International Journal of Pattern Recognition and Artificial Intelligence

This study aims to detect objects quickly and accurately and then start the necessary obstacle avoidance procedure when the uncrewed vessel is at sea. This study uses a multivariate convolutional neural network (CNN) to perform automatic object detection and direction prediction for uncrewed vessels. This study is divided into three parts for processing. The first part of this process uses camera calibration technology to correct the image. Discrete cosine transform (DCT) is then used to detect sea level. Finally, this study uses Kalman filtering and affine transformation to stabilize images taken by uncrewed vessels at sea. The second part of the processing system uses a CNN to detect sea objects automatically. The third part of the process uses the dual-lens camera installed on the vessel to detect the distance and direction of objects at sea. In the experiment, the detection rate can reach more than 91% in this study method. In the experiment on image stabilization, this study’s method can also effectively improve video instability. In the experiment involving the distance and direction of the object, the experimental results show that the distance and direction of the object obtained by this study method have a distance error value of less than 10%, and the prediction results have a good effect no matter whether the object is at a short or long distance. It is hoped that this paper’s method can be applied to the automatic obstacle avoidance of unmanned vessels.

Most of the person re-identification (re-ID) algorithms based on deep learning mainly learn the global feature representation of pedestrians, while ignoring the important role of fine-grained pedestrian attribute features on re-ID tasks. Pedestrian attributes are middle-level semantic features, which have invariance in different poses, camera views, and illumination conditions. Considering the robustness and promotion of pedestrian attributes for person re-ID task, we propose an Attribute-guided Global and Part-level identity Network (AGPNet), which consists of a global identity task, a part-level identity task, and a pedestrian attributes learning task. AGPNet takes advantage of perceived semantic information of pedestrian attributes and deploys them as guidance to attend to human body regions and learn robust feature representation in the feature representation construction stage. Extensive experiments on two large-scale person re-ID datasets (Market-1501 and DukeMTMC-reID) show the effectiveness of our method, which is competitive with the state-of-the-art algorithms.

Mutation testing is widely used in the research of evaluation and optimization of test set quality, and has been paid attention to the study of bug localization and fixing. But one inherent problem of mutation testing is huge computation cost. Selective mutation is an important method to reduce mutation testing cost. However, the existing selective mutation researches indicate that there is no universal selection strategy. This paper proposes a method which integrates with historical software data mining to recommend suitable mutator for program under test. The basic idea of the method is in the software version control system, any pair of nonfixed and post-fixed programs can be original program and mutant of each other. The edit performed during bug fixing contains mutators, and such mutator can guide buggy program to correct program, hence it is called Golden Mutator. The basic method is to compare the buggy files and the corresponding fixed files in the version control system, obtain historical faulty statements and their fix editing operations, thereby accumulating the bug-fix instance base, and then mining mutation patterns from it. Before mutating the target statement, we first find out faulty instances similar to the target statement, and use their fixing edits to match with the mutation pattern, so as to get the golden mutator applicable to the target statement. This paper uses Defects4j dataset in the test, verifies the accuracy of the proposed recommendation method, and further uses the method in bug localization. Compared to fixedly selected mutators, when applying the golden mutator recommended by using the proposed method, the average accuracy of bug localization is higher.

China Agricultural Outlook publishes the outlook report on major agricultural commodities and the estimates of the supply–demand balance sheet of agricultural commodities in the next decade, so as to release agricultural information and guide the development of modern agriculture. A large amount of data has been accumulated in the work associated with agriculture outlook in recent years. This paper will optimize the agriculture commodity model through deep learning and create an analysis tool based on long short-term memory (LSTM) deep learning by comprehensively considering the key factors of supply and demand of agricultural commodities including the output, consumption and price and combining the impact of complex natural, social and economic factors. In this way, the close relevance of different varieties and multi-variable strong coupling of the analysis and prediction model of major agricultural commodities can be solved. The “random sampling” and “stress on causality” of traditional complex agriculture analysis models are replaced by the “whole data” and “emphasis on relevance over causality”. The intelligent decision method of agricultural commodity model based on deep learning proposed in this study can effectively improve the analysis efficiency and accuracy of multi-variety coupling model of agricultural commodities (at least by 15%), and enhance the intelligence of the supply–demand analysis and prediction, especially with the accumulation of future data, the prediction accuracy will continue to improve. Machine learning has been regarded as an effective method to provide forecast and early-warning of future agricultural development in a timely manner based on real-time monitoring of agricultural data.

Mixture of multiple language forms in spoken Chinese is a common but unfavorable issue.. It increases the difficulty of intent understanding and leads to inconvenience for information communication. Existing studies on intent recognition mainly focus on single language form or parallel multilingual language while paying little attention to spoken texts including multiple language forms. In considering that it is hard to capture the semantics of an expression with multiple language forms, it is important to study the problem. To solve this issue, a text representation model for the spoken Chinese expression mixed with English and Chinese Pinyin is proposed. And the feature matrix is built to mine the composition information of English and Pinyin. Besides, the model can efficiently distinguish English from Chinese Pinyin even though both fragments are composed of English letters. Meanwhile, it can effectively process the problem of hidden text information since the problem has been transformed into the Chinese translation task of English and Pinyin. In addition, to verify the performance of the model, the texts processed by this model are used as the input of the classifier. extensive experiments on a large online logistics manual customer service corpus show that this text representation model is correct and effective. It can not only eliminate the obstacles of the mixing of multiple language forms but also bring better results for intent understanding.

Three-dimensional (3D) text appearing in natural scene images is common due to 3D cameras and the capture of text from different angles, which presents new problems for text detection. This is because of the presence of depth information, shadows, and decorative characters in the images. In this work, we consider those images where 3D text appears with depth, as well as shadow information for text detection. We propose a novel method based on local resultant gradient vector difference (LRGVD), inpainting and a deep learning model for detecting 3D as well as two-dimensional (2D) texts in natural scene images. The boundary of components that are invariant to the above challenges is detected by exploring LRGVD. The LRGVD uses gradient magnitude and direction in a novel way for detecting the boundary of the components. Further, we propose an inpainting method in a new way for restoring the character background information using boundaries. For a given region and the input image, the inpainting method divides the whole image into planes and then propagates the values in the planes into the missing region based on posterior probabilities and neighboring information. This results in text regions with false positives. Then, the differential binarization network (DB-Net) is proposed for detecting text irrespective of orientation, background, 3D or 2D, etc. Experiments conducted on our 3D text images and standard datasets of natural scene text images, namely ICDAR 2019 MLT, ICDAR 2019 ArT, DAST1500, Total-Text and SCUT-CTW1500, show that the proposed method is effective in detecting 3D and 2D texts in the images.

Recently, many studies have shown that deep convolutional neural network can achieve superior performance in image super resolution (SR). The majority of current CNN-based SR methods tend to use deeper architecture to get excellent performance. However, with the growing depth and width of network, the hierarchical features from low-resolution (LR) images cannot be exploited effectively. On the other hand, most models lack the ability of discriminating different types of information and treating them equally, which results in limiting the representational capacity of the models. In this study, we propose the multi-attention residual network (MARN) to address these problems. Specifically, we propose a new multi-attention residual block (MARB), which is composed of attention mechanism and multi-scale residual network. At the beginning of each residual block, the channel importance of image features is adaptively recalibrated by attention mechanism. Then, we utilize convolutional kernels of different sizes to adaptively extract the multi-attention features on different scales. At the end of blocks, local multi-attention features fusion is applied to get more effective hierarchical features. After obtaining the outputs of each MARB, global hierarchical feature fusion jointly fuses all hierarchical features for reconstructing images. Our extensive experiments show that our model outperforms most of the state-of-the-art methods.

Clouds frequently affect optical remote sensing pictures throughout the gathering process, resulting in low-resolution images that affect judgment and subsequent use of ground data. Because of the thick cloud cover, the ground surface information below is entirely incorrect. This kind of end-to-end image problem should not be dismissed as a simple task of image inpainting or image translation. Therefore, this paper proposes a multi-head self-attention module based on the encoding–decoding generative adversarial network, considering the redundant information of the deep network, furthermore this paper introduces Ghost convolution to effectively solve the influence of redundant feature maps in the network on the increase of time consumption and parameters. The method in this paper can solve the problem of cloud occlusion. By considering spatial information, it can better complete the prediction of cloud removal. It can reduce the amount of network calculations and parameters while maintaining the effect. In addition, Feature Fusion Module is proposed to integrate high-level features with low-level features, so that the network can extract enough feature information and better supplement the details to complete the cloud removal. The method in this paper has achieved excellent results on the RICE1 and RICE2 datasets.

In this paper, we propose an end-to-end multi-resolution three-dimensional (3D) capsule network for detecting actions of multiple actors in a video scene. Unlike previous capsule, network-based action recognition does not specifically concern with the individual action of multiple actors in a single scene, our 3D capsule network takes advantage of multi-resolution technique to detect different actions of multiple actors that have different sizes, scales, and aspect ratios. Our 3D capsule network is built on top of 3D convolutional neural network (3DCNN) that extracts spatio-temporal features from video frames inside regions of interest generated by Faster RCNN object detection. We first apply our method to the problem of detecting illegal cheating activities in a classroom examination scene with multiple subjects involved. Second, we test our system on the publicly available and extensively studied UCF-101 dataset. We compare our method with several state-of-the-art 3DCNN-based methods, first the multi-resolution 3DCNN, the single-resolution 3D capsule network, and a combination of both these models. We show that models containing 3D capsule networks have a slight advantage over the conventional 3DCNN and multi-resolution 3DCNN. Our 3D capsule networks not only perform a classification of said actions but also generate videos of single actions. Our experimental results show that the use of multi-resolution pathways in the 3D capsule networks make the result even better. Such findings also hold even when we use pre-trained C3D (convolutional 3D) features to train these networks. We believe that the multiple resolutions capture lower-level features at different scales. At the same time, the 3D capsule layers combine these features in more complex ways than conventional convolutional models.

Human action recognition has attracted extensive research efforts in recent years, in which traffic police gesture recognition is important for self-driving vehicles. One of the crucial challenges in this task is how to find a representation method based on spatial-temporal features. However, existing methods performed poorly in spatial and temporal information fusion, and how to extract features of traffic police gestures has not been well researched. This paper proposes an attention mechanism based on the improved spatial-temporal convolutional neural network (AMSTCNN) for traffic police gesture recognition. This method focuses on the action part of traffic police and uses the correlation between spatial and temporal features to recognize traffic police gestures, so as to ensure that traffic police gesture information is not lost. Specifically, AMSTCNN integrates spatial and temporal information, uses weight matching to pay more attention to the region where human action occurs, and extracts region proposals of the image. Finally, we use Softmax to classify actions after spatial-temporal feature fusion. AMSTCNN can strongly make use of the spatial-temporal information of videos and select effective features to reduce computation. Experiments on AVA and the Chinese traffic police gesture datasets show that our method is superior to several state-of-the-art methods.

Pupil center recognition and location is an essential branch of ergonomics. It can be applied to emotion analysis and attention judgment. How to get the position of the pupil center from eye photos is the core of this field. Previous studies provided a helpful method, using scale-invariant feature transform (SIFT) to extract relevant features and combine them with the K-Nearest Neighbor (KNN) classifier. However, this method’s accuracy is not satisfying, and under some conditions, it will be position drift and other problems. We put forward a new idea to solve it by using Oriented FAST and Rotated BRIEF (ORB) features and Random Forest (RF) classifies. It is proved by experiment that our method improves the robustness of localization and the use of isophotes yields low computational cost, allowing for real-time processing. Meanwhile, we found that the ORB and RF are nearly as good, yielding an accuracy of 92.88% (BioID database).

Facial expression recognition (FER) has been a challenging task in the field of artificial intelligence. In this paper, we propose a novel model, named regional self-attention convolutional neural network (RSACNN), for FER. Different from the previous methods, RSACNN makes full use of the facial texture of expression salient region, so yields a robust feature representation for FER. The proposed model contains two novel parts: regional local multiple pattern (RLMP) based on the improved K-means algorithm and the regional self-attention module (RSAM). First, RLMP uses the improved K-means algorithm to dynamically cluster the pixels to ensure the robustness of texture features with expression salient variation. Besides, the texture description is enhanced by extending the binary pattern to the multiple patterns and integrating the information of gray difference between pixels in the region. Next, RSAM can adaptively form weights for each region through the self-attention mechanism, and use rank regularization loss (RRLoss) to constrain the weights of different regions. By jointly combining RLMP and RSAM, RSACNN can effectively enhance the feature representation of expression salient regions, so that the performance of expression recognition can be improved. Extensive experiments on public datasets, i.e. CK$+$, Oulu-CASIA, Fer2013 and SFEW, prove the superiority of our method over state-of-the-art approaches.

Vocal fold, a significant body structure, is accountable for phonation, which regulates air motion within and out of the lungs. The disorders in the vocal fold influence the quality of life. Thus, diagnosis of vocal fold disorders has a significant need, and CT of the neck is employed for an effective imaging scheme. Accordingly, this paper proposes an advanced multi-resolution algorithm (MRA) that optimally identifies and classifies pathologies. The vocal regions are acquired using the genetic k-means algorithm. The pathology features are generated using the local directional pattern (LDP) fed to pathology classification using moth search-rider optimization-based deep convolutional neural networks (MRA-based DCNN). The hybrid optimization (MRA), integrates the standard rider optimization algorithm (ROA) and moth search algorithm (MS) that trains deep learning classifier (DCNN). The analysis using the real databases regarding the performance metrics divulge that the proposed pathology detection module obtained the accuracy, specificity, and sensitivity of 97.020%, 91.698%, and 96.624%.

In passive location systems on the ground, the judgment and location of multi-target is more challenging compared with the case of single target. In this paper, we propose a method for multi-target identification and location in an arbitrary structure with three base stations (BSs). First of all, we discuss the scene of multi-targets judgment based on geometric dilution of precision (GDOP) value. Secondly, we propose an algorithm that calculates the system coverage radius based on arbitrary three BS structures. The algorithm helps to identify the number of targets for unsupervised learning. Finally, we locate each target individually located again based on the linear constrained minimum variance (LCMV) beam former and time difference of arrival (TDOA) algorithm. In the simulations, we analyzed the location dispersion under different signal-to-noise ratio (SNR), then calculated the termination threshold of the k-means algorithm under different SNR. The simulation results show that, compared to the probability hypothesis density (PHD) filter and TDOA-angle-of-arrival (AOA) joint algorithm, the proposed method can increase more than 12.5% and 15.6% points. With the increase of the number of targets, the running time of our algorithm is controllable with better stability.