Search

Along with the improvement of quality requirements industrial production, surface inspection of workpiece has gradually become an indispensable and important process in the production of the workpiece. Aiming at the traditional methods in textured paper inspection, there are problems of low efficiency and large error based on machine vision, we propose a "photometric stereo vision + fast Fourier enhancement + feature fusion" composite structure inspection method. First, as the traditional CCD camera produces obvious noise and scratches, which are difficult to distinguish from the background texture area, we propose combining the photometric stereo vision measurement algorithm to get the surface gradient information of the textured paper to obtain more gradient texture information; and then realize the secondary enhancement of the image through Fourier transform in spatial and frequency domains. Second, as the textured paper scratches are difficult to detect, the features are difficult to extract, and the threshold boundary is difficult to define, we propose dynamic threshold segmentation through multi-feature fusion to realize the surface scratch detection work of textured paper. We designed experiments using more than 300 different textured papers; and the results show that the composite structure detection method proposed in this paper is feasible and has advantages. [ABSTRACT FROM AUTHOR]

The growth in academic and scientific publications has increased very rapidly. Researchers must choose a representative and significant literature for their research, which has become challenging worldwide. Usually, the paper citation number indicates this paper's potential influence and importance. However, this standard metric of citation numbers is not suitable to assess the popularity and significance of recently published papers. To address this challenge, this study presents an effective prediction method called FutureCite to predict the future citation level of research articles. FutureCite integrates machine learning with text and graph mining techniques, leveraging their abilities in classification, datasets in-depth analysis, and feature extraction. FutureCite aims to predict future citation levels of research articles applying a multilabel classification approach. FutureCite can extract significant semantic features and capture the interconnection relationships found in scientific articles during feature extraction using textual content, citation networks, and metadata as feature resources. This study's objective is to contribute to the advancement of effective approaches impacting the citation counts in scientific publications by enhancing the precision of future citations. We conducted several experiments using a comprehensive publication dataset to evaluate our method and determine the impact of using a variety of machine learning algorithms. FutureCite demonstrated its robustness and efficiency and showed promising results based on different evaluation metrics. Using the FutureCite model has significant implications for improving the researchers' ability to determine targeted literature for their research and better understand the potential impact of research publications. [ABSTRACT FROM AUTHOR]

The Braille machines are expensive and as a result, are not accessible to many. In particular, there is a need for a portable text reader that is affordable and readily available to the blind community. The designed system is a portable, low-cost, reading device made for blind people. The designed system also overcomes the limitation of a conventional Braille machine by making it affordable for the common masses. This paper proposes a smart reader for visually challenged people using raspberry pi. This paper addresses the integration of a complete Text Read-out system designed for the visually challenged. The designed system consists of a webcam interfaced with raspberry pi which accepts a page of printed text. This system uses OCR technology to convert images into text and reads out the text by using Text-to-Speech conversion. This system supports audio output via Speakers as well as headphones. The OCR (Optical Character Recognition) package is installed in raspberry pi which scans it into a digital document which is then subjected to skew correction, and segmentation, before feature extraction to perform classification. Once classified, the text is read out by a text-to-speech conversion unit (TTS engine) installed in raspberry pi. The output is fed to an audio amplifier before it is read out. [ABSTRACT FROM AUTHOR]

As more people express their thoughts on products on various online shopping platforms, the feelings expressed in these opinions are becoming a significant source of information for marketers and buyers. These opinions have a big impact on consumers' decision to buy the best quality product. When there are too many features or a small number of records to analyze, the decision-making process gets difficult. A recent stream of study has used the conventional quantitative star score ratings and textual content reviews in this context. In this research, a decision-making framework is proposed that relies on feature-based opinions to analyze the textual content of reviews and classify buyer's opinions, thereby assisting consumers in making long-term purchases. The framework is proposed in this paper for product purchase decision making based on feature-based opinions and deep learning. Framework consists of four components: i) Pre-processing, ii) Feature extraction, iii) Feature-based opinion classification, and iv) Decision-making. Web scraping is used to obtain the dataset of Smartphone reviews, which is subsequently clean and pre-processed using tokenization and POS tagging. From the tagged dataset, noun labeled words are retrieved, and then the probable product's features are extracted. These feature-based sentences or reviews are processed using a word embedding to generate review vectors that identify contextual information. These word vectors are used to construct hidden vectors at the word and sentence levels using a hierarchical attention method. With respect to each feature, reviews are divided into five classes: extremely positive, positive, extremely negative, negative, and neutral. The proposed method may readily detect a customer's opinion on the quality of a product based on a certain attribute, which is beneficial in making a purchase choice. [ABSTRACT FROM AUTHOR]

Copyright of China Pulp & Paper is the property of China Pulp & Paper Magazines Publisher and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

To address the low level of intelligence and low utilization of logs in current rotary cutting equipment, this paper proposes a digital twin-based system for optimizing the rotary cutting of logs using a five-dimensional model of digital twins. The system features a log perception platform to capture three-dimensional point cloud data, outlining the logs' contours. Utilizing the Delaunay3D algorithm, this model performs a three-dimensional reconstruction of the log point cloud, constructing a precise digital twin. Feature information is extracted from the point cloud using the least squares method. Processing parameters, determined through the kinematic model, are verified in rotary cutting simulations via Bool operations. The system's efficacy has been substantiated through experimental validation, demonstrating its capability to output specific processing schemes for irregular logs and to verify these through simulation. This approach notably improves log recovery rates, decreasing volume error from 12.8% to 2.7% and recovery rate error from 23.5% to 5.7% The results validate the efficacy of the proposed digital twin system in optimizing the rotary cutting process, demonstrating its capability not only to enhance the utilization rate of log resources but also to improve the economic efficiency of the factory, thereby facilitating industrial development. [ABSTRACT FROM AUTHOR]

Single-frame infrared small target detection is affected by the low image resolution and small target size, and is prone to the problems of small target feature loss and positional offset during continuous downsampling; at the same time, the sparse features of the small targets do not correlate well with the global-local linkage of the background features. To solve the above problems, this paper proposes an efficient infrared small target detection method. First, this paper incorporates BlurPool in the feature extraction part, which reduces the loss and positional offset of small target features in the process of convolution and pooling. Second, this paper designs an interactive attention deep feature fusion module, which acquires the correlation information between the target and the background from a global perspective, and designs a compression mechanism based on deep a priori knowledge, which reduces the computational difficulty of the self-attention mechanism. Then, this paper designs the context local feature enhancement and fusion module, which uses deep semantic features to dynamically guide shallow local features to realize enhancement and fusion. Finally, this paper proposes an edge feature extraction module for shallow features, which utilizes the complete texture and location information in the shallow features to assist the network to initially locate the target position and edge shape. Numerous experiments show that the method in this paper significantly improves nIoU, F1-Measure and AUC on IRSTD-1k Datasets and NUAA-SIRST Datasets. [ABSTRACT FROM AUTHOR]

In recent years, the custom wardrobe market has been steadily developing. While meeting the functional needs of users, it is gradually shifting towards aesthetic preferences. Rapidly grasping users’ preferences for the appearance of custom wardrobes is a key focus of current research. This study collected a large number of decorative surface images of custom wardrobes and objectively analyzed the design features based on color moments and Tamura texture feature data in computer image analysis methods. K-means cluster analysis was performed on the feature data. Collected images of the points closest to the cluster centers were further screened to select representative finish images, and finally a questionnaire survey was conducted at Nanjing Forestry University, with the help of semantic differential method and factor analysis. The characteristics of the samples were comprehensively summarized to infer design elements. The study found that warm-toned, medium-low saturation, and medium brightness surfaces were preferred by the panel. Different colors, contrasts, saturations, brightness, element features, and arrangements have significantly different effects on visual perception. These conclusions can provide a reference for subsequent custom wardrobe design.

• Qualitative and quantitative detection of amine gases could be achieved by CSA. • A visible detection of beef freshness using the amine-responsive CSA was proposed. • ResNet34 had the best performance for beef freshness detection based on CSA. • T-SNE could further visualize and understand the classification process of DL. This study developed an innovative approach that combines a colourimetric sensor array (CSA) composed of twelve pH-response dyes with advanced algorithms, aiming to detect amine gases and assess the freshness of chilled beef. With the assistance of multivariate statistical analysis, the sensor array can effectively distinguish five amine gases and enable rapid quantification of trimethylamine vapour with a limit of detection (LOD) of 8.02 ppb and visually monitor the fresh levels of chilled beef. Moreover, the utilization of deep learning models (ResNet34, VGG16, and GoogleNet) for chilled beef freshness evaluation achieved an overall accuracy of 98.0 %. Furthermore, t -distributed stochastic neighbour embedding (t -SNE) visualized the feature extraction process and provided explanations to understand the classification process of deep learning. The results demonstrated that applying deep learning techniques in the process of pattern recognition of CSA can help in realizing the rapid, robust, and accurate assessment of chilled beef freshness. [ABSTRACT FROM AUTHOR]

With the popularisation of intelligent power, power devices have different shapes, numbers and specifications. This means that the power data has distributional variability, the model learning process cannot achieve sufficient extraction of data features, which seriously affects the accuracy and performance of anomaly detection. Therefore, this paper proposes a deep learning-based anomaly detection model for power data, which integrates a data alignment enhancement technique based on random sampling and an adaptive feature fusion method leveraging dimension reduction. Aiming at the distribution variability of power data, this paper developed a sliding window-based data adjustment method for this model, which solves the problem of high-dimensional feature noise and low-dimensional missing data. To address the problem of insufficient feature fusion, an adaptive feature fusion method based on feature dimension reduction and dictionary learning is proposed to improve the anomaly data detection accuracy of the model. In order to verify the effectiveness of the proposed method, we conducted effectiveness comparisons through elimination experiments. The experimental results show that compared with the traditional anomaly detection methods, the method proposed in this paper not only has an advantage in model accuracy, but also reduces the amount of parameter calculation of the model in the process of feature matching and improves the detection speed. [ABSTRACT FROM AUTHOR]

In this study, a prediction model for visual fatigue is developed. As input, frequential and nonlinear features are extracted from multichannel EEG, and then dimensionally reduced. In the model, bidirectional LSTM and attention layers are combined for effective learning. As a result, 82.90% accuracy, 85.26% weighted precision, 82.90% weighted recall, and 84.02% weighted F1‐score were obtained. [ABSTRACT FROM AUTHOR]

The Ultra-Wideband (UWB) Location-Based Service is receiving more and more attention due to its high ranging accuracy and good time resolution. However, the None-Line-of-Sight (NLOS) propagation may reduce the ranging accuracy for UWB localization system in indoor environment. So it is important to identify LOS and NLOS propagations before taking proper measures to improve the UWB localization accuracy. In this paper, a deep learning-based UWB NLOS/LOS classification algorithm called FCN-Attention is proposed. The proposed FCN-Attention algorithm utilizes a Fully Convolution Network (FCN) for improving feature extraction ability and a self-attention mechanism for enhancing feature description from the data to improve the classification accuracy. The proposed algorithm is evaluated using an open-source dataset, a local collected dataset and a mixed dataset created from these two datasets. The experiment result shows that the proposed FCN-Attention algorithm achieves classification accuracy of 88.24% on the open-source dataset, 100% on the local collected dataset and 92.01% on the mixed dataset, which is better than the results from other evaluated NLOS/LOS classification algorithms in most scenarios in this paper. [ABSTRACT FROM AUTHOR]

In this paper, we propose a novel method for monocular depth estimation using the hourglass neck module. The proposed method has the following originality. First, feature maps are extracted from Swin Transformer V2 using a masked image modeling (MIM) pretrained model. Since Swin Transformer V2 has a different patch size for each attention stage, it is easier to extract local and global features from images input by the vision transformer (ViT)-based encoder. Second, to maintain the polymorphism and local inductive bias of the feature map extracted from Swin Transformer V2, a feature map is input into the hourglass neck module. Third, deformable attention can be used at the waist of the hourglass neck module to reduce the computation cost and highlight the locality of the feature map. Finally, the feature map traverses the neck and proceeds through a decoder, comprised of a deconvolution layer and an upsampling layer, to generate a depth image. To evaluate the objective reliability of the proposed method in this paper, we used the NYU Depth V2 dataset to compare and evaluate the methods published in other papers. As a result of the experiment, the RMSE value of the novel method for monocular depth estimation using the hourglass neck module proposed in this paper was 0.274, which was lower than those published in other papers. The lower the RMSE value, the better the depth estimation method; therefore, its efficiency compared to other techniques has been proven. [ABSTRACT FROM AUTHOR]

Millimeter-wave radar-based identification technology has a wide range of applications in persistent identity verification, covering areas such as security production, healthcare, and personalized smart consumption systems. It has received extensive attention from the academic community due to its advantages of being non-invasive, environmentally insensitive and privacy-preserving. Existing identification algorithms mainly rely on a single signal, such as breathing or heartbeat. The reliability and accuracy of these algorithms are limited due to the high similarity of breathing patterns and the low signal-to-noise ratio of heartbeat signals. To address the above issues, this paper proposes an algorithm for multimodal fusion for identity recognition. This algorithm extracts and fuses features derived from phase signals, respiratory signals, and heartbeat signals for identity recognition purposes. The spatial features of signals with different modes are first extracted by the residual network (ResNet), after which these features are fused with a spatial-channel attention fusion module. On this basis, the temporal features are further extracted with a time series-based self-attention mechanism. Finally, the feature vectors of the user's vital sign modality are obtained to perform identity recognition. This method makes full use of the correlation and complementarity between different modal signals to improve the accuracy and reliability of identification. Simulation experiments show that the algorithm identity recognition proposed in this paper achieves an accuracy of 94.26% on a 20-subject self-test dataset, which is much higher than that of the traditional algorithm, which is about 85%. [ABSTRACT FROM AUTHOR]

Dissolved oxygen is one of the most important water quality parameters in aquaculture, and the level determines whether fish can grow healthily. Since there is a delay in equipment control in the aquaculture environment, dissolved oxygen prediction is needed to reduce the loss due to low dissolved oxygen. To solve the problem of insufficient accuracy and poor interpretability of traditional methods in predicting dissolved oxygen from multivariate water quality parameters, this paper proposes an improved N-Beats-based prediction network. First, the maximum expectation algorithm [expectation–maximization (EM)] was used to fill in the original data by fitting the missing values. Second, the discrete wavelet transform (DWT) was used to reduce the overall noise of the sample, then the gated recurrent unit (GRU) feature extraction network was employed to extract the water quality information from the temporal dimension, the N-Beats was utilized to predict the preprocessed data, and the residual operation through Stack was performed to obtain the prediction results. The improved algorithm overcomes the challenge of insufficient prediction accuracy of the traditional algorithm. The GRU–N-Beats network proposed in this paper can extract features from multivariate time dimensions for prediction. The values of root mean square error (RMSE), mean absolute error (MAE), mean absolute percentage error (MAPE), and R² for the proposed algorithm were 0.171, 0.120, 0.015, and 0.97, respectively. In particular, they were 28.5%, 32.1%, 51.6%, 24.3%, 14.9%, 36.4%, and 19.3% higher than those of long short-term memory (LSTM), GRU, temporal convolutional network (TCN), LSTM–TCN, PatchTST, backpropagation neural network (BPNN), and N-Beats on RMSE, respectively. [ABSTRACT FROM AUTHOR]

Due to the large intra-class differences between the same categories and the scale imbalance between different categories in the remote sensing image dataset, the semantic segmentation task presents the problem of small-scale object information loss, the imbalance between foreground and background, and simultaneously the background dominates, which seriously affects the performance of the network model. To solve the above problems, this paper proposes an efficient bilateral branch depth neural network model based on the U-Net depth neural network, named BBU-Net. Firstly, one branch of the network learns the distribution characteristics of the original data, and the other focuses on difficult samples. Then the two branches improve the representation and classification ability of the neural network by accumulating learning strategies. Finally, considering the geometric diversity of remote sensing images, this paper adopts test time augmentation and reflection padding strategies and proposes a balanced weighted loss function named CombineLoss to alleviate the imbalance in the training process. The depth neural network proposed in this paper was first tested on the Inria Aerial Image Labeling Dataset, and 87.53% of mean intersection over union and 97.4% of mean pixel accuracy were obtained, respectively. At the same time, to verify the model's complexity, the model proposed in this paper is compared with the neural network based on integrated learning. The comparison results show that the spatial complexity of the network proposed in this paper is much lower than the neural network obtained by integrated learning, and the parameters are also much smaller than the neural network based on integrated learning. Then use the satellite building dataset I in the WHU Building Dataset and mainstream semantic segmentation methods for multiple groups of comparative experiments. The experimental results show that the method proposed in this paper can effectively extract the semantic information of remote sensing images, significantly improve the imbalance of remote sensing image data, improve the performance of the network model, and achieve a good semantic segmentation effect, which fully proves the effectiveness of this method. [ABSTRACT FROM AUTHOR]

This paper introduces an innovative method to improve loop closure detection within the domain of Simultaneous Localization And Mapping (SLAM) by integrating a Multi-Scale Attention and Semantic Guidance (MSA-SG) framework. In SLAM systems, accurate loop closure detection is essential for minimizing localization errors over time and ensuring the reliability of the constructed maps in robotics navigation through uncharted environments. Our proposed method utilizes EfficientNet-EA for robust feature extraction and introduces MSA-SG, a novel mechanism that synergizes multiscale attention with semantic guidance to focus on critical semantic features essential for loop closure detection. This approach ensures the prioritization of static environmental landmarks over transient and irrelevant objects, significantly enhancing the accuracy and efficiency of loop closure detection in complex and dynamic settings. Experimental validations on recognized datasets underscore the superiority of our approach, demonstrating marked improvements in precision, recall, and overall SLAM performance. This research highlights the significant benefits of leveraging semantic insights and attentional focus in advancing the capabilities of loop closure detection for SLAM applications. [ABSTRACT FROM AUTHOR]

In response to the issue that the fusion process of infrared and visible images is easily affected by lighting factors, in this paper, we propose an adaptive illumination perception fusion mechanism, which was integrated into an infrared and visible image fusion network. Spatial attention mechanisms were applied to both infrared images and visible images for feature extraction. Deep convolutional neural networks were utilized for further feature information extraction. The adaptive illumination perception fusion mechanism is then integrated into the image reconstruction process to reduce the impact of lighting variations in the fused images. A Median Strengthening Channel and Spatial Attention Module (MSCS) was designed to be integrated into the backbone of YOLOv8. In this paper, we used the fusion network to create a dataset named ivifdata for training the target recognition network. The experimental results indicated that the improved YOLOv8 network saw further enhancements of 2.3%, 1.4%, and 8.2% in the Recall, mAP50, and mAP50-95 metrics, respectively. The experiments revealed that the improved YOLOv8 network has advantages in terms of recognition rate and completeness, while also reducing the rates of false negatives and false positives. [ABSTRACT FROM AUTHOR]

Defect detection constitutes one of the most crucial processes in industrial production. With a continuous increase in the number of defect categories and samples, the defect detection model underpinned by deep learning finds it challenging to expand to new categories, and the accuracy and real-time performance of product defect detection are also confronted with severe challenges. This paper addresses the problem of insufficient detection accuracy of existing lightweight models on resource-constrained edge devices by presenting a new lightweight YoloV5 model, which integrates four modules, SCDown, GhostConv, RepNCSPELAN4, and ScalSeq. Here, this paper abbreviates it as SGRS-YoloV5n. Through the incorporation of these modules, the model notably enhances feature extraction and computational efficiency while reducing the model size and computational load, making it more conducive for deployment on edge devices. Furthermore, a cloud-edge collaborative defect detection system is constructed to improve detection accuracy and efficiency through initial detection by edge devices, followed by additional inspection by cloud servers. An incremental learning mechanism is also introduced, enabling the model to adapt promptly to new defect categories and update its parameters accordingly. Experimental results reveal that the SGRS-YoloV5n model exhibits superior detection accuracy and real-time performance, validating its value and stability for deployment in resource-constrained environments. This system presents a novel solution for achieving efficient and accurate real-time defect detection. [ABSTRACT FROM AUTHOR]

In pediatric rehabilitation medicine, manual assessment methods for visual–motor integration result in inconsistent scoring standards. To address these issues, incorporating artificial intelligence (AI) technology is a feasible approach that can reduce time and improve accuracy. Existing research on visual–motor integration scoring has proposed a framework based on convolutional neural networks (CNNs) for the Beery–Buktenica developmental test of visual–motor integration. However, as the number of training questions increases, the accuracy of this framework significantly decreases. This paper proposes a new architecture to reduce the number of features, channels, and overall model complexity. The architectureoptimizes input features by concatenating question numbers with answer features and selecting appropriate channel ratios and optimizes the output vector by designing the task as a multi-class classification. This paper also proposes a model named improved DenseNet. After experimentation, DenseNet201 was identified as the most suitable pre-trained model for this task and was used as the backbone architecture for improved DenseNet. Additionally, new fully connected layers were added for feature extraction and classification, allowing for specialized feature learning. The architecture can provide reasons for unscored results based on prediction results and decoding rules, offering directions for children's training. The final experimental results show that the proposed new architecture improves the accuracy of scoring 6 question graphics by 12.8% and 12 question graphics by 20.14% compared to the most relevant literature. The accuracy of the proposed new architecture surpasses the model frameworks of the most relevant literature, demonstrating the effectiveness of this approach in improving scoring accuracy and stability. [ABSTRACT FROM AUTHOR]

In order to solve the problem of inaccurate image feature extraction caused by traditional extraction methods, this paper proposes a remote sensing image size extraction method based on nonlinear multi feature fusion for topographic maps. In this paper, SVM and DS evidence theory are combined to extract image features and classify pre processed remote sensing images. Based on the classification results, basic probability distributions are constructed, and a DS fusion algorithm using matrix analysis is introduced to simplify the complexity of decision level fusion algorithms; We use a multi feature fusion algorithm based on feature proximity, using the proximity vector formed by the attraction between the feature vector and the original graphics pattern as the fusion feature to complete the extraction of remote sensing image features. The simulation results show that after using this method, its soft threshold classifier outputs 0.9865, 0.9965, 0.7852, 0.9921, 0.9847, 0.6879, -0.5898, -0.5678, -0.6897, -0.4785. The algorithm in this paper can distinguish the shape features of terrain images well, and can extract the features of terrain images more accurately, which has strong feasibility. [ABSTRACT FROM AUTHOR]

The performance of the existing student classroom behavior detection model is affected by various aspects such as dataset, algorithm and height as well as the differences between different classrooms, and there are problems such as a single dataset, low accuracy and low efficiency. In order to improve the accuracy of student classroom behavior detection algorithm, this paper proposes a student classroom behavior detection method based on improved convolutional neural network algorithm. Firstly, the student behavior detection dataset is constructed, and the student classroom behavior detection technology scheme is designed; secondly, in order to improve the detection accuracy, the features are extracted by using the new jumping bi-directional paths, and the attention mechanism module is added at different positions to improve the path aggregation network; weekly, the embedding positions of the attention mechanism strategy are determined by analyzing multiple sets of experiments, and the proposed student classroom behavior detection algorithm's effectiveness and superiority. [ABSTRACT FROM AUTHOR]

This paper proposes applying a novel deep‐learning model, TBDLNet, to recognize CT images to classify multidrug‐resistant and drug‐sensitive tuberculosis automatically. The pre‐trained ResNet50 is selected to extract features. Three randomized neural networks are used to alleviate the overfitting problem. The ensemble of three RNNs is applied to boost the robustness via majority voting. The proposed model is evaluated by five‐fold cross‐validation. Five indexes are selected in this paper, which are accuracy, sensitivity, precision, F1‐score, and specificity. The TBDLNet achieves 0.9822 accuracy, 0.9815 specificity, 0.9823 precision, 0.9829 sensitivity, and 0.9826 F1‐score, respectively. The TBDLNet is suitable for classifying multidrug‐resistant tuberculosis and drug‐sensitive tuberculosis. It can detect multidrug‐resistant pulmonary tuberculosis as early as possible, which helps to adjust the treatment plan in time and improve the treatment effect. [ABSTRACT FROM AUTHOR]

Semantic segmentation of target objects in power transmission line corridor point cloud scenes is a crucial step in powerline tree barrier detection. The massive quantity, disordered distribution, and non-uniformity of point clouds in power transmission line corridor scenes pose significant challenges for feature extraction. Previous studies have often overlooked the core utilization of spatial information, limiting the network's ability to understand complex geometric shapes. To overcome this limitation, this paper focuses on enhancing the deep expression of spatial geometric information in segmentation networks and proposes a method called BDF-Net to improve RandLA-Net. For each input 3D point cloud data, BDF-Net first encodes the relative coordinates and relative distance information into spatial geometric feature representations through the Spatial Information Encoding block to capture the local spatial structure of the point cloud data. Subsequently, the Bilinear Pooling block effectively combines the feature information of the point cloud with the spatial geometric representation by leveraging its bilinear interaction capability thus learning more discriminative local feature descriptors. The Global Feature Extraction block captures the global structure information in the point cloud data by using the ratio between the point position and the relative position, so as to enhance the semantic understanding ability of the network. In order to verify the performance of BDF-Net, this paper constructs a dataset, PPCD, for the point cloud scenario of transmission line corridors and conducts detailed experiments on it. The experimental results show that BDF-Net achieves significant performance improvements in various evaluation metrics, specifically achieving an OA of 97.16%, a mIoU of 77.48%, and a mAcc of 87.6%, which are 3.03%, 16.23%, and 18.44% higher than RandLA-Net, respectively. Moreover, comparisons with other state-of-the-art methods also verify the superiority of BDF-Net in point cloud semantic segmentation tasks. [ABSTRACT FROM AUTHOR]

Logistics parcel detection technology is critical for unmanned sorting. YOLOv5, the state-of-the-art (SOTA) object detection model, is a classic network widely used in engineering. However, on the premise of fast and accurate detection of logistics parcels, it confronts certain challenges of high computing load and parameter quantity. To address these issues, this paper proposes a two-scale lightweight deep learning model named SFYOLOv5. In this article, a lightweight feature extraction module called Pruned-Shuffle-Block (PSB) is proposed. Meanwhile, a double-layer pyramid structure for the medium and large target detection is created in accordance with the image size distribution of logistics parcels. With this structure, the Floating Point of Operations (FLOPs) and parameters in feature extraction were significantly reduced. In addition, a down sampling module named Focus for Downsampling (FFD) is designed and attention modules are introduced to extract high-level semantic information in logistics parcels. These modules not only compensate for the loss caused by down sampling but also improve the mean Average Precision (mAP). Finally, the comparison experiment is performed by using the self-built logistics parcel dataset. The results show that the mAP of the model reaches 99.1%, the number of model parameters decreased by 92.14%, and the FLOPs decreased by 89.57% compared with the existing SOTA model. This model can be used in logistics parcel intelligent sorting. [ABSTRACT FROM AUTHOR]

Automatic Speech Emotion Recognition (ASER) has recently garnered attention across various fields including artificial intelligence, pattern recognition, and human–computer interaction. However, ASER encounters numerous challenges such as a shortage of diverse datasets, appropriate feature selection, and suitable intelligent recognition techniques. To address these challenges, a systematic literature review (SLR) was conducted following established guidelines. A total of 60 primary research papers spanning from 2011 to 2023 were reviewed to investigate, interpret, and analyze the related literature by addressing five key research questions. Despite being an emerging area with applications in real-life scenarios, ASER still grapples with limitations in existing techniques. This SLR provides a comprehensive overview of existing techniques, datasets, and feature extraction tools in the ASER domain, shedding light on the weaknesses of current research studies. Additionally, it outlines a list of limitations for consideration in future work. [ABSTRACT FROM AUTHOR]

Manuscript binarization is used to convert each pixel in the script image into text and background. Many manuscript binarization methods have been proposed, such as the Otsu, Bernsen, Sauvola, Niblack, Phansalkar and Singh methods. These methods only focus on one problem of a degraded manuscript. In this research, a deep learning approach based on the U-Net method is applied for binarization of degraded ancient manuscripts. Adding layers to the U-Net architecture can cause more parameters and excessive computational calculations. Residual U-Net (ResUNet) is a development of the U-Net method. ResUNet, with its residual blocks, enables efficient and effective feature extraction, capturing fine details of degraded documents. This is important for identifying and distinguishing text from various artifacts and noise in the document. ResUNet can handle various types of image degradation thanks to its residual blocks that prevent gradient loss and strengthen features over the network. Convolutional Long Short-Term Memory (ConvLSTM) is a variant of LSTM (Long Short-Term Memory) designed for spatial data such as images. ConvLSTM combines the ability of LSTM to learn long-term dependencies with the power of CNN in processing spatial data. The combination of ResUNet and ConvLSTM for binarization of degraded documents is a powerful strategy that leverages the power of both architectures to improve quality and accuracy in separating text from degraded background. The aim of this research is to determine the performance evaluation results of the combination of ResUNet and ConvLSTM architectures on the binarization of degraded ancient Javanese manuscripts. The trial was conducted using datasets taken from several museums. The dataset consists of 1200 images of Javanese ancient manuscripts that were damaged in the form of perforated paper, ink bleed through from the previous page, and red or brownish spots. The proposed method produces a loss value of 0.0559, F-Measure 92.89%, PSNR 18.52 dan IoU 0.85. [ABSTRACT FROM AUTHOR]

An optimization algorithm based on the LMPEC algorithm is proposed to rectify the nameplate image to address the problem that overexposure and underexposure of the nameplate image of electrical equipment will make subsequent nameplate recognition difficult. In the network structure, the PS-UNet++ network is based on the sub-pixel convolution upsampling module, and the UNet++ network is constructed as the feature extraction sub-network of the optimization algorithm to extract more detailed information from the model. Smooth L1 loss is substituted for L1 loss in the loss function to prevent model oscillation. In addition, to increase the robustness of the model, an improved method built on the multi-scale training method is applied. The experimental results indicate that, among all comparison algorithms, the optimized algorithm performs the best on the data set of electrical equipment nameplate exposure the experimenter generated. Compared to the original LMPEC algorithm, the SSIM, PSNR, and PI image evaluation indices are enhanced by 5.6%, 5.1%, and 7.96%, respectively. [ABSTRACT FROM AUTHOR]

Advanced Persistent Threat (APT) attacks are causing a lot of damage to critical organizations and institutions. Therefore, early detection and warning of APT attack campaigns are very necessary today. In this paper, we propose a new approach for APT attack detection based on the combination of Feature Intelligent Extraction (FIE) and Representation Learning (RL) techniques. In particular, the proposed FIE technique is a combination of the Bidirectional Long Short-Term Memory (BiLSTM) deep learning network and the Attention network. The FIE combined model has the function of aggregating and extracting unusual behaviors of APT IPs in network traffic. The RL method proposed in this study aims to optimize classifying APT IPs and normal IPs based on two main techniques: rebalancing data and contrastive learning. Specifically, the rebalancing data method supports the training process by rebalancing the experimental dataset. And the contrastive learning method learns APT IP's important features based on finding and pulling similar features together as well as pushing contrasting data points away. The combination of FIE and RL (abbreviated as the FIERL model) is a novel proposal and innovation and has not been proposed and published by any research. The experimental results in the paper have proved that the proposed method in the paper is correct and reasonable when it has shown superior efficiency compared to some other studies and approaches over 5% on all measurements. [ABSTRACT FROM AUTHOR]

In this paper, a method based on transfer learning is proposed to recover the three-dimensional shape of cultural relics faces from a single old photo. It can simultaneously reconstruct the three-dimensional facial structure and align the texture of the cultural relics with fewer training samples. The UV position map is used to represent the three-dimensional shape in space and act as the output of the network. A convolutional neural network is used to reconstruct the UV position map from a single 2D image. In the training process, the human face data is used for pre-training, and then a small amount of artifact data is used for fine-tuning. A deep learning model with strong generalization ability is trained with fewer artifact data, and a three-dimensional model of the cultural relic face can be reconstructed from a single old photograph. The methods can train more complex deep networks without a large amount of cultural relic data, and no over-fitting phenomenon occurs, which effectively solves the problem of fewer cultural relic samples. The method is verified by restoring a Chinese Terracotta Warrior with facial cracks and a Buddha with a broken nose. Other applications can be used in the fields such as texture recovery, facial feature extraction, and three-dimensional model estimation of the damaged cultural relics or sculptures in the photos. [ABSTRACT FROM AUTHOR]

In this paper, we introduce a novel Multi-scale and Auto-tuned Semi-supervised Deep Subspace Clustering (MAS-DSC) algorithm, aimed at addressing the challenges of deep subspace clustering in high-dimensional real-world data, particularly in the field of medical imaging. Traditional deep subspace clustering algorithms, which are mostly unsupervised, are limited in their ability to effectively utilize the inherent prior knowledge in medical images. Our MAS-DSC algorithm incorporates a semi-supervised learning framework that uses a small amount of labeled data to guide the clustering process, thereby enhancing the discriminative power of the feature representations. Additionally, the multi-scale feature extraction mechanism is designed to adapt to the complexity of medical imaging data, resulting in more accurate clustering performance. To address the difficulty of hyperparameter selection in deep subspace clustering, this paper employs a Bayesian optimization algorithm for adaptive tuning of hyperparameters related to subspace clustering, prior knowledge constraints, and model loss weights. Extensive experiments on standard clustering datasets, including ORL, Coil20, and Coil100, validate the effectiveness of the MAS-DSC algorithm. The results show that with its multi-scale network structure and Bayesian hyperparameter optimization, MAS-DSC achieves excellent clustering results on these datasets. Furthermore, tests on a brain tumor dataset demonstrate the robustness of the algorithm and its ability to leverage prior knowledge for efficient feature extraction and enhanced clustering performance within a semi-supervised learning framework. [ABSTRACT FROM AUTHOR]

Because of the difficulty in fault detection for and diagnosing the adjustment hydraulic servomotor, this paper uses feature extraction technology to extract the time domain and frequency domain features of the pressure signal of the adjustment hydraulic servomotor and splice the features of multiple pressure signals through the Multi-source Information Fusion (MSIF) method. The comprehensive expression of device status information is obtained. After that, this paper proposes a fault detection Algorithm GA-SVDD-neg, which uses Genetic Algorithm (GA) to optimize Support Vector Data Description with negative examples (SVDD-neg). Through joint optimization with the Mutual Information (MI) feature selection algorithm, the features that are most sensitive to the state deterioration of the adjustment hydraulic servomotor are selected. Experiments show that the MI algorithm has a better performance than other feature dimensionality reduction algorithms in the field of the abnormal detection of adjustment hydraulic servomotors, and the GA-SVDD-neg algorithm has a stronger robustness and generality than other anomaly detection algorithms. In addition, to make full use of the advantages of deep learning in automatic feature extraction and classification, this paper realizes the fault diagnosis of the adjustment hydraulic servomotor based on 1D Convolutional Neural Network (1DCNN). The experimental results show that this algorithm has the same superior performance as the traditional algorithm in feature extraction and can accurately diagnose the known faults of the adjustment hydraulic servomotor. This research is of great significance for the intelligent transformation of adjustment hydraulic servomotors and can also provide a reference for the fault warning and diagnosis of the Electro-Hydraulic (EH) system of the same type of steam turbine. [ABSTRACT FROM AUTHOR]

The fire heat release rate (HRR) is a crucial parameter for describing the combustion process and its thermal effects. In recent years, some studies have employed fire scene images and deep learning algorithms to predict real-time fire HRR, which has led to the advancement of HRR prediction in terms of both lightweightness and real-time monitoring. Nevertheless, the development of an early-stage monitoring system for fires and the ability to predict future HRR based on current moment data represents a crucial foundation for evaluating the scale of indoor fires and enhancing the capacity to prevent and control such incidents. This paper proposes a deep learning model based on continuous fire scene images (containing both flame and smoke features) and their time-series information to predict the future transient fire HRR. The model (Att-BiLSTM) comprises three bi-directional long- and short-term memory (Bi-LSTM) layers and one attention layer. The model employs a bidirectional feature extraction approach, followed by the introduction of an attention mechanism to highlight the image features that have a critical impact on the prediction results. In this paper, a large-scale dataset is constructed by collecting 27,231 fire scene images with instantaneous HRR annotations from 40 different fire trials from the NIST database. The experimental results demonstrate that Att-BiLSTM is capable of effectively utilizing fire scene image features and temporal information to accurately predict future transient HRR, including those in high-brightness fire environments and complex fire source situations. The research presented in this paper offers novel insights and methodologies for fire monitoring and emergency response. [ABSTRACT FROM AUTHOR]

Target detection algorithms can greatly improve the efficiency of tomato leaf disease detection and play an important technical role in intelligent tomato cultivation. However, there are some challenges in the detection process, such as the diversity of complex backgrounds and the loss of leaf symmetry due to leaf shadowing, and existing disease detection methods have some disadvantages in terms of deteriorating generalization ability and insufficient accuracy. Aiming at the above issues, a target detection model for tomato leaf disease based on deep learning with a global attention mechanism, TDGA, is proposed in this paper. The main idea of TDGA includes three aspects. Firstly, TDGA adds a global attention mechanism (GAM) after up-sampling and down-sampling, as well as in the SPPF module, to improve the feature extraction ability of the target object, effectively reducing the interference of invalid targets. Secondly, TDGA uses a switchable atrous convolution (SAConv) in the C3 module to improve the model's ability to detect. Thirdly, TDGA adopts the efficient IoU loss (EIoU) instead of complete IoU loss (CIoU) to solve the ambiguous definition of aspect ratio and sample imbalance. In addition, the influences of different environmental factors such as single leaf, multiple leaves, and shadows on the performance of tomato disease detection are extensively experimented with and analyzed in this paper, which also verified the robustness of TDGA. The experimental results show that the average accuracy of TDGA reaches 91.40%, which is 2.93% higher than that of the original YOLOv5 network, which is higher than YOLOv5, YOLOv7, YOLOHC, YOLOv8, SSD, Faster R-CNN, RetinaNet and other target detection networks, so that TDGA can be utilized for the detection of tomato leaf disease more efficiently and accurately, even in complex environments. [ABSTRACT FROM AUTHOR]

Malicious encrypted traffic detection is a critical component of network security management. Previous detection methods can be categorized into two classes as follows: one is to use the feature engineering method to construct traffic features for classification and the other is to use the end‐to‐end method that directly inputs the original traffic to obtain traffic features for classification. Both of the abovementioned two methods have the problem that the obtained features cannot fully characterize the traffic. To this end, this paper proposes a hierarchical multimodal deep learning model (HMMED) for malicious encrypted traffic detection. This model adopts the abovementioned two feature generation methods to learn the features of payload and header, respectively, then fuses the features to get the final traffic features, and finally inputs the final traffic features into the softmax classifier for classification. In addition, since traditional deep learning is highly dependent on the training set size and data distribution, resulting in a model that is not very generalizable and difficult to adapt to unseen encrypted traffic, the model proposed in this paper uses a large amount of unlabeled encrypted traffic in the pretraining layer to pretrain a submodel used to obtain a generic packet payload representation. The test results on the USTC‐TFC2016 dataset show that the proposed model can effectively solve the problem of insufficient feature extraction of traditional detection methods and improve the ACC of malicious encrypted traffic detection. [ABSTRACT FROM AUTHOR]

To digitize and preserve the cultural heritage in the form of Indian classical dance become apparent area of research. Sattriya classical dance of North-East India (Assam) is one of the eight Indian classical dance forms that requires immediate preservation. Sattriya classical dance consists of 29 Asamyukta hastas (single-hand gestures) and 14 Samyukta hastas (double-hand gestures). Moreover, the foundation of Samyukta hasta depends on understanding Asamyukta hasta. Therefore, the paper aims to classify single-hand gestures of Sattriya classical dance only. Although, a solution based on two level classification method to classify the Sattriya classical dance is available in recent literature, but it requires a trial and error method to select the optimized features. Since, Asamyukta hastas can appear closely similar to each other and therefore misclassification chances are very high. Thus, accuracy rate obtained for the two level classification method was only 75.45%. So, to address this issues in this paper, a Multilevel Classification Model with Vision based Features (MCM- V b F) has been proposed to classify the Asamyukta hastas of Sattriya classical dance. This model uses two types of feature matching, viz., high-level feature matching and low-level feature matching. To extract the high-level features and low-level features different algorithm has been proposed. In this model, features are automatically selected. This proposed MCM- V b F model is also tested on Asamyukta hasta mudras of Bharatanatyam classical dance of South India (Tamil Nadu). This model obtain an accuracy 94.12%, 87.14% for Sattriya classical dance Single-Hand Gestures (SSHG) dataset and Bharatnatyam classical dance Single-Hand Gestures (BHSG) dataset respectively. This paper also provides the comparative study of the proposed model MCM- V b F with traditional bench-mark classifier model such as Naive Bayes, Decision Tree and Support Vector Classifier (SVM) etc. Article highlights: A Multilevel Classification Model with Vision based Features (MCM-VbF) has been proposed to classify twenty nine (29) single-hand gestures of Sattriya classical dance. To enhance this proposed model eight vision based features has been extracted. Details of each features extraction method have been explained. This feature has been explained. This feature has Evaluating the proposed MCM-VbF model on three benchmark classifier such as Naive bayes, Decision tree and Support Vector Machine (SVM) which gives promising performance for all cases. The proposed model gives better performance for both own generated Sattriya classical dance Single-Hand Gestures (SSHG) dataset as well as Bharatnatyam classical dance Single-Hand Gestures (BSHG) dataset. [ABSTRACT FROM AUTHOR]

As unmanned vehicle technology advances rapidly, obstacle recognition and target detection are crucial links, which directly affect the driving safety and efficiency of unmanned vehicles. In response to the inaccurate localization of small targets such as pedestrians in current object detection tasks and the problem of losing local features in the PointPillars, this paper proposes a three-dimensional object detection method based on improved PointPillars. Firstly, addressing the issue of lost spatial and local information in the PointPillars, the feature encoding part of the PointPillars is improved, and a new pillar feature enhancement extraction module, CSM-Module, is proposed. Channel encoding and spatial encoding are introduced in the new pillar feature enhancement extraction module, fully considering the spatial information and local detailed geometric information of each pillar, thereby enhancing the feature representation capability of each pillar. Secondly, based on the fusion of CSPDarknet and SENet, a new backbone network CSE-Net is designed in this paper, enabling the extraction of rich contextual semantic information and multi-scale global features, thereby enhancing the feature extraction capability. Our method achieves higher detection accuracy when validated on the KITTI dataset. Compared to the original network, the improved algorithm's average detection accuracy is increased by 3.42%, it shows that the method is reasonable and valuable. [ABSTRACT FROM AUTHOR]

The existing deep estimation networks often overlook the issue of computational efficiency while pursuing high accuracy. This paper proposes a lightweight self-supervised network that combines convolutional neural networks (CNN) and Transformers as the feature extraction and encoding layers for images, enabling the network to capture both local geometric and global semantic features for depth estimation. First, depth-separable convolution is used to construct a dilated convolution residual module based on a shallow network to improve the shallow CNN feature extraction receptive field. In the transformer, a multidepth separable convolution head transposed attention module is proposed to reduce the computational burden of spatial self-attention. In the feedforward network, a two-step gating mechanism is proposed to improve the nonlinear representation ability of the feedforward network. Finally, the CNN and transformer are integrated to implement a depth estimation network with a local-global context interaction function. Compared with other lightweight models, this model has fewer model parameters and higher estimation accuracy. It also has better generalizability for different outdoor datasets. Additionally, the inference speed can reach 87 FPS, achieving better real-time performance and accounting for both inference speed and estimation accuracy. [ABSTRACT FROM AUTHOR]

Deep learning-based hyperspectral remote sensing image classification methods are currently a research hotspot. However, they suffer from issues such as large feature network parameter size, complex calculations, and the need for a large number of training data to achieve good classification results. Moreover, hyperspectral remote sensing images face challenges such as difficulty in obtaining the ground truth of land cover, limited availability of effective datasets for training, and endmember spectral variability, making it difficult for existing algorithm models to be widely adopted. To address these issues, this paper proposes a multi-branch classification model with multi-dimensional feature fusion, constructing lightweight deep network models for one-dimensional spectral, two-dimensional spatial, and three-dimensional depth feature extraction, respectively. This enriches feature information while reducing the parameters of each branch's deep model, effectively improving the land cover classification accuracy using hyperspectral remote sensing images under limited training sample conditions. Experimental verification with open-source hyperspectral remote sensing datasets shows that the proposed classification method can obtain over 90% classification accuracy when the training set account for only 5% of the total dataset, which is significantly better than current mainstream deep network classification models. [ABSTRACT FROM AUTHOR]

The model network based on YOLOv8 for detecting race cones and buckets in the Formula Unmanned Competition for Chinese university students needs help with problems with complex structure, redundant number of parameters, and computation, significantly affecting detection efficiency. A lightweight detection model based on YOLOv8 is proposed to address these problems. The model includes improving the backbone network, neck network, and detection head, as well as introducing knowledge distillation and other techniques to construct a lightweight model. The specific improvements are as follows: firstly, the backbone network for extracting features is improved by introducing the ADown module in YOLOv9 to replace the convolution module used for downsampling in the YOLOv8 network, and secondly, the FasterBlock in FasterNet network was introduced to replace the fusion module in YOLOv8 C2f, and then the self-developed lightweight detection head was introduced to improve the detection performance while achieving lightweight. Finally, the detection performance was further improved by knowledge distillation. The experimental results on the public dataset FSACOCO show that the improved model's accuracy, recall, and average precision are 92.7%, 84.6%, and 91%, respectively. Compared with the original YOLOv8n detection model, the recall and average precision increase by 2.7 and 1.2 percentage points, the memory is half the original, and the model computation is 51%. The model significantly reduces the misdetection and leakage of conical buckets in real-vehicle tests and, at the same time, ensures the detection speed to satisfy the deployment requirements on tiny devices. Satisfies all the requirements for deployment of tiny devices in the race car of the China University Student Driverless Formula Competition. The improved method in this paper can be applied to conebucket detection in complex scenarios, and the improved idea can be carried over to the detection of other small targets. [ABSTRACT FROM AUTHOR]

With the development of image feature matching technology, feature matching algorithms based on deep learning have achieved excellent results, but in scenarios with low texture or extreme perspective changes, the matching accuracy is still difficult to guarantee. In this paper, a superresolution reconstruction method based on a Residual-ESPCN (efficient subpixel convolutional neural network) approach is proposed based on LoFTR (local feature matching with transformers). The superresolution method is used to improve the interpolation method used in ASFF (adaptive spatial feature fusion) to increase the image resolution, enhance the detailed information of the image, and make the extracted features richer. Then, ASFF is introduced into the local feature extraction module of LoFTR, which can alleviate the inconsistency problem of information transmission between different scale features of the feature pyramid and lessen the amount of information lost during transmission from low- to high-resolution levels. Moreover, to improve the adaptability of the algorithm to different scenarios, OTSU is introduced to adaptively calculate the threshold of feature matching. The experimental results show that in different indoor or outdoor scenarios, our proposed algorithm for matching features can effectively improve the adaptability of feature matching and can achieve good results in terms of the area under the curve (AUC), accuracy and recall. [ABSTRACT FROM AUTHOR]

Addressing the challenges in detecting surface floating litter in artificial lakes, including complex environments, uneven illumination, and susceptibility to noise and weather, this paper proposes an efficient and lightweight Ghost-YOLO (You Only Look Once) v8 algorithm. The algorithm integrates advanced attention mechanisms and a small-target detection head to significantly enhance detection performance and efficiency. Firstly, an SE (Squeeze-and-Excitation) mechanism is incorporated into the backbone network to fortify the extraction of resilient features and precise target localization. This mechanism models feature channel dependencies, enabling adaptive adjustment of channel importance, thereby improving recognition of floating litter targets. Secondly, a 160 × 160 small-target detection layer is designed in the feature fusion neck to mitigate semantic information loss due to varying target scales. This design enhances the fusion of deep and shallow semantic information, improving small target feature representation and enabling better capture and identification of tiny floating litter. Thirdly, to balance performance and efficiency, the GhostConv module replaces part of the conventional convolutions in the feature fusion neck. Additionally, a novel C2fGhost (CSPDarknet53 to 2-Stage Feature Pyramid Networks Ghost) module is introduced to further reduce network parameters. Lastly, to address the challenge of occlusion, a new loss function, WIoU (Wise Intersection over Union) v3 incorporating a flexible and non-monotonic concentration approach, is adopted to improve detection rates for surface floating litter. The outcomes of the experiments demonstrate that the Ghost-YOLO v8 model proposed in this paper performs well in the dataset Marine, significantly enhances precision and recall by 3.3 and 7.6 percentage points, respectively, in contrast with the base model, mAP@0.5 and mAP@0.5:0.95 improve by 5.3 and 4.4 percentage points and reduces the computational volume by 1.88 MB, the FPS value hardly decreases, and the efficient real-time identification of floating debris on the water's surface can be achieved cost-effectively. [ABSTRACT FROM AUTHOR]

Many government institutions in developing countries such as the Kurdistan Region of Iraq (KRI) keep a variety of paper-based records that are available in printed or handwritten format. The need for technology that turns handwritten writing into digital text is therefore highly demanded. E-government in developed and developing countries is a crucial facilitator for the provision of such services. This paper aims to develop a deep learning model based on the mask region convolutional neural network (mask-RCNN) to effectively digitize kurdish handwritten text recognition (KHTR). In this research, typical datasets, which includes the isolated handwritten Central Kurdish character images, an extensive database of 40,410 images, and 390 native writers have been produced to determine the developed approach's performance in terms of identification rates. This approach achieves adequate outcomes in terms of training time and accuracy. The proposed model gives higher performance for detection, localization, and recognition when using a dataset containing many challenges, the results were 80%, 96%, and 87.6 for precision, recall, and F-score respectively. The findings revealed that the proposed model obtained better results compared to other similar works. The accuracy of optical character recognition (OCR) is more than 99%. [ABSTRACT FROM AUTHOR]

In order to address the challenge of the small-scale, small-target, and complex scenes often encountered in offshore remote sensing image datasets, this paper employs an interpolation method to achieve super-resolution-assisted target detection. This approach aligns with the logic of popular GANs and generative diffusion networks in terms of super-resolution but is more lightweight. Additionally, the image count is expanded fivefold by supplementing the dataset with DOTA and data augmentation techniques. Framework-wise, based on the Faster R-CNN model, the combination of a residual backbone network and pyramid balancing structure enables our model to adapt to the characteristics of small-target scenarios. Moreover, the attention mechanism, random anchor re-selection strategy, and the strategy of replacing quantization operations with bilinear interpolation further enhance the model's detection capability at a low cost. Ablation experiments and comparative experiments show that, with a simple backbone, the algorithm in this paper achieves a mAP of 71.2% on the dataset, an improvement in accuracy of about 10% compared to the Faster R-CNN algorithm. [ABSTRACT FROM AUTHOR]

Timely discovery and disposal of road risk sources constitute the cornerstone of road operation safety. Presently, the detection of road risk sources frequently relies on manual inspections via inspection vehicles, a process that is both inefficient and time-consuming. To tackle this challenge, this paper introduces a novel automated approach for detecting road risk sources, termed the multi-scale lightweight network (MSLN). This method primarily focuses on identifying road surfaces, potholes, and scattered objects. To mitigate the influence of real-world factors such as noise and uneven brightness on test results, pavement images were carefully collected. Initially, the collected images underwent grayscale processing. Subsequently, the median filtering algorithm was employed to filter out noise interference. Furthermore, adaptive histogram equalization techniques were utilized to enhance the visibility of cracks and the road background. Following these preprocessing steps, the MSLN model was deployed for the detection of road risk sources. Addressing the challenges associated with two-stage network models, such as prolonged training and testing times, as well as deployment difficulties, this study adopted the lightweight feature extraction network MobileNetV2. Additionally, transfer learning was incorporated to elevate the model's training efficiency. Moreover, this paper established a mapping relationship model that transitions from the world coordinate system to the pixel coordinate system. This model enables the calculation of risk source dimensions based on detection outcomes. Experimental results reveal that the MSLN model exhibits a notably faster convergence rate. This enhanced convergence not only boosts training speed but also elevates the precision of risk source detection. Furthermore, the proposed mapping relationship coordinate transformation model proves highly effective in determining the scale of risk sources. [ABSTRACT FROM AUTHOR]