PCP-GAN generates realistic pore-scale images with controlled geological properties

Researchers have developed a conditional Generative Adversarial Network (cGAN) called PCP-GAN to generate realistic pore-scale images with controlled geological properties. This framework addresses the challenge of spatial heterogeneity and data scarcity in subsurface characterization by training on thin section images and conditioning on porosity and depth. The generated images accurately reflect bulk formation properties, preserving critical mineralogical information and morphological characteristics, which is valuable for applications like carbon storage and geothermal energy. AI

IMPACT Enables more accurate subsurface characterization for energy and environmental applications by generating realistic geological images.

RANK_REASON The cluster contains a research paper detailing a new model (PCP-GAN) for image reconstruction. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• Ali Sadeghkhani
• Generative Adversarial Networks
• PCP-GAN

• paper
• model release

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →