International Journal of Pattern Recognition and Artificial Intelligence

Traditional classifiers are trapped by the class-imbalanced problem due to the fact that they are biased toward the majority class. Oversampling methods can improve imbalanced classification by creating synthetic minority class samples. Noise generation has been a great challenge in oversampling methods. Filtering-based and direction-change methods are proposed against noise generation. Yet, the adopted noise filters in filtering-based methods are biased to the majority class. Besides, the $k$-nearest neighbor (KNN)-based interpolation in filtering-based and direction-change methods is susceptible to abnormal samples (e.g. outliers, noise or unsafe borderline samples). To overcome noise generation while solving the above shortcomings of filtering-based and direction-change methods, this work presents a new synthetic minority oversampling technique based on local means-based KNN (SMOTE-LMKNN). In SMOTE-LMKNN, the local mean-based KNN (LMKNN) is first introduced to describe the local characteristic of imbalanced data. Second, a new LMKNN-based noise filter is proposed to remove noise and unsafe borderline samples. Third, the interpolation between a base sample and its LMKNN is proposed to create synthetic minority class samples. Empirical results of extensive experiments with 18 data sets show that SMOTE-LMKNN is competent compared with seven popular oversampling methods in training KNN classifier and classification and regression tree (CART).

Numerous research based on offline Tamil recognition deals only with few Tamil characters since it becomes extremely complicated in distinguishing small variations in large handwritten document. The writer’s complexity affects the overall formation of the characters. Such types of complexities are due to discontinuation of structures, unnecessary over loops, variation in shapes as well as irregular curves. This complex issue results in enhanced error value rate. Therefore, to conquer such issues, this paper proposes a novel approach to enhance the offline Tamil handwritten character recognition by utilizing four principal steps: pre-processing, segmentation, feature extraction and classification. For optimal segmentation of Tamil characters, this paper utilizes the Tsallis entropy approach-based atom search (TEAS) optimization algorithm. Then a Newton algorithm based deep convolution extreme learning (DELM) approach is utilized for the extraction and classification of input images. Finally, experiments are carried out for numerous Tamil handwritten recognition-based approaches. The proposed Tamil character recognition utilizes the datasets of isolated Tamil handwritten characters established by HP lab India to evaluate the efficiency of the system.

The automatic detection and recognition for motorcycle license plates present a very challenging task since they appear more compact and versatile than vehicle license plates. In this paper, we present an efficient detection and recognition system for motorcycle license plates based on decision tree and deep learning. It can be successfully carried out under various conditions, such as frontal, horizontally or vertically skewed, blurry, poor illumination, large viewing distances or angles, distortions, multiple license plates in an image, at night or interfered with brake lights, and headlights. Experimental results show that our system performs the best when testing with multiple license plates images under different conditions as compared against six state-of-the-art methods. Furthermore, our detection and recognition system have shown more accurate results than three commercial automatic license plate recognition systems in evaluation using accuracy, precision, recall, and F1 rates.

Recently, information networks are becoming a significant part of daily life, so keeping the system’s security is necessary for security tools, such as firewalls and encryption. However, because of the weaknesses of the existing tools, the Intrusion Detection System (IDS) has been implemented to solve the problem. In the application of IDS, feature classification and data analysis are the two most important steps. In this paper, by using the Logit regression model, we attempt to search for the optimal cutting value based on the relationship between cutting value and accuracy index and put forward an input-output port crossed (IOPC) classification for IDS to distinguish the new intrusion features. First, we discuss whole features and propose a taxonomy of IOPC classification for CIC-IDS2017 that is different from other former studies, which can reduce the data space. Second, we compute the distribution curve of cutting values varied with the accuracy index, the purpose of which is to search for the optimal cutting values. Finally, utilizing IOPC classification, the difference between the distribution of the cutting values under the attacks of distributed denial of service (DDoS) and PortScan in CIC-IDS2017 is discussed, which highlights the characteristic that cutting values besieged the attack by PortScan has a conditional distribution compared with DDoS.

Machine reading comprehension (MRC) is a crucial and challenging task in natural language processing (NLP). With the development of deep learning, language models have achieved excellent results. However, these models still cannot answer complex questions. Currently, researchers often utilize structured knowledge, such as knowledge bases (KBs), as external knowledge by directly extracting triples to enhance the results of machine reading. Although they can support certain background knowledge, the triples are limited to the interrelationships among entities or words. Unlike structured knowledge, unstructured knowledge is rich and extensive. However, these methods ignore unstructured knowledge resources, such as Wikipedia. In addition, the effect of combining the two types of knowledge is still not known. In this study, we first attempt to explore the usefulness of combining them. We introduce a fusion mechanism into a rich knowledge fusion layer (RKF) to obtain more useful and relevant knowledge from different external knowledge resources. Further to promote interaction among different types of knowledge, a bi-matching layer is added. We propose the RKF-NET framework based on BERT, and our experimental results demonstrate the effectiveness of two classic datasets: SQuAD1.1 and the Easy-Challenge (ARC).

This paper presents a new variant of cuckoo search (CS) algorithm named neighborhood learning-based CS (NLCS) to address global optimization problems. Specifically, in this modified version, each individual learns from the personal best solution rather than the best solution found so far in the entire population to discourage premature convergence. To further enhance the performance of CS on complex multimode problems, each individual is allowed to learn from different learning exemplars on different dimensions. Moreover, the exemplar individual is chosen from a predefined neighborhood to further maintain the population diversity. This scheme enables each individual to interact with the historical experience of its own or its neighbors, which is controlled by using a learning probability. Extensive comparative experiments are conducted on 39 benchmark functions and two application problems of neural network training. Comparison results indicate that the proposed NLCS algorithm exhibits competitive convergence performance.

It is important to locate the fault distance and identify the fault types quickly, take effective measures to maintain line stability, and minimize the losses timely when there are short-circuit faults in transmission lines. For this purpose, a method based on deep learning is proposed for short-circuit faults identification in the transmission line. According to the similarity of samples in the reconstruction phase, a minimum neighborhood sample set is selected from the massive samples firstly, and then, the samples are trained using the back propagation algorithm along time in a recurrent neural network (RNN) with long-short term memory (LSTM) units. Compared with existing algorithms, the experimental results show that this algorithm meets the requirements of rapid fault diagnosis in the case of variable parameters, and higher fault type recognition accuracy and lower fault distance error can be obtained.

Artificial Intelligence (AI) has promoted countless contributions in the field of healthcare and medical imaging. In this paper, we thoroughly analyze peer-reviewed research findings/articles on AI-guided tools for Covid-19 analysis/screening using chest X-ray images in the year 2020. We discuss on how far deep learning algorithms help in decision-making. We identify/address data collections, methodical contributions, promising methods, and challenges. However, a fair comparison is not trivial as dataset sizes vary over time, throughout the year 2020. Even though their unprecedented efforts in building AI-guided tools to detect, localize, and segment Covid-19 cases are limited to education and training, we elaborate on their strengths and possible weaknesses when we consider the need of cross-population train/test models. In total, with search keywords: (Covid-19 OR Coronavirus) AND chest x-ray AND deep learning AND artificial intelligence AND medical imaging in both PubMed Central Repository and Web of Science, we systematically reviewed 58 research articles and performed meta-analysis.

This paper proposes a flexible example-based color transfer system, providing an automatic mode and an advanced mode, for novices and expert users, respectively. The experimental results show that the proposed color transfer system not only can introduce natural results with simple operation by novices in the first use, but also can produce various desired results by users with short-term learning.

The issue of image attribute migration is one of the hot research topics in the field of computer vision, which has received extensive research interest. However, current unsupervised image attribute migration models using symmetric generative adversarial network structure do not work well on datasets with large geometric variations, where the results lack diversity and are of low quality. To address these problems, we present an image attribute migration model based on decoupling and adaptive layer instance normalization. First, a codec structure based on a decoupled representation is constructed as the generator, and an adaptive layer instance normalization operation is used in the decoder. Then, the iterations of the model are constrained by various improved loss functions. We conducted controlled experiments and compared the results of our method with other methods using several datasets with large geometric variations. The experimental results demonstrate that the proposed method can achieve high quality and diverse image attribute migration.

With the usage of computer visual technology in civil engineering, pavement crack survey with imaging sensor gained wide attention in the past years. Unfortunately, it is still a challenge to achieve the satisfying results. This paper presents a pavement crack survey approach based on edge detection for laser data. At first, the LS-40 line-laser scanner is implemented to achieve 3D pavement surface data. With the advantage of the various depth information exhibited in the pavement data, an enhanced edge detection based on fractional differential is proposed for 3D crack segmentation. The proposed method could effectively enhance the crack boundary and maintain texture details, which can guarantee the high accuracy in crack segmentation. Moreover, a novel plane fitting method based on dynamic threshold is studied to calculate crack depth information. It can not only identify and remove invalid points effectively, but also accomplish plane calculation with errors in three directions. Experiments verify that the proposed approach can achieve the satisfying result and can work well in F-measure system.

Ontology matching is able to identify the entity correspondences between two heterogeneous ontologies, which is an effective method to solve the data heterogeneous problem on the Semantic Web. Traditional fully-automatic ontology matching techniques suffer from the limitation of similarity measure, whose alignment’s quality cannot be ensured. To overcome this drawback, in this work, an Interactive Compact Genetic Algorithm (ICGA)-based ontology matching technique is proposed, which utilizes both the compact encoding mechanism and expert interacting mechanism to improve the algorithm’s performance and the alignment’s quality. In addition, an optimization model is established to formally define the ontology entity matching problem, and an efficient interacting strategy is proposed, which is able to reduce the expert’s workload and maximize his working value. The experiment uses Ontology Alignment Evaluation Initiative (OAEI)’s benchmark to test our proposal’s performance. The experimental results show that our approach is able to make use of the expert knowledge to improve the alignment’s quality, and it also outperforms OAEI’s participants.

Social media are digitally mediated platforms that allow people to create and exchange content, professional interests, ideas, and other forms of expression through virtual networks. Users often utilize web-based programs on their PCs and laptops to visit social media sites, or they download programs that provide their devices social media capabilities. As users connect with these platforms, groups, organizations, and individuals can upload, co-create, discuss, engage in, and update self-curated or user-generated information. Although the platforms such as Facebook, Twitter, Instagram, etc., aid in the communication purposes, it also has some demerits like cyber-crime, hacking, etc. The growing number of crimes through these platforms needs to be deducted by predicting the crimes. For the crime prediction, the data acquired from Twitter is pre-processed for the data cleansing process. Later the features are extracted using various techniques like bag of words (BoW), Glove, term frequency-inverse document frequency (TF-IDF), and feature hashing. The feature selection is done using a modified tree growth algorithm (MTGA) and clustering is performed using the fuzzy manta ray foraging (FMRF). Finally, the crime detection is done using hybrid wavelet convolutional neural network with world cup organization (WCNN-WCO). The PYTHON tool is used for the implementation and the Twitter user dataset is used for analysis. The results showed that the proposed method outperforms the existing method in terms of precision, accuracy, $F$1 measure, and recall.

Situational awareness is the ability of pilots to master flight status, which is of great significance to aviation flight safety and flight effect. According to the information processing model, the pilot’s main steps of processing information are feeling, perception and execution. There are many problems in situation awareness analysis guided by visual gaze, such as large analysis deviation and high delay due to various influencing factors and complex characteristics. In order to solve this problem, this paper proposes a situation awareness assessment method based on artificial intelligence neural network and integrating visual gaze and flight control. First, this paper carries out simulated flight training experiments for flight cadets, and collects the data of eye movement, line of sight tracking, flight control and flight parameters of pilot cadets. Then, aiming at the flight subjects, a situation awareness analysis method based on events is established, and the situation awareness state in the experiment is evaluated and analyzed through the flight parameter data. Then, the visual gaze and flight control data are sliced in the unit of situational awareness events, and the data set is constructed. Finally, this paper designs a multi-channel sequence data classification and analysis model based on transformer, in which the situation awareness characteristics of visual gaze and operation behavior are analyzed through the attention mechanism. The experimental results show that the accuracy of situation awareness classification of the designed neural network model to the experimental data set is 96%, and can classify and evaluate the pilot’s situation awareness state in 5s.