Search

In this paper, a high accuracy and reduced processing time block based classification method for computer screen images is presented. This method classifies blocks into five types: smooth, sparse, fuzzy, text and picture blocks. In a computer screen compression application, the choice of block compression algorithm is made based on these block types. The classification method presented has four novel features. The first novel feature is a combination of Discrete Wavelet Transform (DWT) and colour counting classification methods. Both of these methods have only been used for computer image compression in isolation in previous publications but this paper shows that combined together more accurate results are obtained overall. The second novel feature is the classification of the image blocks into five block types. The addition of the fuzzy and sparse block types make the use of optimum compression methods possible for these blocks. The third novel feature is block type prediction. The prediction algorithm is applied to a current block when the blocks on the top and the left of the current block are text blocks or smooth blocks. This new algorithm is designed to exploit the correlation of adjacent blocks and reduces the overall classification processing time by 33%. The fourth novel feature is down sampling of the pixels in each block which reduces the classification processing time by 62%. When both block prediction and down sampling are enabled, the classification time is reduced by 74% overall. The overall classification accuracy is 98.46%.

The key to an accurate understanding of terrain is to extract the informative features from the multi-modal data obtained from different devices. Sensors, such as RGB cameras, depth sensors, vibration sensors, and microphones, are used as the multi-modal data. Many studies have explored ways to use them, especially in the robotics field. Some papers have successfully introduced single-modal or multi-modal methods. However, in practice, robots can be faced with extreme conditions; microphones do not work well in crowded scenes, and an RGB camera cannot capture terrains well in the dark. In this paper, we present a novel framework using the multi-modal variational autoencoder and the Gaussian mixture model clustering algorithm on image data and audio data for terrain type clustering by forcing the features to be closer together in the feature space. Our method enables the terrain type clustering even if one of the modalities (either image or audio) is missing at the test-time. We evaluated the clustering accuracy with a conventional multi-modal terrain type clustering method and we conducted ablation studies to show the effectiveness of our approach.

Deep convolutional neural networks (DCNNs) have attracted extensive attention in synthetic aperture radar (SAR) image target detection. However, there is usually a lack of labeled data in SAR images which results in difficulties to improve the detection performance with DCNN based methods. Transfer learning borrows generic knowledge from other tasks for the target task and can be an effective solution to the small sample problem, but the features’ generality and specificity across various models are not discussed in target detection. In addition, conventional DCNN based models have too many parameters related to the prior information of the targets which increase the difficulty of transfer learning between different tasks. In this paper, we mainly conduct research in three folds to address these mentioned issues. First, the generality and specificity of features in different network models are analyzed in the field of SAR target detection. Second, a transitive transfer framework is developed for detection under small sample condition. Third, an anchor free rotatable detector with flexible structure is designed for SAR images detection to suit the transfer detection framework proposed in this paper and a reasonable explanation is given for its upper hand with small samples. Experimental results demonstrate that: 1) transitive transfer framework is valid in SAR target detection; 2) the anchor free detector in this framework is superior to the anchor based one; 3) the proposed detector achieves better performance under small sample condition by comparison with state-of-the-art methods.

Anomaly detection has become an indispensable tool for modern society, applied in a wide range of applications, from detecting fraudulent transactions to malignant brain tumours. Over time, many anomaly detection techniques have been introduced. However, in general, they all suffer from the same problem: a lack of data that represents anomalous behaviour. As anomalous behaviour is usually costly (or dangerous) for a system, it is difficult to gather enough data that represents such behaviour. This, in turn, makes it difficult to develop and evaluate anomaly detection techniques. Recently, generative adversarial networks (GANs) have attracted a great deal of attention in anomaly detection research, due to their unique ability to generate new data. In this paper, we present a systematic literature review of the applications of GANs in anomaly detection, covering 128 papers on the subject. The goal of this review paper is to analyze and summarize: (1) which anomaly detection techniques can benefit from certain types of GANs, and how, (2) in which application domains GAN-assisted anomaly detection techniques have been applied, and (3) which datasets and performance metrics have been used to evaluate these techniques. Our study helps researchers and practitioners to find the most suitable GAN-assisted anomaly detection technique for their application. In addition, we present a research roadmap for future studies in this area., Comment: 34 pages,5 figures

The new coronavirus, which has become a global pandemic, has confirmed more than 88 million cases worldwide since the first case was recorded in December 2019, causing over 1.9 million deaths. Since COIVD-19 lesions have clear imaging features on CT images, it is suitable for the auxiliary diagnosis and treatment of COVID-19. Deep learning can be used to segment the lesions areas of COVID-19 in CT images to help monitor the epidemic situation. In this paper, we propose a multi-point supervision network (MPS-Net) for segmentation of COVID-19 lung infection CT image lesions to solve the problem of a variety of lesion shapes and areas. A multi-scale feature extraction structure, a sieve connection structure (SC), a multi-scale input structure and a multi-point supervised training structure were implemented into MPS-Net. In order to increase the ability to segment various lesion areas of different sizes, the multi-scale feature extraction structure and the sieve connection structure will use different sizes of receptive fields to extract feature maps of various scales. The multi-scale input structure is used to minimize the edge loss caused by the convolution process. In order to improve the accuracy of segmentation, we propose a multi-point supervision training structure to extract supervision signals from different up-sampling points on the network. Experimental results showed that the dice similarity coefficient (DSC), sensitivity, specificity and IOU of the segmentation results of our model are 0.8325, 0.8406, 09988 and 0.742, respectively. The experimental results demonstrated that the network proposed in this paper can effectively segment COVID-19 infection on CT images. It can be used to assist the diagnosis and treatment of new coronary pneumonia.

In recent years, a variety of supervised manifold learning techniques have been proposed to outperform their unsupervised alternative versions in terms of classification accuracy and data structure capturing. Some dissimilarity measures have been used in these techniques to guide the dimensionality reduction process. Their good performance was empirically demonstrated; however, the relevant analysis is still missing. This paper contributes to a theoretical analysis on a) how dissimilarity measures affect maintaining manifold neighbourhood structure and b) how supervised manifold learning techniques could contribute to the reduction of classification error. This paper also provides a cross-comparison between supervised and unsupervised manifold learning approaches in terms of structure capturing using Kendall’s Tau coefficients and co-ranking matrices. Four different metrics (including three dissimilarity measures and Euclidean distance) have been considered along with manifold learning methods such as Isomap, ${t}$ -Stochastic Neighbour Embedding ( ${t}$ -SNE), and Laplacian Eigenmaps (LE), in two datasets: Breast Cancer and Swiss-Roll. This paper concludes that although the dissimilarity measures used in the manifold learning techniques can reduce classification error, they do not learn well or preserve the structure of the hidden manifold in the high dimensional space, but instead, they destroy the structure of the data. Based on the findings of this paper, it is advisable to use supervised manifold learning techniques as a pre-processing step in classification. In addition, it is not advisable to apply supervised manifold learning for visualization purposes since the two-dimensional representation using supervised manifold learning does not improve the preservation of data structure.

The epidemic increase in online reviews’ growth made the sentiment classification a fascinating domain in academic and industrial research. The reviews assist several domains, which is complicated to gather annotated training data. Several sentiment classification methodologies are devised for performing the sentiment analysis, but retrieval of information is not accurately performed, less effective, and less convergence speed. In this paper, we propose a sentiment paper proposes a sentiment classification model, namely Spider Monkey Crow Optimization algorithm (SMCA), for training the deep recurrent neural network (DeepRNN). In this method, the telecom review is employed to remove stop words and stemming to eliminate inappropriate data to minimize user’s seeking time. Meanwhile, the feature extraction is performed using SentiWordNet to derive the sentiments from the reviews. The extracted SentiWordNet features and other features, like elongated words, punctuation, hashtag, and numerical values, are employed in the DeepRNN for classifying sentiments. To retrieve the required review, the Fuzzy K-Nearest neighbor (Fuzzy-KNN) is employed to retrieve the review based on a distance measure. With rigorous assessments and experimentation, it is observed that the proposed SMCA-based DeepRNN performs better in terms of accuracy of 97.7%, precision of 95.5%, recall of 94.6%, and F1-score 96.7%, respectively.

In order to solve the problems of unbalanced sample data and the lack of consideration of temporal information in existing Siamese-based trackers, this paper proposes a Siamese recurrent neural network and region proposal network (Siamese R-RPN), which can be trained in an end-to-end manner. Siamese R-RPN is consisted of Siamese network, recurrent neural network and region proposal network. Image features extracted by the Siamese network are strengthened by the channel and spatial attention mechanisms, and are sent to the RPN for classification and regression. Temporal information is processed by a recurrent neural network-based Long Short-Term Memory (LSTM) to predict the rough location of the target, it is mapped to the anchor feature map of the RPN for anchor selection. This makes the positive and negative samples participating in the training procedure to become more balanced and representative. Because of the collaborative use of temporal and spatial information, the tracker proposed in this paper has achieved state-of-the-art performance on three large tracking benchmarks—OTB 2015, VOT2016 and VOT 2018—where this verifies its effectiveness.

Several hypovigilance detection systems (HDx) were developed to avoid road-side accidents due to driver fatigue. They have suffered from several limitations. Notably many of these are focused on center-head position to define an area of interest (often referred to as PERCLOS (percentage eye closure)) without considering the face occlusion problem, light illumination, and suffer poor response time. These HDx systems mostly depend on image processing, vision-based, and multisensor-based features. To address these problems, the author utilized vision, sensors, environmental, and vehicular-based features that integrated together by fusion to predict multistage of HDx. Lately, few studies have utilized the combination of multimodal features and deep learning (DL) architectures. Those multimodal-based features (M-HDx) were feasible to predict stages of driver fatigue (multi-stage). However, there is a need to critically measure the performance of these M-HDx by carrying out a comparative analysis to recognize multi-stage of fatigue in terms of hardware-based benchmarks. Moreover, it is important to evaluate the M-HDx systems using different features-set with respect to traditional and advanced machine learning techniques. Therefore, the primary aim of this work is in algorithm and feature modeling, then compare the advantages and differences with other work. In this paper, a different study is conducted compare to state-of-the-art survey articles by statistically measuring the performance. After experiments on M-HDx systems, this paper concludes that there is still a research gap to real-time development of multistage M-HDx systems. In the end, the paper summarizes the directions, challenges, and applications in the development of HDx systems to assist other researchers for further research.

Terrain generation aims to automatize the procedure of creating landscapes using a computer system. The generation models must follow different terrains’ topographical features, such as areas with river deltas and other regions where water bodies affect the natural landscapes. It is possible to generate more realistic terrains thanks to improvements in computer graphics techniques and deep learning models that use specific hardware. However, the advance on the generation of terrains that include river deltas, fjords, and waterfalls has not had the same pace as other more studied landscapes. Therefore, as a contribution to the advance of the research of terrain generation with water bodies using generative models, this paper presents the DRCA2020 dataset, which is useful for supervised training. The proposed dataset contains eight different types of real-world satellite images. These images are grouped by the same geographical location. There are 13,184 groups; each one has three RGB surface images, a water coverage map, three binarizations of water coverage, and a digital elevation model (DEM). Additionally, this paper proposes the use of a cGAN composite model, trained with the DRCA2020 dataset, for generating synthetic DEMs from water coverage images and therefore, to create realistic surface texture images with promising validation results.

Neuromorphic vision sensors provide low power sensing and capture salient spatial-temporal events. The majority of the existing neuromorphic sensing work focus on object detection. However, since they only record the events, they provide an efficient signal domain for privacy aware surveillance tasks. This paper explores how the neuromorphic vision sensor data streams can be analysed for human action recognition, which is a challenging application. The proposed method is based on handcrafted features. It consists of a pre-processing step for removing the noisy events followed by the extraction of handcrafted local and global feature vectors corresponding to the underlying human action. The local features are extracted considering a set of high-order descriptive statistics from the spatio-temporal events in a time window slice, while the global features are extracted by considering the frequencies of occurrences of the temporal event sequences. Then, low complexity classifiers, such as, support vector machines (SVM) and K-Nearest Neighbours (KNNs), are trained using these feature vectors. The proposed method evaluation uses three groups of datasets: Emulator-based, re-recording-based and native NVS-based. The proposed method has outperformed the existing methods in terms of human action recognition accuracy rates by 0.54%, 19.3%, and 25.61% for E-KTH, E-UCF11 and E-HMDB51 datasets, respectively. This paper also reports results for three further datasets: E-UCF50, R-UCF50, and N-Actions, which are reported for the first time for human action recognition on neuromorphic vision sensor domain.

The increasing significance of technology in daily lives led to the need for the development of convenient methods of human-computer interaction (HCI). Given that the existing HCI approaches exhibit various limitations, hand gesture recognition-based HCI may serve as a more intuitive mode of human-machine interaction in many situations. In addition, the system has to be deployable on low-power devices for applicability in broadly defined Internet of Things (IoT) and smart home solutions. Recent advances exhibit the potential of deep learning models for gesture classification, whereas they are still limited to high-performance hardware. Embedded neural network accelerators are constrained in terms of available memory, central processing unit (CPU) clock speed, graphics processing unit (GPU) performance, and a number of supported operations. The aforementioned problems are addressed in this paper by namely two approaches - simplifying the signal processing pipeline to avoid recurrent structures and efficient topological design. This paper employs an intuitive scheme allowing for the generation of the data in the compressed form from the sequence of range-Doppler images (RDI). Thus, it allows for the design of a neural classifier avoiding the usage of recurrent layers. The proposed framework has been optimized for Intel® Neural Compute Stick 2 (Intel® NCS 2), at the same time achieving promising classification accuracy of 97.57%. To confirm the robustness of the proposed algorithm, five independent persons have been involved in the algorithm testing process.

Research on the particle size of blast piles has always been an essential issue in mining engineering. Reasonable blasting parameters can reduce mining costs and reduce the workload of secondary crushing, which can significantly improve mining efficiency. The usual particle size analysis methods include the sieving method, the large particle size statistical method and other manual measurement methods. Nevertheless, these methods have the disadvantages of a high labor intensity, low efficiency and low precision. This paper analyzes UAV image information based on the single-picture photogrammetry method of computer image processing technology. A two-dimensional empirical wavelet transform (EWT) is used for image noise reduction. The nonlocal multiscale fractional differential (NMFD) enhances the texture of dark images and uses superpixel image segmentation technology so that the processed image can meet the granularity statistical study requirements of blast piles. The research results show that the accuracy of the ore particle size distribution by the method proposed in this paper is more than 90%.

The standardization and digitalization of nine constitutions of traditional Chinese medicine (TCM) have promoted the development of TCM automation. The keys to constitution identification are tongue diagnosis and consultation diagnosis. In this paper, a tongue detection method based on the combination of the histogram of oriented gradients (HOG) and the support vector machine (SVM) is proposed. To separate the tongue body and tongue coating, a k-means segmentation method based on the Lab color space is proposed. Based on the clustering analysis of the difference between the color components of the tongue body and tongue coating in the Lab color space, the separation is realized. Thus, the relationship between the tongue image and constitution can be analyzed. The method for consultation diagnosis is divided into a questionnaire scale and a question answering system. The questionnaire includes 29 quantified items with corresponding weights and answer scores, so the final score of each type of constitution is calculated. The final constitution type is determined on the basis of the scores of both diagnoses. In addition, the key of question answering system is text similarity calculation. First, a sentence similarity calculation method combining the n-gram language model and word2vec is proposed. Then, a multifeature fusion sentence similarity calculation method combining the Word Mover’s Distance (WMD) and editing distance is proposed. Finally, the experimental results show the effectiveness of the proposed method, and a constitution identification application is developed based on the content of this paper.

Stress is an escalated psycho-physiological state of the human body emerging in response to a challenging event or a demanding condition. Environmental factors that trigger stress are called stressors. In case of prolonged exposure to multiple stressors impacting simultaneously, a person’s mental and physical health can be adversely affected which can further lead to chronic health issues. To prevent stress-related issues, it is necessary to detect them in the nascent stages which are possible only by continuous monitoring of stress. Wearable devices promise real-time and continuous data collection, which helps in personal stress monitoring. In this paper, a comprehensive review has been presented, which focuses on stress detection using wearable sensors and applied machine learning techniques. This paper investigates the stress detection approaches adopted in accordance with the sensory devices such as wearable sensors, Electrocardiogram (ECG), Electroencephalography (EEG), and Photoplethysmography (PPG), and also depending on various environments like during driving, studying, and working. The stressors, techniques, results, advantages, limitations, and issues for each study are highlighted and expected to provide a path for future research studies. Also, a multimodal stress detection system using a wearable sensor-based deep learning technique has been proposed at the end.

The application of Artificial Intelligence (AI) to financial investment is a research area that has attracted extensive research attention since the 1990s, when there was an accelerated technological development and popularization of the personal computer. Since then, countless approaches have been proposed to deal with the problem of price prediction in the stock market. This paper presents a systematic review of the literature on Artificial Intelligence applied to investments in the stock market based on a sample of 2326 papers from the Scopus website between 1995 and 2019. These papers were divided into four categories: portfolio optimization, stock market prediction using AI, financial sentiment analysis, and combinations involving two or more approaches. For each category, the initial introductory research to its state-of-the-art applications are described. In addition, an overview of the review leads to the conclusion that this research area is gaining continuous attention and the literature is becoming increasingly specific and thorough.

Aiming at the difference in density between the light guide point distribution of the light guide plate (LGP) images, the different size, shape and brightness of LGP defects, and the limitation of a small number of defect samples, our paper proposes a two-stage multiscale residual attention network based on “segmentation + decision” for LGP defect detection. In the first stage, a segmentation subnet is constructed to achieve accurate location of multiscale defects by using the U-shaped structure and the designed Multiscale Residual Attention Unit (MRAU). In the second stage, a decision subnet is constructed to achieve accurate decision whether the LGP image is defective under the guidance of features extracted by segmentation subnet. The design of the network enables the model to be trained using a small number of samples, which is a significant requirement for practical applications. Finally, a lot of experimental studies were conducted on two datasets, including compared with related deep learning methods, proving the effectiveness of residual attention mechanism and proving the effectiveness of multiscale mechanism. Experimental results show that the detection method in this paper has advantages in both detection accuracy and universality under the limitation of a small number of samples. In particular, the detection indicator F1-Score respectively reached 99.85% and 95.46% on the two datasets. Combined with an image acquisition platform, a surface defect detection system is implemented to realize the real-time production quality inspection on the production lines.

Polarimetric synthetic aperture radar (PolSAR) image classification has become a hot research topic in recent years. Sparse representation plays an important role in image processing. However, almost all the existing dictionary learning methods are linear transformation in the original data space, so they cannot capture the nonlinear relationship of the input data. The recently proposed projective dictionary pair learning (DPL) method has acquired good performance in classification result and time consumption. In this paper, we propose the nonlinear projective dictionary pair learning (NDPL) model, which introduced the nonlinear transformation to the DPL model. Our method can adaptively obtain the nonlinear relationship between the elements of input data, and it also has the excellent performance of DPL model. In this paper, we use three PolSAR images to test the performance of our proposed method. Compared with several state-of-the-art methods, our proposed method has obtained promising results in solving the task of PolSAR image classification.

Semantic segmentation is a task that covers most of the perception needs of intelligent vehicles in an unified way. Recent studies witnessed that attention mechanisms achieve impressive performance in computer vision task. Current attention mechanisms based segmentation methods differ with each other in position and form of the attention mechanism, and perform differently in practice. This paper firstly introduces the effectiveness of multi-scale context features and attention mechanisms in segmentation tasks. We find that multi-scale and channel attention can play a vital role in constructing effective context features. Based on this analysis, this paper proposes an efficient attention pyramid network (EAPNet) for semantic segmentation. Specifically, to efficient handle the problem of segmenting objects at multiple scales, we design efficient channel attention pyramid (ECAP) which employ atrous convolution with channel attention in cascade or in parallel to capture multi-scale context by using multiple atrous rates. Furthermore, we propose a residual attention fusion block (RAFB), whose purpose is to simultaneously focus on meaningful low-level feature maps and spatial location information. At the same time, we will explore different channel attention modules and spatial attention modules, and describe their impact on network performance. We empirically evaluate our EAPNet on two semantic segmentation datasets, including PASCAL VOC 2012 and Cityscapes datasets. Experimental results show that without MS COCO pre-training and any post-processing, EAPNet achieved 81.7% mIoU on the PASCAL VOC 2012 validation set. With deeplabv3+ as the benchmark, EAPNet improve the model performance of more than 1.50% mIoU.

In order to realize the automatic detection of surface defects of lithium battery pole piece, a method for detection and identification of surface defects of lithium battery pole piece based on multi-feature fusion and PSO-SVM was proposed in this paper. Firstly, image subtraction and contrast adjustment were used to preprocess the defect image to weaken the influence of non-defective areas and enhance the defect features. Then, Canny algorithm and the AND logical operation were used to extract the image of defect area. Next, the texture feature, edge feature, and HOG feature were combined to extract the feature of the defect area image. Finally, the support vector machine (SVM) optimized by particle swarm optimization (PSO) was used to automatically identify and classify defect images. The experimental results show that the proposed method in this paper can effectively detect surface multiple types defects of lithium battery pole piece, and the average recognition rate of defects reaches 98.3%, which is an effective and feasible automatic defect detection and identification method.

This paper attempts to analyze the Korean sentence classification system. Sentence classification is the task of classifying an input sentence based on predefined categories. However, spelling or space error contained in the input sentence causes problems in morphological analysis and tokenization. This paper proposes a novel approach of Integrated Eojeol (Korean syntactic word separated by space) Embedding to reduce the effect of poorly analyzed morphemes on sentence classification. The paper also proposes two noise insertion methods that further improve classification performance. Our evaluation results indicate that by applying the proposed methods on the existing sentence classifiers, the sentence classification accuracy on erroneous sentences is increased by 8% to 15%.

Wireless communication systems play a very crucial role in modern society for entertainment, business, commercial, health and safety applications. These systems keep evolving from one generation to next generation and currently we are seeing deployment of fifth generation (5G) wireless systems around the world. Academics and industries are already discussing beyond 5G wireless systems which will be sixth generation (6G) of the evolution. One of the main and key components of 6G systems will be the use of Artificial Intelligence (AI) and Machine Learning (ML) for such wireless networks. Every component and building block of a wireless system that we currently are familiar with from our knowledge of wireless technologies up to 5G, such as physical, network and application layers, will involve one or another AI/ML techniques. This overview paper, presents an up-to-date review of future wireless system concepts such as 6G and role of ML techniques in these future wireless systems. In particular, we present a conceptual model for 6G and show the use and role of ML techniques in each layer of the model. We review some classical and contemporary ML techniques such as supervised and un-supervised learning, Reinforcement Learning (RL), Deep Learning (DL) and Federated Learning (FL) in the context of wireless communication systems. We conclude the paper with some future applications and research challenges in the area of ML and AI for 6G networks.

The use rate of urban land is a significant sign to evaluate urban construction, and scene recognition has important application value in improving urban land use rate. In this paper, a new lightweight model based on VGG16 is proposed to extract distinct features of remote sensing images through five convolution modules. This model uses depthwise separable convolution to reduce the network parameters. An adaptive pooling layer is added to solve the inherent non-adaptive problem of the convolution network. It makes the network insensitive to the size of the input image. The global average pooling layer is used to sum the information to make the input spatial transformation more stable. This paper conducts training and testing on two data sets, NWPU-RESISC45 Dataset and SIRI-WHU Dataset, and the recognition scenarios are 13 and 12 categories. Experimental results show that this method is better than other models in recognition accuracy and model size.

Machine learning (ML) is becoming an integral part of networks security arsenal, where Internet of Things (IoT) structures play an increasingly important role. However, IoT networks have many specific requirements, mostly due to limited energy availability and stringent computing resources. This results in limitations for traditional ML approaches to security, in particular for anomaly detection. Consequently, new focuses for solutions that range from architectural to data processing ones are necessary. Therefore, appropriate lightweight ML algorithms have to be designed and deployed in appropriate architectural settings, which is the main contribution of this paper. In addition, insights into ML functioning are needed to better understand the observed anomalies. To enable these insights (and support a wider applicability of ML based approaches), the results have to be as explainable as possible. The research presented in this paper addresses this problem through the functional and data transparency of ML applications, tailored to the specifics of anomaly detection in IoT networks. To tackle accordingly also the architectural issues, the presented approach builds on the well-established layering principle from computer communications reference models. This principle not only supports flexibility but also increases security in these new environments of growing importance.

There are many brands and types of liquor in China. Yet, some merchants use low-quality liquor for high-quality sales, which contains methanol that is harmful to humans. Therefore, the identification of fake liquor is of great significance to human health and brand protection. This paper presents a laser-induced fluorescence (which has very high sensitivity and rapidity) recognition liquor. Specifically, it first sets up a LIF spectrum acquisition system for liquor in the laboratory, uses Principal Component Analysis (PCA) to reduce the dimension of LIF spectral data, and then collects the dimension reduction data for Random Forest (RF) identification. The results show that the average recognition accuracy of RF combined with PCA for liquor is as high as 99.80%. This paper combines PCA-RF with LIF to accurately identify different brands and types of liquor, which provides a new method for the recognition of fake liquor.

Understanding the peak time of popularity evolution can provide insights on recommendation systems and online advertising campaigns. Although popularity evolution has been largely studied, the problem of how to predict its peak time remains unexplored. Taking Twitter hashtags as case study, the goal of this study is to predict when popularity reaches the peak for Twitter hashtags, from the perspective of an online social network application, in the context of the Twitter social network. On the whole, this paper includes three research aspects. Firstly, this paper investigates how early popularity reaches its peaks. Then, it is found that popularity tends to peak in the early stage of its evolution. Secondly, this paper discusses when a peak time prediction should be triggered. Thirdly, this paper designs a multi-modal based deep learning method, where the state-of-art deep learning techniques, such as multi-modal embedding and attention mechanisms, are adopted. We find that in the early stage of popularity evolution, no matter which factor is used as the input, the prediction effect is poor. By contrast, the hashtag string factor has the weakest contribution to popularity prediction in the middle and late stages of popularity evolution. The overall performance of our proposed method is evaluated in terms of the minimum, quartiles, and maximum values of absolute errors. From the experimental results, the prediction method we designed is superior.

Face recognition technology is a biometric technology, which is based on the identification of facial features of a person. People collect the face images, and the recognition equipment automatically processes the images. The paper introduces the related researches of face recognition from different perspectives. The paper describes the development stages and the related technologies of face recognition. We introduce the research of face recognition for real conditions, and we introduce the general evaluation standards and the general databases of face recognition. We give a forward-looking view of face recognition. Face recognition has become the future development direction and has many potential application prospects.

Analysis and collection of time-series data as a major role of machine learning has been emphasized with an important key in cognitive science. Because the cognitive mechanisms such as human sensation and perception from cognitive science are fast responses ranging from a few milliseconds to hundreds of milliseconds, the method of pattern recognition and analysis of these brain signals must be done and it is necessary to derive some information. In this paper, we investigated time-series data of cognitive function of the brain obtained using a non-invasive technique on multiple channels via signal classification and analysis, using a cognitive science approach and experiments. The test dataset was collected in 19 channels using functional near-infrared spectroscopy (fNIRS) and electroencephalography (EEG) techniques with multiple rests and working conditions on eight subjects. From this perspective, the main contributions of this paper are that it completes the collection and analysis of cognitive-scientific time-series data and has scientific implications that extend to other integrated domains, energy, manufacturing, bioinformatics, and finance area. The use of Shapelet and DTW (Dynamic Time Warping) classification techniques on brain signal time-series shows the potential to identify neuro-biological phenomena that can proactively signal a disease or disorder. EEG bandwidth and frequency-specific data have also been categorized as machine learning algorithms and have shown accurate patterns and trends in measuring cognitive functions of scientific, biological and academic importance.

Nature-inspired optimization is a modern technique in the past decades. Researchers report their successful applications in various fields such as manufacturing, biomedical, and environmental engineering, while other researchers doubt its applicability. In this paper, we collect newly emerging nature-inspired optimization algorithms proposed after 2008, present them in a unified way, implement them, and evaluate them on benchmark functions. Moreover, we optimize the behavioural parameters for these algorithms. Since it is impossible to cover all interesting topics regarding nature-inspired optimization, this paper only focuses on the continuous encoding algorithms for single objective global problems, which is fundamental for other related topics.

Recently, radar high-resolution range profile (HRRP) recognition based on convolutional neural networks (CNNs) has received considerable attention due to its robustness to translation and amplitude changes. Most of the existing methods require that sufficient labeled data with complete aspect angles be used as training data, which is a difficult task in practice. In addition, HRRP signals have a high sensitivity to the aspect angle. Therefore, the representative and discriminative powers of the features extracted from typical CNN models are reduced due to incomplete aspect angles in the training data, which significantly limit the recognition performance. This paper first considers the problem of HRRP recognition with incomplete aspect angle training data and addresses the problem by a deep transfer learning framework. Specifically, the two proposed methods enhance the recognition performance by exploring the discriminative power and the intraclass consistency with auxiliary data, which have HRRP signals with complete aspect angles. This paper generates a simulated HRRP dataset from public data to validate the proposed work. The comparisons of the recognition results demonstrate that the proposed framework outperforms the latest CNN-based models.

Traditional image classification technology has become increasingly unable to meet the changing needs of the era of big data. With the open source use of a large number of marked databases and the development and promotion of computers with high performance, deep learning has moved from theory to practice and has been widely used in image classification. This paper takes big data image classification as the research object, selects distributed deep learning tools based on Spark cluster platform, and studies the image classification algorithm based on distributed deep learning. Aiming at the problems that the Labeled Structural Deep Network Embedding (LSDNE) model is applied to the attribute network and generates a large number of hyperparameters and the model complexity is too high, inspired by the Locally Linear Embedding (LLE) algorithm, this paper proposes a semi-supervised network based on the neighbor structure learning model. This model will add the neighbor information of the node at the same time when learning the network representation. Through the node vector reconstruction, the node itself and the neighbor node together constitute the next layer of representation. On the basis of Structural Labeled Locally Deep Nonlinear Embedding (SLLDNE), the node attribute is further added to propose Structural Informed Locally Distributed Deep Nonlinear Embedding (SILDDNE), and how the model combines the structural characteristics of the node with the attribute characteristics is explained in detail. The SVM classifier classifies the known labels, and SILDDNE fuses the network structure, labels, and node attributes into the deep neural network. The experimental results on the CIFAR-10 and CIFAR-100 datasets for image classification standard recognition tasks show that the proposed network achieves good classification performance and has a high generalization ability. Experiments on the CIFAR-10 data set show that the 34-layer SLLDNE pruned 40-layer Dense Net compresses about 50% of the parameter amount, increases the computational complexity efficiency by about 8 times, and reduces the classification error rate by 30%. Experiments on the CIFAR-100 data set show that the 34-layer SLLDNE parameter volume is compressed by about 16 times compared to the 19-layer VGG parameter volume, the computational complexity efficiency is increased by about 6 times, and the classification error rate is reduced by 14%.

With the rapid development of machine learning, its powerful function in the machine vision field is increasingly reflected. The combination of machine vision and robotics to achieve the same precise and fast grasping as that of humans requires high-precision target detection and recognition, location and reasonable grasp strategy generation, which is the ultimate goal of global researchers and one of the prerequisites for the large-scale application of robots. Traditional machine learning has a long history and good achievements in the field of image processing and robot control. The CNN (convolutional neural network) algorithm realizes training of large-scale image datasets, solves the disadvantages of traditional machine learning in large datasets, and greatly improves accuracy, thereby positioning CNNs as a global research hotspot. However, the increasing difficulty of labeled data acquisition limits their development. Therefore, unsupervised learning, self-supervised learning and reinforcement learning, which are less dependent on labeled data, have also undergone rapid development and achieved good performance in the fields of image processing and robot capture. According to the inherent defects of vision, this paper summarizes the research achievements of tactile feedback in the fields of target recognition and robot grasping and finds that the combination of vision and tactile feedback can improve the success rate and robustness of robot grasping. This paper provides a systematic summary and analysis of the research status of machine vision and tactile feedback in the field of robot grasping and establishes a reasonable reference for future research.

In recent years, the class imbalance problem that aims to correctly classify imbalanced data sets and improve the classification performance of minority instances has received attention. Such problem can be roughly described as one of the class(es) termed as minority class(es) contains much smaller instances than the others, also referred as majority class(es). To address this problem, a novel classification method called focused online random forest based on synthetic minority oversampling technique (FORF-SMOTE) is proposed in this paper and simply expressed as FORF-S, which constructs two online random forests respectively trained by original training dataset and new generated dataset, then further jointly consititute the model. Instead of oversampling the minority instances in data level, the algorithm in this paper is motivated by making the sampling strategies integrated into algorithm level to create classifiers, which can better identify the minority class. Moreover, the method is also compared with other state-of-the-art methods and the results have demonstrated that the proposed algorithm takes advantages of the aforementioned methods.

In view of the shortcomings of traditional fault diagnosis methods based on time domain vibration analysis, which require complicated calculations of feature vectors, and are over-dependent on a prior diagnosis knowledge, effort for the design of the feature extraction algorithms, and have limited ability to extract the complex relationships in fault signals, in this paper we propose a convolutional neural network (CNN) framework for machine health monitoring based on the encoding of one-dimension (1-D) time series to two-dimension (2-D) images. This paper defines a new Gram matrix and through the Python programming environment, we emulate the new Gram matrix in 2-D images, thus feature extraction and recognition can be performed by CNNs. The proposed method which is tested on two datasets, including multi-stage centrifugal fan dataset for our lab, motor bearing dataset for Case Western Reserve University, has achieved prediction average accuracy of 94.07% and 96.28%, respectively. The results have been compared with other deep learning and traditional methods, including Recurrent neural network (RNN), Support Vector Machines (SVM), Multi-Genetic algorithm, shallow CNN and BP neural network. The results show that the method can improve fault diagnosis accuracy in an effective way and stability than other advanced techniques.

In recent years, linear representation-based methods have been widely researched and applied in the image classification field. Generally speaking, there are three steps within linear representation-based classification (LRC) algorithms. The first step is coding, which uses all training samples to represent the test sample in a linear combination. The second step is subspace approximation, where residuals between the test sample and the linear combination of each class are calculated. The third step is classification, which assigns the class label to the minimum class-specific residual. We classify the LRC methods into six categories: 1) linear representation-based classification methods with norm minimizations, 2) linear representation-based classification methods with constraints, 3) linear representation-based classification methods with feature spaces, 4) linear representation-based classification methods with structural information, 5) linear representation with subspace learning, and 6) linear representation in semi-supervised learning and unsupervised learning. The purpose of this paper is to: 1) make an accurate and clear definition of the linear representation-based method, 2) provide a categorization and a comprehensive survey of the existing linear representation-based classification methods for image classification, 3) Summarize the main applications of linear representation-based methods, 4) provide extensive classification results and a discussion of the linear representation-based methods. Furthermore, this paper summarizes specific applications of the linear representation-based methods. Particularly, we performed extensive experiments to compare thirteen linear representation-based classification methods on seven image classification datasets.

Gait recognition aims to identify person by walking pattern of individuals, which has a valuable application prospect in the fields of identity identification, public security and medical diagnosis. At present, most of the gait recognition methods are designed based on the human gait information. However, this paper focuses on the feature space of gait recognition feature learning, and optimizes the feature space by introducing some general criteria to improve the discrimination of depth features. In recent literature, optimizing the feature space has been proven to be beneficial for personal identification. Taken above insights together, we propose a novel gait recognition method based on joint metric learning and hierarchical network in an end-to-end manner, for improving the ability of feature identification. The contribution of this paper is two-fold. First, a joint metric learning and hierarchical network is used to realize the learning of fine-grained information and accelerated the discriminative feature learning. Second, to learn discriminative feature, we use a fuse loss function, the triplet loss and classifier loss. It guides the training process, pulls the samples in the same set close to each other, pushes the samples in different sets away from each other, and takes the silhouettes in each set as a whole. We evaluate the proposed method on two widely-used gait datasets, ie. CASIA-B and OU-LP-Bag. Experimental results demonstrate the effectiveness of the proposed method, significantly improve gait recognition performances under cross-view and bag-carrying walking conditions, respectively. Particularly, we achieves Rank-1 accuracy of 93.2% on OU-LP-Bag dataset, better than existing state-of-the art.

Images of different modalities play important roles in the fields of military, navigation, and target detection, including visible and infrared images. Existing fusion methods can reach relatively good fusion effects, but often make the processing speed slow. To achieve the purpose of faster fusion, this paper proposes a fast fusion of visible and infrared images (FFVI) based on Fourier transform and difference minimization. First, both visible and infrared images are transformed using Fourier transform, and the contour information of infrared image is extracted out at the same time. Then the two transformed images are added together to obtain a composite image, the inverse Fourier transform and the grayscale normalization are performed on the composite transformation image, which will generate an image that reflects the target features. Finally, the resulting image is reconstructed by fusing visible and the generated images according to the principle of “difference minimization”. The experimental results show that the novel method FFVI proposed in this paper can attain remarkable fusion effects in terms of the subjective qualitative and the objective quantitative analysis. In addition, this method outperforms several representative image fusion algorithms in the computational speed, which will be beneficial to practical application.

Sentiment analysis is one of the significant application fields of natural language processing. Researchers divide the sentiment into explicit sentiment and implicit sentiment according to whether the expression contains sentiment words or not. Most of the existing research focuses on explicit sentiment tasks. However, implicit sentiment analysis is difficult to study because of its implicit tendency to express sentiment and the problem of “weak features” and “multi-confounding weak features”. We develop a novel model based on “Hierarchical Knowledge Enhancement and Multi-pooling”(HKEM), which fully integrates the knowledge information of different levels in the text through hierarchical knowledge enhancement to alleviate the “weak feature” problem. At the same time, the multi-pooling method is used to extract and analyze multiple features, and the “multi-confounding weak features” are modeled. In this paper, comparative experiments have been conducted on the SMP-ECISA2019 dataset, and the results show that the accuracy of this model is comparable to that of the current best model, and the F1 score is 5.9% higher than that of the best model, indicating that the model proposed in this paper can accurately analyze Chinese implicit sentiment and achieve the state-of-the-art in the comprehensive performance evaluation, which proves the effectiveness and superiority of the model in this paper.

Given the similarity of handwritten formula symbols and various handwriting styles, this paper proposes a squeeze-extracted multi-feature convolution neural network (SE-MCNN) to improve the recognition rate of handwritten formula symbols. The system proposed in this paper integrates the eight-directional feature of the original sequence in the convolutional layer, which significantly compensates for the lost dynamic trajectory information in the handwritten formula symbol. Meanwhile, the joint loss is constructed to improve the discriminability of features in the way of supervised learning, which enlarges the inter-class difference and decreases inner-class similarity. The standard mathematical formula symbol library provided by the Competition Organization on Recognition of Online Handwritten Mathematical Expression (CROHME) is used to verify the effectiveness of the proposed algorithm. Experiments show that the proposed SE-MCNN approach outperforms the state-of-the-art methods even at the condition of without using the data augmentation.

Charging of large-scale electric vehicles (EVs) will have a serious impact on the power grid. In the environment of power market, the orderly scheduling of EVs through agents is an effective way to solve this problem. Due to the uncertainty of EV travels, how to participate and profit in demand response has become urgent. In this paper, a set of optimal bidding method for agents participating in the market is proposed. Firstly, because the EV service provider cannot collect the historical travel data of all the owners in the whole region, a multi-layer machine learning algorithm is used to simulate the travel behavior of EV in this paper. Secondly, based on the demand response rules of the current electricity market, the optimal bidding model for EV service providers is proposed. In addition, considering the offset between the predicted and the actual electricity price, the model is designed based on interval optimization. Finally, the behavior model and the bidding method are applied to IEEE30 bus system and a provincial practical system. It is shown that the proposed method can bring greater benefits to EV service providers.

Identification of growing points of cotton main stem is the key to realize intelligent and precise machine topping and chemical topping. The identification of growing points by applying the traditional target detection model is subject to a series of shortcomings such as low accuracy, slow identification speed, large number of model parameters, huge storage cost and high calculation workload. On the basis of the advantages and disadvantages of YOLOv3, this paper proposed a modified YOLOv3 lightweight model for identifying the growing points of cotton main stem, which realized the multiplexing integration of low-level semantic features and high-level semantics features by adding dense connection modules and modifying the nonlinear transformation within the dense modules. This model significantly reduced the number of model parameters by utilizing deep separable convolution and improved the learning ability of multi-scale features by applying the hierarchical multi-scale method. Our model achieved an accuracy rate of 90.93% based on a self-prepared dataset, which is higher than the accuracy of the original YOLOv3 model by 1.64%, while the number of training parameters was significantly reduced by 48.90%. Compared with other target detection models under different illumination conditions and actual complex environments, the modified YOLOv3 model proposed in this paper showed better robustness, higher accuracy and higher speed in identifying the growing points of cotton main stem.

The existing deep learning methods for human fall detection have difficulties to distinguish falls from similar daily activities such as lying down because of not using the 3D network. Meanwhile, they are not suitable for mobile devices because they are heavyweight methods and consume a large number of memories. In order to alleviate these problems, a two-stream approach to fall detection with the MobileVGG is proposed in this paper. One stream is based on the motion characteristics of the human body for detection of falls, while the other is an improved lightweight VGG network, named the MobileVGG, put forward in the paper. The MobileVGG is constructed as a lightweight network model through replacing the traditional convolution with a simplified and efficient combination of point convolution, depth convolution and point convolution. The residual connection between layers is designed to overcome the gradient disappeared and the obstruction of gradient reflux in the deep model. The experimental results show that the proposed two-stream lightweight fall classification model outperforms the existing methods in distinguishing falls from similar daily activities such as lying and reducing the occupied memory. Therefore, it is suitable for mobile devices.

Wireless technology and the latest developments in a mobile object, has led to a Mobile Ad Hoc network (MANET), which is a collection of mobile nodes that are communicating with each other without requiring any fixed infrastructure. Due to the dynamic nature with a decentralized system, these networks are susceptible to different attacks such as Black Hole Attack (BHA), Gray Hole Attack (GHA), Sink Hole Attack (SHA) and many more. Several researchers have worked for the detection and mitigation of individual attacks, either GHA or BHA nodes. But the protection of MANET against a dual-threat is scarce. In this paper, the protection against dual attacks has been presented for BHA and GHA by using the concept of Artificial Neural Network (ANN) as a deep learning algorithm along with the swarm-based Artificial Bee Colony (ABC) optimization technique. The performance of the system has been increased by the selection of appropriate and best nodes for data packets transmission which is explained in the result section of this paper. For the network designing and simulation purposes, MATLAB software is used with communication and neural network toolboxes. The examined results show that the presented protocol performs better in contrast to the existing work under black hole as well as gray hole attack condition. A mobile ad hoc network (MANET) is a collection of mobile nodes that dynamically form a temporary network without using any existing network infrastructure.

Classification is the key and most widely studied paradigm in machine learning community. The selection of appropriate classification algorithm for a particular problem is a challenging task, formally known as algorithm selection problem (ASP) in literature. It is increasingly becoming focus of research in machine learning community. Meta-learning has demonstrated substantial success in solving ASP, especially in the domain of classification. Considerable progress has been made in classification algorithm recommendation and researchers have proposed various methods in literature that tackles ASP in many different ways in meta-learning setup. Yet there is a lack of survey and comparative study that critically analyze, summarize and assess the performance of existing methods. To fill these gaps, in this paper we first present a literature survey of classification algorithm recommendation methods. The survey shed light on the motivational reasons for pursuing classifier selection through meta-learning and comprehensively discusses the different phases of classifier selection based on a generic framework that is formed as an outcome of reviewing prior works. Subsequently, we critically analyzed and summarized the existing studies from the literature in three important dimensions i.e., meta-features, meta-learner and meta-target. In the second part of this paper, we present extensive comparative evaluation of all the prominent methods for classifier selection based on 17 classification algorithms and 84 benchmark datasets. The comparative study quantitatively assesses the performance of classifier selection methods and highlight the limitations and strengths of meta-features, meta-learners and meta-target in classification algorithm recommendation system. Finally, we conclude this paper by identifying current challenges and suggesting future work directions. We expect that this work will provide baseline and a solid overview of state of the art works in this domain to new researchers, and will steer future research in this direction.

This paper proposes an adaptive image enhancement method for electrical impedance tomography (EIT). The images are enhanced based on a steerable and multi-scale resolution enhancement algorithm. It is initiated by capturing the spatial variations in decomposition orientations, and decomposition scales of the EIT image. The interpretation of projected image sub-bands is translated into resolution through statistical processes. A steerable filter containing Gaussian basis function derivatives captures the statistical information. Using the regional quantization method (RQM) proposed in this paper, projection weights are computed through spatial statistics of the image sub-bands and tuned adaptively. RQM assigns more resolution to those directional edges which have higher standard deviation and embeds high-order curvatures into the EIT images while suppressing noise. Comparison with conventional image enhancement methods demonstrates the superior performance of RQM. Using RQM it is shown that for 16, 32 and 64 electrode configurations with noise-free recording of 32 × 32 EIT images the number of electrodes can be reduced by 5, 7 and 12 respectively without loss of detail.

In this paper, we are interesting in inferring 3D pose estimation of aircraft object leveraging 2D key-points localization. Monocular vision based pose estimation for aircraft can be widely utilized in airspace tasks like flight control system, air traffic management, autonomous navigation and air defense system. Nonetheless, prior methods using directly regression or classification can not meet the requirements of high precision in aircraft pose estimation context, other approaches using PnP algorithms that need additional information such as template 3D model or depth as prior knowledge. These methods do not exploit to full advantage the correlation information between 2D key-points and 3D pose. In this paper, we present a multi-branch network, named AirPose network, using convolutional neural network to address 3D pose estimation based on 2D key-points information. In the meantime, a novel feature fusion method is explored to enable orientation estimation branch adequately exploit key-points information. Our feature fusion method significantly decreases 3D pose estimation error also avoids the involvement of RANSAC based PnP algorithms. To address the problem that there is no available dedicated aircraft 3D pose dataset for training and testing, we build a visual simulation platform on Unreal Engine 4 applying domain randomization (DR) skill, named AKO platform, which generates aircraft images automatically labeled with 3D orientation and key-points location. The dataset is called AKO dataset. We implement a series of ablation experiments to evaluate our framework for aircraft object detection, key-points localization and orientation estimation on AKO dataset. Experiments show that our proposed AirPose network leveraging AKO dataset can achieve convincing results for each of the tasks.

This paper proposes a novel parametric identification approach for linear systems using Deep Learning (DL) and the Modified Relay Feedback Test (MRFT). The proposed methodology utilizes MRFT to reveal distinguishing frequencies about an unknown process; which are then passed to a trained DL model to identify the underlying process parameters. The presented approach guarantees stability and performance in the identification and control phases respectively, and requires few seconds of observation data to infer the dynamic system parameters. Quadrotor Unmanned Aerial Vehicle (UAV) attitude and altitude dynamics were used in simulation and experimentation to verify the presented methodology. Results show the effectiveness and real-time capabilities of the proposed approach, which outperforms the conventional Prediction Error Method in terms of accuracy, robustness to biases, computational efficiency and data requirements., 13 pages, 9 figures. Submitted to IEEE access. A supplementary video for the work presented in this paper can be accessed at: https://www.youtube.com/watch?v=dz3WTFU7W7c. This version includes minor style edits for appendix and references

The development of convolutional neural networks has promoted the progress of computer-aided diagnostic systems. Details in medical image, such as the texture and tissue structure, are crucial features for diagnosis. Therefore, large input images combined with deep convolution neural networks are adopted to boost the performance in recent research of chest X-ray diagnosis. Meanwhile, due to the variable sizes of thoracic diseases, many researchers have worked to introduce additional module to capture multi-scale feature of images in CNN. However, these efforts hardly consider the computational costs of large inputs and introduced additional modules. This paper aims to automatically diagnose diseases on chest X-rays images quickly and effectively. We propose the multi-kernel depthwise convolution(MD-Conv) which contains depthwise convolution kernels with different filter sizes in one depthwise convolution layer. MD-Conv has high calculation efficiency and few parameters. Because its ability to learn multi-scale feature based on the multi-size kernels, it is appropriate for medical images diagnosis tasks in which abnormalities varied in sizes. In addition, larger depthwise convolution kernels are adopted in MD-Conv to obtain a larger receptive field efficiently, which can ensure sufficient receptive field for high resolution inputs. MD-Conv can be easily applied in modern lightweight networks to replace the normal depthwise convolution layer. We conduct experiments on the Chest X-ray 14 Dataset, which is the largest available chest x-ray dataset, and obtain competitive results. We also evaluate the MD-Conv on the new released dataset for pediatric pneumonia diagnosis. We obtain a better performance of 98.3% AUC than original paper (96.8%) for recognize pneumonia versus normal. Meanwhile we compare the FLOPs and Params of different models to show their efficiency for chest X-rays recognition.

In this paper, a twin hyper-ellipsoidal support vector machine (TESVM) for binary classification of data is presented. Similar to twin support SVM(TWSVM) and twin hypersphere SVM (THSVM), as in the literature, our proposed method finds two hyper-ellipsoidals by solving two related SVM-type quadratic programming problem (QPPs), each of which is smaller than that of the classical SVM, causing it to achieve higher speed. The main idea of this paper is to employ Mahalanobis distance-based kernels for two classes of data in the THSVM algorithm to improve its generalization performance. Since the kernel used in SVM, TWSVM, and THSVM is based on Euclidean distance, it is assumed that the data points have been distributed in a hyper-spherical region, while the data points of two classes have been distributed in two different hyper-ellipsoidal regions. As mentioned in the literature, to work with hyper-ellipsoidal areas, Mahalanobis distance is a better choice than Euclidean distance. The effect of computational results of SVM, TWSVM, THSVM, and TESVM in terms of generalization performance and central processing unit (CPU) learning time on several benchmarks as well as synthetic and image datasets indicates, TESVM achieves fast learning speed along with higher generalization.

The accuracy of defects detection for logistics packaging box is a critical factor to ensure the quality of goods under edge computing environment. Now, there are few works on this issue. This paper designs an image acquisition process system and then proposes a novel approach in addressing logistics packaging box defect detection (LPDD) on the basis of support vector machine (SVM). Firstly, this paper designs a new mean denoising template and Laplace sharpening template, which are more suitable for logistics packaging based on image preprocessing, image enhancement and other relevant technical theories. Then in the stage of noise removal, this paper proposes an improved morphological method and a gray morphological edge detection algorithm. The edge defect detection of a gray image is carried out by combining the above two methods. Hence, LPDD extracts the features of logistics packaging box by using scale-invariant feature transform (SIFT) algorithm and designs SVM classifiers to classify the logistics package defects. This paper uses a large number of samples to train, learn and test the designed SVM classifier. The simulation results show that the proposed LPDD method can accurately detect two common types of defects in logistics packaging boxes with higher accuracy and less computational costs, which meets the requirements of manufacturers on the classification and recognition of defects in machine vision detection system.