Search

Traditional point cloud simplification methods are slow to process large point clouds and prone to losing small features, which leads to a large loss of point cloud accuracy. In this paper, a new point cloud simplification method using a three-step strategy is proposed, which realizes efficient reduction of large point clouds while preserving fine features through point cloud down-sampling, normal vector calibration, and feature extraction based on the proposed feature descriptors and neighborhood subdivision strategy. In this paper, we validate the method using measured point clouds of large co-bottomed component surfaces, visualize the errors, and compare it with other methods. The results demonstrate that this method is well-suited for efficiently reducing large point clouds, even those on the order of ten million points, while maintaining high accuracy in feature retention, refinement precision, efficiency, and robustness to noise. [ABSTRACT FROM AUTHOR]

Forest fire is natural disasters that are sudden, destructive and difficult to handle and rescue, with millions of hectares of forests burned every year all over the world, causing serious ecological damage, loss of life and property. Therefore, timely detection and treatment of early fire is of positive and important significance for forest fire early control. The fire detection method based on image processing is one of the most important means of preventing the occurrence of large-scale forest fires at present by extracting the flame and smoke features in the image and quickly determining the location of the fire. The current deep learning-based forest fire early detection methods have problems such as high false alarm rate due to small detection targets and complex environmental background, and large number of detection model parameters. In view of this, this paper proposes a lightweight early forest fire and smoke detection method based on GS-YOLOv5. Firstly, this paper proposes a novel Super-SPPF structure to replace the SPPF structure in YOLOv5, by which the output of the feature extraction network is used as input, and the input is divided into two branches to retain more semantic information, and the GhostConv operation is performed separately to reduce the number of model parameters. The Super-SPPF structure performs the serial MaxPooling operation on one of the branches, which improves the computation speed by choosing a smaller pooling kernel, and then fuses the outputs of the two branches to reduce the false alarm rate of the detection model. Secondly, C3Ghost is utilized instead of the C3 module in YOLOv5 to further reduce the number of detection model parameters. Finally, the coordinate attention (CA) module is introduced in backbone of YOLOv5 to obtain the relationship between channels and space, which enables the network to obtain the location information of interest more accurately and further improves the detection accuracy of early fires. In this paper, a self-constructed DL-Fire dataset is used to verify the performance of the GS-YOLOv5 detection model by collecting environmental interference samples and combining them with the D-Fire dataset. The experimental results show that the detection accuracy of GS-YOLOv5 is 95.9%, and the model size is 10.58 mb. Compared with YOLOv5, false alarm rate is reduced from 12% to 6% and computational complexity is reduced from 16.0 GFlops to 12.8 GFlops by GS-YOLOv5. [ABSTRACT FROM AUTHOR]

Existing biohashing authentication schemes usually store bioinformation in the cloud, whose storage method can face the problems of bioinformation leakage and tampering, and single point of failure. To address the above problems, a ciphertext speech biohashing authentication scheme based on chaotic system and improved public chain is proposed. First, the fusion features are extracted and generated with the inner product of 3D chaotic system to generate the biohashing sequence. Second, the original speech is encrypted and uploaded to the InterPlanetary file system and then, uploaded to the blockchain after the data are encrypted. Finally, the server matches the hash sequence queried on the chain with the to-be-authenticated hash at the user's end using the Hamming distance. The experimental results show that the biohashing algorithm in this paper has better distinguishability, robustness, security and time efficiency. The key space of the speech encryption algorithm is 2 256 , which can well resist the exhaustive attack and ensure the security during speech authentication. In this paper, the combination of on-chain and off-chain storage of speech information can prevent the problem of bioinformation leakage and tampering, as well as avoiding the problem of single point of failure. [ABSTRACT FROM AUTHOR]

Lung cancer is an uncontrolled growth of tissue causing a lump in the human lung. If lung cancer can be detected early, it can increase the survival rate. Therefore, a multi-classification approach of lung nodule detection with high computational effectiveness is required. In this paper, a multi-classification approach of lung nodule detection and classification is proposed using artificial intelligence on computed tomography (CT) scan images. Different preprocessing steps are applied for resizing, smoothing, and enhancement of the CT images. Then, two different approaches for feature extraction using VGG16 transfer learning and morphological segmentation are proposed. Morphological segmentation and feature extraction are applied for the segmentation of the region of interest and to extract the distinct features. Finally, the proposed deep learning architecture and seven different machine learning algorithms are applied on the preprocessed data and the extracted features for the classification of lung nodules into three classes: malignant, benign, and normal. It is observed that the stacked ensemble model of deep learning convolutional neural network (CNN) and VGG16 transfer learning models (CNN+VGG16) can achieve 99.55% accuracy using preprocessed data. It is also observed that all the ML algorithms perform with reasonably high accuracy using the low-dimensional morphological features. It is observed from the fivefold cross-validation results that logistic regression performs with 99.36% accuracy in 23.71 s time using the preprocessed data. Whereas, using the morphological features, k-nearest neighbor, and the support vector machine perform with the highest accuracy of 99.76% with very reduced computational time of 0.017 and 0.008 s, respectively. [ABSTRACT FROM AUTHOR]

Danmaku-enabled videos have revolutionized the way online users engage with and enjoy video content, and creating a dynamic and interactive atmosphere for users. In this paper, we investigate a new task of extract entity and relation on video-sync comments, which holds the potential to facilitate the creation of semantic-driven video knowledge graphs and video content analysis. However, the task is difficult since video-sync comments often contain intricate, ambiguous relationships, and their syntactic structure may be incomplete or inconsistent. One possible method is to extract relationships by using distant supervision or an external knowledge base system (Kbs). Regrettably, the domain of online videos suffers from a lack of comprehensive external KBs. To tackle the problem, we propose an innovative strategy that combines syntactic rule-based techniques with advanced pre-trained language models. We aim to encapsulate both syntactic cues and distinctive features within latent spaces. Our approach comprises two key facets. Firstly, we employ syntactic rules to extract open relations from the video-sync comments. Simultaneously, we introduce a BERT-based model that integrates relationship category constraints. This model not only fuses extraction features but also autonomously learns intricate relationships, enhancing the accuracy and granularity of the extraction process. For demonstration, we conduct experiments on large-scale danmaku data and results demonstrate that our proposed approach can achieve competitive performance. [ABSTRACT FROM AUTHOR]

The unsupervised anomaly detection in high-dimensional and complex settings poses a formidable challenge. To tackle the challenges associated with the recognition of high-dimensional data, this paper proposes a feedback channel between the Variational Auto Encoder and the Gaussian process to enhance its data feature extraction capabilities. In order to alleviate the impact of common mislabeling issues within unsupervised contexts, the proposed model introduces a method for the selection of representative data. Subsequently, these representative data points are employed as anchors for the construction of a full-data Gaussian process model covering the entire dataset. To further disentangle the intertwined normal and abnormal data, the proposed model employs a Bayesian framework to guide the displacement of each data point within the latent space toward its position with the highest confidence. This strategic approach effectively discriminates between normal and abnormal data. This innovative method significantly enhances data consolidation within the latent space, thus preserving robustness, particularly when processing high-complexity data. Also, the numerical experiments not only validate the method's effectiveness but also showcase the model's robustness. These experiments highlight that even a small number of incorrect labels do not significantly impact the accuracy of anomaly detection. [ABSTRACT FROM AUTHOR]

With the rapid development of artificial intelligence and Internet of Things (IoT) technology, increasingly edge devices have entered people's daily lives. However, due to the limited performance of edge devices, complex models can affect the response speed and efficiency of the whole system. Existing research still cannot simultaneously satisfy the demand for accuracy and response speed of edge devices. This paper proposes a lightweight and highly accurate object detection model that uses the Transformer to address edge devices' limited computational capacity and storage space. Specifically, the proposed model adopts the Swin Transformer for multi-scale feature extraction to achieve better global modeling capability. In addition, we propose the Neck module with path aggregation network (PAN), which is designed with a two-feature pyramid structure capable of combining semantic and localization information in order to improve the operational performance by exploiting the underlying location features. A lightweight detection head is then developed using group convolution, fusing the two localization branches and removes the additional decoupling operation. Finally, we conduct comparative experiments on three datasets: the Retail-cabinet dataset, the Roadsign dataset, and the Pascal VOC dataset. Experimental results show that compared with the baseline model, our model achieves an 11.8% improvement in mAP on the Retail-cabinet dataset while reducing Params and FLOPs by 23.19% and 71.50%, respectively. The proposed model effectively reduces the model's computational complexity and improves detection performance, thereby possessing high practical value. This code is released on https://github.com/ydlam/ST-YOLOX. [ABSTRACT FROM AUTHOR]

Microelectronics production failure analysis is an important step in improving product quality and development. In fact, the understanding of the failure mechanisms and therefore the implementation of corrective actions on the cause of the failure depend on the results of this analysis. These analyses are saved under textual features format. Then such data need first to be preprocessed and vectorized (converted to numeric). Second, to overcome the curse of dimensionality caused by the vectorisation process, a dimension reduction is applied. A two-stage variable selection and feature extraction is used to reduce the high dimensionality of a feature space. We are first interested in studying the potential of using an unsupervised variable selection technique, the genetic algorithm, to identify the variables that best demonstrate discrimination in the separation and compactness of groups of textual data. The genetic algorithm uses a combination of the K-means or Gaussian Mixture Model clustering and validity indices as a fitness function for optimization. Such a function improves both compactness and class separation. The second work looks into the feasibility of applying a feature extraction technique. The adopted methodology is a Deep learning algorithm based on variational autoencoder (VAE) for latent space disentanglement and Gaussian Mixture Model for clustering of the latent space for cluster identification. The last objective of this paper is to propose a new methodology based on the combination between variational autoencoder (VAE) for the latent space disentanglement, and genetic algorithm (GA) to find, in an unsupervised way, the latent variables allowing the best discrimination of clusters of failure analysis data. This methodology is called VAE-GA. Experiments on textual datasets of failure analysis demonstrate the effectiveness of the VAE-GA proposed method which allows better discrimination of textual classes compared to the use of GA or VAE separately or the combination of PCA with GA (PCA-GA) or a simple Auto-encoders with GA (AE-GA). [ABSTRACT FROM AUTHOR]

The presence of haze results in diminished image clarity, the loss of image details, color alterations, and a decline in image visibility. This paper proposes an efficient multi-stage feature fusion defogging network based on the attention mechanism for effective defogging in complex environments. To create a multi-branch defogging network, the model integrates several attention techniques to create various sorts of branching network architectures. The image's spatial details and contextual information are supplemented and merged based on the recovered image feature information of different network branches to increase the effectiveness of the network model. The stage attention fusion mechanism created among several network branches can lessen the loss of image data during feature extraction and improve the effectiveness of the image-defogging operation. The experimental results demonstrate that the proposed algorithm has superior defogging performance in both synthetic and real-world scene datasets and performs more admirably in terms of accuracy compared to other sophisticated algorithms, particularly in the RESIDE and O-Haze datasets. The PSNR metrics on the RESIDE and O-Haze datasets are improved by 1.58 dB and 1.61 dB, respectively, compared to the best-advanced technique suggested in this study, and the SSIM metrics on the O-Haze dataset are improved by 3.4%. [ABSTRACT FROM AUTHOR]

Facial expression recognition (FER) is a technology that recognizes human emotions based on biometric markers. Over the past decade, FER has been a popular research area, particularly in the computer vision community. With deep learning (DL) development, FER can achieve impressive recognition accuracy. In addition, favorable Internet-of-Things (IoT) advancements generate massive amounts of visual data needed to enable reliable DL-based emotion analysis. However, training DL models can suffer from significant memory consumption and computational costs, complicating many vision tasks. Additionally, the direct use of RGB images during the training and inference stages might raise privacy concerns in various FER applications. On the other hand, adopting large deep networks hampers quick and accurate recognition on resource-constrained end devices such as smartphones. As a viable solution, edge computing can be employed to bring data storage and computation closer to end devices rather than relying on a central cloud server. As a result, it can potentially facilitate the deployment of real-time FER applications since the latency and efficiency problems are well addressed by utilizing the computing resources at the edge. In this paper, we develop an efficient FER framework that integrates DL with edge computing. Our framework relies on facial landmarks to enable privacy-preserving and low-latency FER. Accordingly, various landmark detection models and feature types are studied empirically to investigate their capabilities in capturing the dynamic information of facial expressions in videos. Then, using the extracted landmark-based features, we design lightweight DL models to classify human emotions on IoT devices. Extensive experiments performed on benchmark datasets further validate the efficiency and robustness of our framework. [ABSTRACT FROM AUTHOR]

Fast and accurate detection of traffic signs is crucial for the development of intelligent transportation systems. To address the issue of false detection and missing detection of small traffic signs in complex scenes, this paper proposes a YOLO-SG model based on YOLOv5. The YOLO-SG approach employs SPD-Conv as a down-sampling structure to mitigate the loss of feature information during the down-sampling process. This enhances the detection performance of small objects in complex scenes and improves the generalization and robustness of the model. The feature extraction architecture uses GhostNet, which effectively reduces the number of model parameters and weight, enhancing the feasibility of practical model deployment. Furthermore, this study optimizes the output feature structure by introducing a small object detection layer and removing the large object detection layer, enabling the detection of small objects. Extensive experiments conducted on the GTSDB and TT100K datasets demonstrate that YOLO-SG exhibits excellent detection performance. On the GTSDB dataset, YOLO-SG achieved a 2.3% increase in mAP compared to the baseline network, while reducing the number of parameters by 42%. Similarly, on the TT100K dataset, YOLO-SG increased mAP by 6.3% and reduced the number of parameters by 43%. These experimental results showcase the effectiveness and accuracy of YOLO-SG, particularly in detecting small traffic signs. [ABSTRACT FROM AUTHOR]