depth convolution neural network is proposed for natural scene image in this paper. In text detection, this method obtains high-level visual features from the bottom pixels by ResNet network, and extracts the context features from character sequences by BLSTM layer, then introduce to the idea of faster R-CNN vertical anchor point to find the bounding box of the detected text, which effectively improves the effect of text object detection. In addition, in text recognition task, DenseNet model is used to construct character recognition based on Kares. Finally, the output of Softmax is used to classify each character. Our method can replace the artificially defined features with automatic learning and context-based features. It improves the efficiency and accuracy of recognition, and realizes text detection and recognition of natural scene images. And on the PAC2018 competition platform, the experimental results have achieved good results.     

A Novel Framework for Windows Malware Detection Using a Deep Learning Approach

Malicious software (malware) is one of the main cyber threats that organizations and Internet users are currently facing. Malware is a software code developed by cybercriminals for damage purposes, such as corrupting the system and data as well as stealing sensitive data. The damage caused by malware is substantially increasing every day. There is a need to detect malware efficiently and automatically and remove threats quickly from the systems. Although there are various approaches to tackle malware problems, their prevalence and stealthiness necessitate an effective method for the detection and prevention of malware attacks. The deep learning-based approach is recently gaining attention as a suitable method that effectively detects malware. In this paper, a novel approach based on deep learning for detecting malware proposed. Furthermore, the proposed approach deploys novel feature selection, feature co-relation, and feature representations to significantly reduce the feature space. The proposed approach has been evaluated using a Microsoft prediction dataset with samples of 21,736 malware composed of 9 malware families. It achieved 96.01% accuracy and outperformed the existing techniques of malware detection.