1-20 OF 257 RESULTS FOR

PhaseNet

Results shown limited to content with bounding coordinates.
Follow your search
Access your saved searches in your account

Would you like to receive an alert when new items match your search?
Close Modal
Sort by
Image
<span class="search-highlight">PhaseNet</span> picking accuracy and probability distribution. (a)    Δ   t  P    ...
Published: 01 November 2023
Figure 4. PhaseNet picking accuracy and probability distribution. (a)  Δ t P distribution, in which Δ t P represents the differences between the arrival times of seismic phases of P waves picked by PhaseNet and manually. Δ t P is positive when picked by PhaseNet
Image
(a,b) Example showing the automatic (PN = <span class="search-highlight">PhaseNet</span>, EQT = EQTransformer) an...
Published: 23 February 2024
Figure 4. (a,b) Example showing the automatic (PN = PhaseNet, EQT = EQTransformer) and manual P and S phase picks at station LOK03. Data are 8–25 Hz band‐pass filtered, and amplitude is scaled to the maximum trace value. (c–j) Time difference between the manual and PhaseNet P (red) and S
Image
Histograms of the differences between matched Kurtosis and <span class="search-highlight">PhaseNet</span> picks. ...
Published: 11 July 2024
Figure 6. Histograms of the differences between matched Kurtosis and PhaseNet picks. The pick time difference is plotted as Kurtosis pick time and PhaseNet pick time. The color version of this figure is available only in the electronic edition.
Image
Transfer‐learning workflow. The original <span class="search-highlight">PhaseNet</span>‐DAS model ( Zhu  et al. ,...
Published: 30 January 2025
Figure 2. Transfer‐learning workflow. The original PhaseNet‐DAS model ( Zhu et al. , 2023 ) is applied to new data from SeaFOAM to generate noisy pseudo labels, which are then filtered by the association algorithm GaMMA ( Zhu et al. , 2022 ) for training. It is optional to repeat
Journal Article
Published: 05 July 2022
Bulletin of the Seismological Society of America (2022) 112 (5): 2344–2363.
...Stephen L. Heck; Christopher J. Young; Ronald Brogan ABSTRACT Advances in deep learning in the past decade have recently been applied to various algorithms in the seismic event monitoring data processing pipeline. In this article, we apply PhaseNet ( Zhu and Beroza, 2018 )—a deep learning model...
FIGURES | View All (17)
Image
Similar to Figure  6  but shown here are the well‐constrained events from t...
Published: 23 February 2024
Figure 8. Similar to Figure  6 but shown here are the well‐constrained events from the automatic catalog compilation approaches: (a,b) Lassie and PhaseNet; (c,d) PhaseNet and GaMMA; (e,f) EQTransformer and GaMMA. The color version of this figure is available only in the electronic edition.
Image
Representative seismograms (vertical component) of a seismic event (    M  ...
Published: 12 November 2024
Figure 6. Representative seismograms (vertical component) of a seismic event ( M L  0.5). The red and blue lines mark the arrival times of P and S phases picked by human and PhaseNet, respectively. (a) Raw data and (b) data filtered within 2–20 Hz band are utilized in PhaseNet
Image
Same as Figure  5  but for the automatic catalog compilation approaches ran...
Published: 23 February 2024
Figure 7. Same as Figure  5 but for the automatic catalog compilation approaches ran on the ∼12 month continuous data set: (a,d,g,j) Lassie with PhaseNet; (b,e,h,k) PhaseNet with GaMMA; and (c,f,i,l) EQTtransformer with GaMMA. The color version of this figure is available only in the electronic
Image
Schematic workflows used for this study. (a) For event detection we used a ...
Published: 23 February 2024
Figure 2. Schematic workflows used for this study. (a) For event detection we used a subset of 11 control days and compared a manually compiled event catalog with the event detections of three automatic detectors (Lassie, PhaseNet/GaMMA, and EQTransformer (EQT) /GaMMA). (b) From the continuous
Image
Waveforms and phase picks for aftershocks detected (a) only by EQTransforme...
Published: 23 January 2024
Figure 5. Waveforms and phase picks for aftershocks detected (a) only by EQTransformer on 9 December 2022 and (b) only by PhaseNet on 26 December 2022. DPZ, DP1, and DP2 denote vertical, radial, and transverse components, respectively, on station 10. (a) The dark red and orange dashed lines
Image
STA&#x2F;LTA‐WCEDS and <span class="search-highlight">PhaseNet</span>‐WCEDS true positive counts ( y  axis) plotted ve...
Published: 05 July 2022
Figure 8. STA/LTA‐WCEDS and PhaseNet‐WCEDS true positive counts ( y axis) plotted versus false positive counts ( x axis). Note that the x and y axes have different scales. The solid blue horizontal line at the top of the figure marks the total UUEB possible (7862) for the region of study
Image
ECS event category counts by day for events in an area around the Circlevil...
Published: 05 July 2022
Figure 11. ECS event category counts by day for events in an area around the Circleville Earthquake for (a) PhaseNet‐WCEDS and (b) STA/LTA‐WCEDS. The gray bars show the number of UUEB associations per day (axis scale on the right side of the plot). The mainshock occurred on 3 January 2011
Image
The precision and recall metrics for the nine LPPN models and <span class="search-highlight">PhaseNet</span>. The...
Published: 22 June 2022
Figure 6. The precision and recall metrics for the nine LPPN models and PhaseNet. The tiny, medium, and large LPPN models are plotted as dashed, dot‐dashed, and dotted lines and the PhaseNet is plotted as solid line. (a) Precision of the P phase. (b) Precision of the S phase. (c) Recall
Image
The time residuals of the predicted and true phase arrivals. (a) The distri...
Published: 22 June 2022
Figure 5. The time residuals of the predicted and true phase arrivals. (a) The distributions for the P picks. The subfigures on the first row are for tiny LPPN models with stride S = 8, 16, 32 from left to right, and that from PhaseNet using the same dataset. The second and third rows
Image
Performance comparison for the original <span class="search-highlight">PhaseNet</span>‐DAS model and the SeaFOAM‐...
Published: 30 January 2025
Figure 3. Performance comparison for the original PhaseNet‐DAS model and the SeaFOAM‐PhaseNet‐DAS model for an M 5.1 event. The machine learning (ML) models are applied to 0.5 Hz high‐pass filtered strain‐rate data in ∼10‐s‐long moving windows. (a) Results using the original model and (b) new
Image
Event commonality score (ECS) event category counts by day for (a) <span class="search-highlight">PhaseNet</span>...
Published: 05 July 2022
Figure 9. Event commonality score (ECS) event category counts by day for (a) PhaseNet‐WCEDS and (b) STA/LTA‐WCEDS. Different marker shapes denote the different ECS categories. The black line shows the UUEB event counts. PhaseNet‐WCEDS is performing better relative to STA/LTA‐WCEDS in terms
Image
Seismicity associated with the Peace River earthquake sequence for the peri...
Published: 23 January 2024
with M L ≥ 4.0 from 30 November 2022 to 31 March 2023 are shown in gray and scaled by magnitude ( Vasyura‐Bathke et al. , 2023 ). Map and depth view of events detected by (a,c) PhaseNet and (b,d) EQTransformer, with 1,078 events detected by PhaseNet and 1,242 events detected by EQTransformer
Image
Map and cross‐sections of 429 earthquakes common to <span class="search-highlight">PhaseNet</span>, EQTransformer...
Published: 21 December 2023
Figure 6. Map and cross‐sections of 429 earthquakes common to PhaseNet, EQTransformer, and the manual catalogs. Gray circles, manual solutions; red circles, automatic hypocenters. (a) EQTransformer‐derived solutions. (b) PhaseNet’s. Stations are represented with black triangles. Blue lines show
Image
Summary of output picks from <span class="search-highlight">PhaseNet</span>. (a) Hourly picks at all investigated...
Published: 29 May 2024
Figure 2. Summary of output picks from PhaseNet. (a) Hourly picks at all investigated stations in the time interval 4 February 2023 00:00 UTC–11 February 2023 00:00 UTC and on 27 February 2023. In all subplots, magenta and green vertical dashed lines indicate the M w  7.8 and 7.6
Journal Article
Published: 23 January 2024
The Seismic Record (2024) 4 (1): 21–31.
...Figure 5. Waveforms and phase picks for aftershocks detected (a) only by EQTransformer on 9 December 2022 and (b) only by PhaseNet on 26 December 2022. DPZ, DP1, and DP2 denote vertical, radial, and transverse components, respectively, on station 10. (a) The dark red and orange dashed lines...
FIGURES | View All (5)