Synthetic Fault-Cave Seismic Dataset and MTA-UNet Code for Joint Identification

Description

This repository provides the synthetic seismic dataset and PyTorch implementation of MTA-UNet for joint identification of fault and karst cave systems. The dataset includes geology-constrained synthetic seismic volumes and multi-task binary ground-truth labels. The provided code includes the complete architecture with the fault-cave attention framework.

Institutions

• China University of Petroleum, Beijing

  Beijing, Beijing

Categories

Licence