SwePub
Sök i SwePub databas

  Utökad sökning

Träfflista för sökning "WFRF:(Wang N) ;lar1:(hig)"

Sökning: WFRF:(Wang N) > Högskolan i Gävle

  • Resultat 1-2 av 2
Sortera/gruppera träfflistan
   
NumreringReferensOmslagsbildHitta
1.
  • Wang, Yan Ming, et al. (författare)
  • Colorado geoid computation experiment : overview and summary
  • 2021
  • Ingår i: Journal of Geodesy. - : Springer. - 0949-7714 .- 1432-1394. ; 95:12
  • Tidskriftsartikel (refereegranskat)abstract
    • The primary objective of the 1-cm geoid experiment in Colorado (USA) is to compare the numerous geoid computation methods used by different groups around the world. This is intended to lay the foundations for tuning computation methods to achieve the sought after 1-cm accuracy, and also evaluate how this accuracy may be robustly assessed. In this experiment, (quasi)geoid models were computed using the same input data provided by the US National Geodetic Survey (NGS), but using different methodologies. The rugged mountainous study area (730 km x 560 km) in Colorado was chosen so as to accentuate any differences between the methodologies, and to take advantage of newly collected GPS/leveling data of the Geoid Slope Validation Survey 2017 (GSVS17) which are now available to be used as an accurate and independent test dataset. Fourteen groups from fourteen countries submitted a gravimetric geoid and a quasigeoid model in a 1' x 1' grid for the study area, as well as geoid heights, height anomalies, and geopotential values at the 223 GSVS17 marks. This paper concentrates on the quasigeoid model comparison and evaluation, while the geopotential value investigations are presented as a separate paper (Sanchez et al. in J Geodesy 95(3):1. https://doi.org/10.1007/s00190-021-01481-0, 2021). Three comparisons are performed: the area comparison to show the model precision, the comparison with the GSVS17 data to estimate the relative accuracy of the models, and the differential quasigeoid (slope) comparison with GSVS17 to assess the relative accuracy of the height anomalies at different baseline lengths. The results show that the precision of the 1' x 1' models over the complete area is about 2 cm, while the accuracy estimates along the GSVS17 profile range from 1.2 cm to 3.4 cm. Considering that the GSVS17 does not pass the roughest terrain, we estimate that the quasigeoid can be computed with an accuracy of similar to 2 cm in Colorado. The slope comparisons show that RMS values of the differences vary from 2 to 8 cm in all baseline lengths. Although the 2-cm precision and 2-cm relative accuracy have been estimated in such a rugged region, the experiment has not reached the 1-cm accuracy goal. At this point, the different accuracy estimates are not a proof of the superiority of one methodology over another because the model precision and accuracy of the GSVS17-derived height anomalies are at a similar level. It appears that the differences are not primarily caused by differences in theory, but that they originate mostly from numerical computations and/or data processing techniques. Consequently, recommendations to improve the model precision toward the 1-cm accuracy are also given in this paper.
  •  
2.
  • Zhou, Jincheng, et al. (författare)
  • Voice spoofing countermeasure for voice replay attacks using deep learning
  • 2022
  • Ingår i: Journal of Cloud Computing. - : Springer. - 2192-113X. ; 11:1
  • Tidskriftsartikel (refereegranskat)abstract
    • In our everyday lives, we communicate with each other using several means and channels of communication, as communication is crucial in the lives of humans. Listening and speaking are the primary forms of communication. For listening and speaking, the human voice is indispensable. Voice communication is the simplest type of communication. The Automatic Speaker Verification (ASV) system verifies users with their voices. These systems are susceptible to voice spoofing attacks - logical and physical access attacks. Recently, there has been a notable development in the detection of these attacks. Attackers use enhanced gadgets to record users’ voices, replay them for the ASV system, and be granted access for harmful purposes. In this work, we propose a secure voice spoofing countermeasure to detect voice replay attacks. We enhanced the ASV system security by building a spoofing countermeasure dependent on the decomposed signals that consist of prominent information. We used two main features— the Gammatone Cepstral Coefficients and Mel-Frequency Cepstral Coefficients— for the audio representation. For the classification of the features, we used Bi-directional Long-Short Term Memory Network in the cloud, a deep learning classifier. We investigated numerous audio features and examined each feature’s capability to obtain the most vital details from the audio for it to be labelled genuine or a spoof speech. Furthermore, we use various machine learning algorithms to illustrate the superiority of our system compared to the traditional classifiers. The results of the experiments were classified according to the parameters of accuracy, precision rate, recall, F1-score, and Equal Error Rate (EER). The results were 97%, 100%, 90.19% and 94.84%, and 2.95%, respectively.
  •  
Skapa referenser, mejla, bekava och länka
  • Resultat 1-2 av 2

Kungliga biblioteket hanterar dina personuppgifter i enlighet med EU:s dataskyddsförordning (2018), GDPR. Läs mer om hur det funkar här.
Så här hanterar KB dina uppgifter vid användning av denna tjänst.

 
pil uppåt Stäng

Kopiera och spara länken för att återkomma till aktuell vy