algorithms for sound detection and localization are discussed. These algorithms are good fit for real-world scenarios as less assumptions are used, with no big extra add-up in complexity. The algorithms are presented ...
详细信息
ISBN:
(数字)9781624105784
ISBN:
(纸本)9781624105784
algorithms for sound detection and localization are discussed. These algorithms are good fit for real-world scenarios as less assumptions are used, with no big extra add-up in complexity. The algorithms are presented in a way that can be easily programmed and implemented. The event detection algorithm is based on a probability approach that reduces the effect of the noisy features. The localization algorithm is a combination of the delay-and-sum (DAS) algorithm with the cross-correlation (CC) in a way that ensures high robustness. These algorithms are implemented on a microphone phased-array system attached to a quadcopter that is able to detect and localize sound events successfully even with the existence of high propellers noise.
In this paper, a framework will be proposed for the evaluation of detector algorithms in a maritime environment. Performance metrics and test cases will be defined to allow the impartial comparison of different detect...
详细信息
ISBN:
(纸本)9781424458110
In this paper, a framework will be proposed for the evaluation of detector algorithms in a maritime environment. Performance metrics and test cases will be defined to allow the impartial comparison of different detectors. In this framework the main approaches for detector comparison are numerical simulation and the use of recorded sea clutter and boat reflectivity data. Available data suitable to the fair comparison of different algorithms will be highlighted, with results for a selection of algorithms. The proposed framework, performance metrics and baseline cases give researchers and system engineers the ability to quantify system performance in a complex clutter environment and to evaluate the effectiveness of a particular detector (or radar design(s)) as compared to another.
We developed an automated approach for QRS complex detection and QRS duration (QRSd) measurement that can effectively analyze multichannel electrocardiograms (MECGs) acquired during abnormal conduction and pacing in h...
详细信息
We developed an automated approach for QRS complex detection and QRS duration (QRSd) measurement that can effectively analyze multichannel electrocardiograms (MECGs) acquired during abnormal conduction and pacing in heart failure and cardiac resynchronization therapy (CRT) patients to enable the use of MECGs to characterize cardiac activation in such patients. The algorithms use MECGs acquired with a custom 53-electrode investigational body surface mapping system and were validated using previously collected data from 58 CRT patients. An expert cohort analyzed the same data to determine algorithm accuracy and error. The algorithms: 1) detect QRS complexes;2) identify complexes of the most prevalent morphology and morphologic outliers;and 3) determine the array-specific (i.e., anterior and posterior) and global QRS complex onsets, offsets, and durations for the detected complexes. The QRS complex detection algorithm had a positive predictivity and sensitivity of >= 96% for complex detection and classification. The absolute QRSd error was 17 +/- 14 ms, or 12%, for array-specific QRSd and 12 +/- 10 ms, or 8%, for global QRSd. The absolute global QRSd error (12 ms) was less than the interobserver variation in that measurement (15 +/- 10 ms). The sensitivity, positive predictivity, and error of the algorithms were similar to the values reported for current state-of-the-art algorithms designed for and limited to simpler data sets and conduction patterns and within the variation found in clinical 12-lead ECG QRSd measurement techniques. These new algorithms permit accurate, real-time analysis of QRS complex features in MECGs in patients with conduction disorders and/or pacing.
Association rules are widely used to extract patterns from a given database. The association rules are capable of finding correlations among items, making it possible for the user to learn which items are present in t...
详细信息
Association rules are widely used to extract patterns from a given database. The association rules are capable of finding correlations among items, making it possible for the user to learn which items are present in the transactions and which of them have a significant correlation. One of the major problems with association rules is that the number of extracted rules usually exceeds the number of transactions present in the database, also surpassing the user's capability to explore the obtained knowledge. To overcome this problem, the post-processing phase was proposed with the objective of directing the user to the rules that potentially have the most interesting knowledge. One of the used approaches is to divide the association rules into groups (or clusters), so that rules behave similarly are on the same group, facilitating the rule set understanding. In the literature, there are some works that uses clustering algorithms to split the rules while some other works use community detection algorithms. As both approaches obtain groups of association rules, but using different premises, different results can be obtained. No study has been done on the differences among clustering and community detection algorithms, which makes the selection of the algorithm hard, once their behavior is not well known in the association rule post-processing phase. This paper presents an analysis on both approaches, aiming to find the differences and the similarities among them, making it easier to select an approach by knowing its behavior.
Here is presented a method which gives performance improvement for blob detection and tracking in real-time video stream. The algorithm can be used in multi-touch screen based and touch-free gesture based user interfa...
详细信息
Here is presented a method which gives performance improvement for blob detection and tracking in real-time video stream. The algorithm can be used in multi-touch screen based and touch-free gesture based user interfaces.
Community detection is one of the key areas of social network analysis. There are various community detection algorithms available in the literature. Numerous community metrics are also available to evaluate the detec...
详细信息
ISBN:
(纸本)9781538663745;9781538663738
Community detection is one of the key areas of social network analysis. There are various community detection algorithms available in the literature. Numerous community metrics are also available to evaluate the detected communities. In our study, by using synthetic networks, we compare between four well known community metrics, namely; modularity, conductance, coverage and performance. We also compare seven different community detection algorithms based on above mentioned parameters.
Efficient collision detection is important in many robotic tasks, from high-level motion planning in a static environment to low-level reactive behavior in dynamic situations. Specially challenging are problems in whi...
详细信息
ISBN:
(纸本)0780365763
Efficient collision detection is important in many robotic tasks, from high-level motion planning in a static environment to low-level reactive behavior in dynamic situations. Specially challenging are problems in which multiple robots are moving among multiple moving obstacles. In this paper we present a number of collision detection algorithms formulated under the kinetic data structures (KDS) framework, a framework for design and analyzing algorithms for objects in motion. The KDS framework leads to event-based algorithms that sample the state of different parts of the system only as often as necessary for the task at hand. Earlier work has demonstrated the theoretical efficiency of KDS algorithms. In this paper we present new algorithms and demonstrate their practical efficiency as well as by an implementable and direct comparison with classical broad and narrow phase collision detection techniques.
Wearing a helmet is one of the effective measures to protect workers' safety. To address the challenges of severe occlusion, multi-scale, and small target issues in helmet detection, this paper proposes a helmet d...
详细信息
Wearing a helmet is one of the effective measures to protect workers' safety. To address the challenges of severe occlusion, multi-scale, and small target issues in helmet detection, this paper proposes a helmet detection algorithm based on deformable attention transformers. The main contributions of this paper are as follows. A compact end-to-end network architecture for safety helmet detection based on transformers is proposed. It cancels the computationally intensive transformer encoder module in the existing detection transformer(DETR) and uses the transformer decoder module directly on the output of feature extraction for query decoding, which effectively improves the efficiency of helmet detection. A novel feature extraction network named Swin transformer with deformable attention module(DSwin transformer) is proposed. By sparse cross-window attention, it enhances the contextual awareness of multi-scale features extracted by Swin transformer, and keeps high computational efficiency simultaneously. The proposed method generates the query reference points and query embeddings based on the joint prediction probabilities, and selects an appropriate number of decoding feature maps and sparse sampling points for query decoding, which further enhance the inference capability and processing speed. On the benchmark safety-helmet-wearing-dataset(SHWD), the proposed method achieves the average detection accuracy mAP@0.5 of 95.4% with 133.35G floating-point operations per second(FLOPs) and 20 frames per second(FPS), the state-of-the-art method for safety helmet detection.
The integration of Internet of Things (IoT) technology with deep learning (DL) algorithms has revolutionized plant disease detection and crop management and paved the way for sustainable agricultural practices. Real-t...
详细信息
The integration of Internet of Things (IoT) technology with deep learning (DL) algorithms has revolutionized plant disease detection and crop management and paved the way for sustainable agricultural practices. Real-time information on soil moisture, plant health, and environmental conditions can be collected by deploying a network of connected devices and sensors in agricultural fields. DL algorithms, specifically convolutional neural networks (CNN), analyze this massive dataset, facilitating timely and accurate recognition of plant diseases. This early detection allows farmers to implement targeted interventions, like adjustment to irrigation or precision application of pesticides, maximizing crop yield, and minimizing resource wastage. Therefore, this article develops an automated Plant Disease detection and Crop Management using a spotted hyena optimizer with deep learning (APDDCM-SHODL) technique for Sustainable Agriculture. The APDDCM-SHODL approach aims to detect the existence of plant diseases and improve crop productivity in the IoT infrastructure. To achieve this, the APDDCM-SHODL method primarily employs the Vector Median Filter (VMF) technique. In addition, the Densely Connected Networks (DenseNet201) model is deployed for feature extraction. In addition, the SHO technique is exploited for optimum hyperparameter tuning of the DenseNet201 model. Furthermore, the classification algorithm is implemented by using the recurrent spiking neural network (RSNN) model. A brief set of experiments has been made to determine the experimental validation of the APDDCM-SHODL model. The comprehensive results inferred that the APDDCM-SHODL method reaches remarkable performance over other existing methods with the highest accuracy of 98.60%.
Existing frameworks of privacy-preserving multilayer community detection have room for improving detection performance and reducing communication overhead. To address these issues, we propose a novel privacy-preservin...
详细信息
Existing frameworks of privacy-preserving multilayer community detection have room for improving detection performance and reducing communication overhead. To address these issues, we propose a novel privacy-preserving multilayer community detection framework based on federated learning which is called federated multilayer community detection (FMCD). First, we propose a novel aggregation strategy by utilizing the network average degree of local networks to aggregate the parameters uploaded by clients in the step of aggregation, which can improve the performance of community detection. Second, we design a training procedure to complete multilayer community detection in multiorganizations, which can reduce communication overhead by transmitting merged community information instead of the global parameter. Finally, experiment results on synthetic and real networks with different criteria illustrate that FMCD can achieve significant performance gains, compared with state-of-the-art algorithms.
暂无评论