International Journal of Pattern Recognition and Artificial Intelligence

Structural equation modeling (SEM) is a system of two kinds of equations: a linear latent structural model (SM) and a linear measurement model (MM). The latent structure model is a causal model from the latent parent node to the latent child node. Meanwhile, MM’s link is from latent variable parent node to observed variable child node. However, researchers should determine the initial causal order between variables based on experience when applying SEM. The main reason is that SEM does not fully construct causal models between observed variables (OVs) from big data. When the artificial causal order is contrary to the fact, the causal inference from SEM is doubtful, and the implicit causal information between the OVs cannot be extracted and utilized. This study first objectively identifies the causal order of variables using the DirectLiNGAM method widely accepted in recent years. Then traditional SEM is converted to expanded SEM (ESEM) consisting of SM, MM and observation model (OM). Finally, through model testing and debugging, ESEM with good fit with data is obtained.

One of the major problems in biometrics and in document forensics is the offline mode of signature verification. This study aims to present a novel approach of verifying an individual’s signature through offline images of handwriting. The approach proposed here relies on a global method which considers signature images as waveforms. First, image decompositions are in terms of a series of wavelet sub-bands at some specific levels. Wavelet sub-bands are then extended so as to obtain waveforms. Each waveform is quantized by two Nonlinear Dynamics Tools in order to generate feature vectors. Multi-Resolution Box-Counting (MRBC) fractal dimension algorithm as well as probabilistic finite state automata (PFSA) are applied separately to signature waveforms. In the training and verification phase, we propose the one-class Gaussian process (GP) priors based on writer-independent approach. As one of the main parameters, optimal decision threshold is selected from False Accept Rate (FAR) and False Reject Rate (FRR) curves. The presented system was tested on two Persian databases (PHBC and UTSig) as well as on two Latin databases (MCYT-75 and CEDAR). In fact, the results produced by this method were generally better in terms of the four signature databases than the state-of-the-art results.

Based on the process of generating HDR images from LDR image sequences with different light exposures in the same scene, a new fitting method of camera response curves is proposed to solve the problem that the boundary of the fitting algorithm of camera response curves will be blurred and it is difficult to determine and verify the accuracy of the fitting curves. The optimal response curve is fitted by increasing LDR images step by step through considering the pixel value and texture characteristics. In order to validate the fitting effect of curves, we compare the photographed images and the real images in different time intervals on the basis of HDR images and response curves. We use RGB and gray image experiments to compare the current mainstream algorithms and the accuracy of our proposed algorithm can reach 96%, which has robustness.

Early detection of water or steam leaks into sodium in the steam generator units of nuclear reactors is an important requirement from safety and economic considerations. Automated defect detection and classification algorithm for categorizing the defects in the steam generator tube (SGT) of nuclear power plants using magnetic flux leakage (MFL) technique has been developed. MFL detection is one of the most prevalent methods of pipeline inspection. Comsol 4.3a, a multiphysics modeling software has been used to obtain the simulated MFL defect images. Different thresholding methods are applied to segment the defect images. Performance metrics have been computed to identify the better segmentation technique. Shape-based feature sets such as area, perimeter, equivalent diameter, roundness, bounding box, circularity ratio and eccentricity for defect have been extracted as features for defect detection and classification. A feed forward neural network has been constructed and trained using a back-propagation algorithm. The shape features extracted from Tsallis entropy-based segmented MFL images have been used as inputs for training and recognizing shapes. The proposed method with Tsallis entropy segmentation and shape-based feature set has yielded the promising results with detection accuracy of 100% and average classification accuracy of 96.11%.

Traditional automatic lip-reading systems generally consist of two stages: feature extraction and recognition, while the handcrafted features are empirical and cannot learn the relevance of lip movement sequence sufficiently. Recently, deep learning approaches have attracted increasing attention, especially the significant improvements of convolution neural network (CNN) applied to image classification and long short-term memory (LSTM) used in speech recognition, video processing and text analysis. In this paper, we propose a hybrid neural network architecture, which integrates CNN and bidirectional LSTM (BiLSTM) for lip reading. First, we extract key frames from each isolated video clip and use five key points to locate mouth region. Then, features are extracted from raw mouth images using an eight-layer CNN. The extracted features have the characteristics of stronger robustness and fault-tolerant capability. Finally, we use BiLSTM to capture the correlation of sequential information among frame features in two directions and the softmax function to predict final recognition result. The proposed method is capable of extracting local features through convolution operations and finding hidden correlation in temporal information from lip image sequences. The evaluation results of lip-reading recognition experiments demonstrate that our proposed method outperforms conventional approaches such as active contour model (ACM) and hidden Markov model (HMM).

Full-color three-dimensional (3D) printing technology is a powerful process to manufacture intelligent customized colorful objects with improved surface qualities; however, poor surface color optimization methods are the main impeding factors for its commercialization. As such, the paper explored the correlation between microstructure and color reproduction, then an assessment and prediction method of color optimization based on microscopic image analysis was proposed. The experimental models were divided into 24-color plates and 4-color cubes printed by ProJet 860 3D printer, then impregnated according to preset parameters, at last measured by a spectrophotometer and observed using both a digital microscope and a scanning electron microscope. The results revealed that the samples manifested higher saturation and smaller chromatic aberration ($\mathrm{\Delta }E$) after postprocessing. Moreover, the brightness of the same color surface increased with the increasing soaked surface roughness. Further, reduction in surface roughness, impregnation into surface pores, and enhancement of coating transparency effectively improved the accuracy of color reproduction, which could be verified by the measured values. Finally, the chromatic aberration caused by positioning errors on different faces of the samples was optimized, and the value of $\mathrm{\Delta }E$ for a black cube was reduced from 8.12 to 0.82, which is undetectable to human eyes.

This paper proposes a simultaneous localization and mapping (SLAM)-based markerless mobile-end tracking registration algorithm to address the problem of virtual image drift caused by fast camera motion in mobile-augmented reality (AR). The proposed algorithm combines the AGAST-FREAK-SLAM algorithm with inertial measurement unit (IMU) data to construct a scene map and localize the camera’s pose. The extracted feature points are matched with feature points in a map library for real-time camera localization and precise registration of virtual objects. Experimental results show that the proposed method can track feature points in real time, accurately construct scene maps, and locate cameras; moreover, it improves upon the tracking and registration robustness of earlier algorithms.

Virtual Reality (VR) 360^∘ video is spherical, and has ultra-high definition. It is time-consuming to code it using High Efficient Video Coding (HEVC) directly. In this paper, a fast intra mode selection algorithm based on Coding Unit (CU) size is proposed for VR 360^∘ video in Cubemap projection (CMP) format. A redefined mode selection strategy for two-stage Rough mode decision (RMD) is designed based on different depth CUs to reduce coding time during intra prediction process. The experimental results show that the proposed algorithm brings 27.70% saving in average coding time at the cost of only 0.8% Bjontegaard delta rate increase in All-Intra mode.

The most significant part of any autonomous intelligent robot is the localization module that gives the robot knowledge about its position and orientation. This knowledge assists the robot to move to the location of its desired goal and complete its task. Visual Odometry (VO) measures the displacement of the robots’ camera in consecutive frames which results in the estimation of the robot position and orientation. Deep Learning, nowadays, helps to learn rich and informative features for the problem of VO to estimate frame-by-frame camera movement. Recent Deep Learning-based VO methods train an end-by-end network to solve VO as a regression problem directly without visualizing and sensing the label of training data in the training procedure. In this paper, a new approach to train Convolutional Neural Networks (CNNs) for the regression problems, such as VO, is proposed. The proposed method first changes the problem to a classification problem to learn different subspaces with similar observations. After solving the classification problem, the problem converts to the original regression problem to solve using the knowledge achieved by solving the classification problem. This approach helps CNN to solve regression problem globally in a local domain learned in the classification step, and improves the performance of the regression module for approximately 10%.

Traffic sign recognition (TSR) is a key technology of intelligent vehicles, which is based on visual perception for road information. In view of the fact that the traditional computer vision identification technology cannot meet the requirements of real-time accuracy, the TSR algorithm has been proposed on the basis of improved Lenet-5 algorithm. Firstly, we performed picture noise elimination and image enhancement on selected traffic sign images. Secondly, we used Gabor filter kernel in the convolution layer for convolution operation. In the convolution process, we added normalization layer Batch Normality (BN) after each convolution layer and reduced the data dimension. In the down-sampling layer, we replaced Sigmoid with the Relu activator. Finally, we selected the expanded GTSRB traffic sign database for the comparison experiment on the Caff platform. The experimental results showed that the proposed improved Lenet-5 network test set had the recognition accuracy of 96%, which was better than the method that combined Gabor with Support Vector Machine (SVM) in terms of recognition accuracy and real-time performance.

Face recognition, as a research hot topic, still faces many challenges. This paper proposes a new face recognition method by fusing the advantages of fuzzy set theory, sub-image method and random sampling technique. In this method, we partition an original image into some sub-images to improve the robustness to different facial variations, and extract local features from each sub-image by using fuzzy 2D-Linear Discriminant analyzis (LDA) which makes use of the class information hidden in neighbor samples. In order to increase the diversity of component classifiers and retain as much as the structural information of the row vectors, we further randomly sample row vectors from each sub-image before performing fuzzy 2D-LDA. Experimental results on Yale A, ORL, AR and Extended Yale B face databases show its superiority to other related state-of-the-art methods on the different variations such as illumination, occlusion and facial expression. Furthermore, we analyze the diversity of our proposed method by virtue of Kappa diversity-error analyzis and frequency histogram and results show that the proposed method can construct more diverse component classifiers than other methods.

Cancer prediction from gene expression data is a very challenging area of research in the field of computational biology and bioinformatics. Conventional classifiers are often unable to achieve desired accuracy due to the lack of ‘sufficient’ training patterns in terms of clinically labeled samples. Active learning technique, in this respect, can be useful as it automatically finds only few most informative (or confusing) samples to get their class labels from the experts and those are added to the training set, which can improve the accuracy of the prediction consequently. A novel active learning technique using fuzzy-rough nearest neighbor classifier (ALFRNN) is proposed in this paper for cancer classification from microarray gene expression data. The proposed ALFRNN method is capable of dealing with the uncertainty, overlapping and indiscernibility often present in cancer subtypes (classes) of the gene expression data. The performance of the proposed method is tested using different real-life microarray gene expression cancer datasets and its performance is compared with five other state-of-the-art techniques (out of which three are active learning-based and two are traditional classification methods) in terms of percentage accuracy, precision, recall, $F_1$-measures and kappa. Superiority of the proposed method over the other counterpart algorithms is established from experimental results for cancer prediction and results of the paired $t$-test confirm statistical significance of the results in favor of the proposed method for almost all the datasets.

Machine Translation (MT) evaluation metrics like BiLingual Evaluation Understudy (BLEU) and Metric for Evaluation of Translation with Explicit Ordering (METEOR) are known to have poor performance for word-order and morphologically rich languages. Application of linguistic knowledge to evaluate MTs for morphologically rich language like Hindi as a target language, is shown to be more effective and accurate [S. Tripathi and V. Kansal, Using linguistic knowledge for machine translation evaluation with Hindi as a target language, Comput. Sist.21(4) (2017) 717–724]. Leveraging the recent progress made in the domain of word vector and sentence vector embedding [T. Mikolov and J. Dean, Distributed representations of words and phrases and their compositionality, Adv. Neural Inf. Process. Syst. 2 (2013) 3111–3119], authors have trained a large corpus of pre-processed Hindi text ($\sim 112$ million tokens) for obtaining the word vectors and sentence vector embedding for Hindi. The training has been performed on high end system configuration utilizing Google Cloud platform resources. This sentence vector embedding is further used to corroborate the findings through linguistic knowledge in evaluation metric. For morphologically rich language as target, evaluation metric of MT systems is considered as an optimal solution. In this paper, authors have demonstrated that MT evaluation using sentence embedding-based approach closely mirrors linguistic evaluation technique. The relevant codes used to generate the vector embedding for Hindi have been uploaded on code sharing platform Github.

Due to the multiple types of support tasks and high-energy efficiency equipment, energy saving research in a Bergepanzer has attracted much attention in recent years. A hydraulic hybrid system is an effective way for these problems of Bergepanzers. The initial problem is to optimize the parameter optimization-matching for a new hydraulic hybrid Bergepanzers (NHHB), which is a relatively complex and variable time-varying nonlinear system. In this paper, an RBF-adaptive artificial immune algorithm (RBF&AAIA) is presented to solve the parameter optimization-matching problem, and the performance of algorithm is analyzed. Finally, the RBF&AAIA is successfully used for the NHHB to optimize the parameters of key components for the minimum energy consumption of system, and the optimal parameter matching group is obtained.

Supervisory Control and Data Acquisition (SCADA) system in the modern industrial automation control network is facing an increasing number of serious security threats. In order to meet the security defense requirements of oil and gas SCADA system, an intelligent security defense model based on multi-agent was designed by analyzing the security risks in oil and gas SCADA system and combining the advantages of multi-agent technology in distributed intrusion detection system. First, the whole structure of this model was divided into three layers: monitoring layer, decision layer and control layer. Then, the defense model was verified by C4.5 decision tree algorithm, and obtained a good result. Finally, the security defense prototype system of large-scale oil and gas SCADA system based on this model was realized. Results demonstrate that the application of multi-agent technology in the security defense of oil and gas SCADA system can achieve more comprehensive defense, more accurate detection, which can handle large-scale distributed attacks and improve the robustness and stability of security defenses. This study makes full use of the multi-agent architecture and has the advantage of accurate detection, high detection efficiency and timely response.