Many indoor fingerprinting localization methods are based on signal-domain distances with large error and low stability. An improved method using a clustering algorithm dynamic compensation was proposed. In the offline stage, fingerprint database built clustered hybrid distance an affinity propagation algorithm. Furthermore, clusters were adjusted transition regions given radius, as well updating corresponding position of cluster centroid. online lost received signal strength (RSS) in...

The purpose of audio-text retrieval is to learn a cross-modal similarity function between audio and text, enabling given audio/text find similar text/audio from candidate set. Recent models aggregate multi-modal features into single-grained representation. However, representation difficult solve the situation that an described by multiple texts different granularity levels, because association pattern text complex. Therefore, we propose adaptive aggregation strategy automatically optimal...

Recently, there has been great interest in feature-based image source identification. Previous statistical learning-based methods usually regarded the identification process as a classification problem. They assumed dependence of features and models. However, two assumptions are problematic because genuine coupling To address issues, this paper, we propose novel scheme. For feature coupling, coupled representation is adopted to analyze interaction among features. The relations their powers...

In this study, the authors present a novel individual retrieval method based on oral cavity point cloud data and correntropy‐based registration algorithm. Since three‐dimensional contains large amount of noise outliers, it may lead to decrease in accuracy, which affects accuracy rate. Therefore, introduce correntropy into rigid algorithm solve problem. Then, they filter matched then use mean squared error judge differences model data. Finally, accurate is realised. Experimental results...

Abstract This paper proposed a novel precise point set registration method based on feature fusion for three-dimensional data. Firstly, the prominent foreground with dense and continuous cluster structure, we propose an automatic extraction combining principal component analysis projection density-based clustering method. Secondly, sets containing noises, introduce correntropy measurement into to weaken their influence. Thirdly, of uneven distribution points in same set, algorithm which is...

In this paper, we propose a precise point set registration algorithm for indoor scene reconstruction. Firstly, as the problem that sensor samples unevenly, point-to-plane metric is introduced into our model, which effectively makes use of surface structure information and improves result. Secondly, to handle with noises outliers in scene, correntropy introduced. To solve variant Iterative Closest Point (ICP) algorithm, iteratively establishes point-to-point correspondence calculates rigid...

This paper proposes a new iterative closest point approach based on correntropy with feature guided. Iterative Closest Point (ICP) algorithm can deal most rigid registration problems, but for sets lots of noise and outliers, ICP cannot achieve high precision. We introduce into to handle this problem by suppressing the influence outliers. In terms contain large proportion planes or curved surface, have single structure, such as three-dimensional model upper jaw, we propose feature-guided...

In recent years, fuzzing has become a powerful tool for security researchers to uncover vulnerabilities. It is used discover software vulnerabilities by continuously generating malformed inputs trigger bugs. Directed grey-box also been widely in the verification of patch testing and vulnerability reproduction. For directed fuzzing, core problem make test cases reach target faster. Selecting seeds that are closer site be mutated first an effective method. this purpose, DGF calculates distance...

The DTMF audio dialing signal is widely used in worldwide scope because of its high speed, easy automatic detection -, identification and extended telephone service. With new theories algorithms continue to appear, digital processing technologies have opened up a lot application fields. This paper gives generation scheme on the theoretical basis D/A, A/D converting, deduces modified Goertzel algorithm based processing's DFT algorithm, finishs using filter banks, finally, realizes MATLAB...

Circuital sinusoidal steady-state analysis is the important content in such curriculums as Analysis and Signals Systems.It essential knowledge for college students majoring electric, signals procession of signals, communication, etc.Studying usage time-domain method or complex frequency domain method, this paper uses numerical computing solving differential equations algebra to explore an approach that can analyse steady response with use function MATLAB software modeling simulation function.

A MHP-RGA (Multiple Hierarchical Projection Grey Relational Analysis) method is proposed for consistency and effectiveness assessment between validation results, which focus on ADAS or autonomous driving functions under designed test case. The complex interactive systemic output trace could be projected calculated to the normalized relevance degree from multi-dimensional concerned levels such as system architecture, component constitution, functional behavior, measurement approaches. It can...

As the core of autonomous vehicle technology, perception chips are essential for data processing and environmental sensing. The efficacy reliability these strongly impacted by various functional risk factors may confront in scenarios real world. This paper provide efficient preventive measures an innovative method addressing safety failures across a variety real-world scenarios. It systematically investigates typical failure as well offers detailed explanation underlying difficulties. A...

Estimating driver gaze accurately is critical for the human-machine cooperative driving, but significant facial appearance diversions caused by background, illumination, personal characteristics, etc. pose a challenge to generalizability of estimation models. In this paper, we propose generative self-adversarial learning mechanism generalized that aims learn general features while eliminating sample-specific and preventing cross-domain feature over-fitting. Firstly, reduce information...

In view of the development high speed network and challenge to congestion control, By deeply researching route buffer packet loss rate performance for throughput N-TCP, which is based on equivalent noise model bandwidth estimation. Simulation results indicate that N-TCP can get more stable higher than TCP Reno significantly in situation with low router buffer, it has better ability RTT fairness compared HSTCP.