PickX™ Neural Net Picker

A Neural Net First-Break Picker from XtremeGeo
Inhuman Results. Inhuman Speed.

Key Features

Easy Model Training
- Train new models directly from Flatirons or Phoenix data.
- Use XtremeGeo-provided models or build your own proprietary models.
Flexible Data Support
- Flatirons project data – no import required into Phoenix.
- Phoenix project data.
- SEG-Y data.
- Additional formats coming soon.
Performance
- Lightning-fast, even on entry-level GPU cards.
- High-quality picks—often as good as or better than human picks, especially on difficult datasets.

Built on LibTorch, the C++ foundation of PyTorch.
Current release requires a GPU.
Future releases will include CPU-only support for even broader accessibility.

⚡ PickX™ delivers the next generation of seismic first-break picking—combining speed, accuracy, and adaptability.

Errors are in milliseconds

PickX picks were not modified in any way
PickX picks are made using the Amplitude Onset approach
Comparison made using “Average Pick Error” of delay time and VNS tomography solutions
User picks were given an advantage- they were usually refined using the delay time stack approach
“User” picks are in general a combination of auto picks, manual picks, etc. and were made in Flatirons