Amplitude-guided deep reinforcement learning for semi-supervised layer segmentation

Enting Gao

doi:10.1016/j.patcog.2026.113204

Amplitude-guided deep reinforcement learning for semi-supervised layer segmentation

2026-01-30

0

PRE

AI

DOI：10.1016/j.patcog.2026.113204

Original

Original request for help

Save

Abstract

En 中文

• An Amplitude-aware DRL Data Augmentation (ADDA) strategy is designed by integrating Fourier transform into a deep reinforcement learning framework. • A Phase Alignment (PHA) strategy is introduced to guide the model’s attention toward structural components in the image, thereby effectively mitigating the impact of noise and artifacts. • A Cross-Power Spectrum Correlation (CPSC) module is proposed to capture global constraints for mitigating misclassification of layer structures.

Keywords:

Amplitude-aware DRL

Phase Alignment

Cross-Power Spectrum Correlation

Semi-supervised learning

Layer segmentation

Amplitude-guided deep reinforcement learning for semi-supervised layer segmentation

Abstract

Journal

Organization