PulseExploreJournal ClubDebatesTrendingResearchersJournals
Instagram
HomeExploreJournal ClubTrending
Synapse
⌘+K
Synapse
April 1, 2026Journal of Geophysical Research Machine Learning and ComputationOpen Access

CNXT‐Ti‐LT–Based Multi‐Scale Feature–Aware Susceptibility Mapping of Rainfall‐Induced Clustered Landslides in Southeast China

View Full Paper
Ask AI
Bookmark
Share

Authors

SLSenlin LuoWMWuwei MaoZYZhiqiang Yang

Discussion

Loading...

Member takes

Overview

A model enhances accuracy for landslide mapping, suggesting improved understanding of rainfall-triggered events.

Key Points

  • This research aims to develop a deep learning model for accurately mapping landslide susceptibility caused by rainfall in Southeast China.
  • Developed a dataset with 13 controlling factors including terrain, geomorphology, and hydrology.
  • Enhanced ConvNeXt-Tiny with feature pyramids and skip connections for better multi-scale feature recognition.
  • Introduced a Lite-Transformer model for capturing global relationships among factors and improved semantic understanding.
  • Evaluated the CNXT-Ti-LT model against existing deep and machine learning models.
  • CNXT-Ti-LT showed improvements across nearly all evaluation metrics compared to other models.
  • The model effectively captured features associated with highly susceptible slopes.
  • The balance between accuracy and robustness was maintained, indicating its practical applicability.

Cite This Study

Luo et al. (2026) studied this question.

synapsesocial.com/papers/69cd7e935652765b073a9856https://doi.org/10.1029/2025jh001115
View Full Paper
Ask AI
Bookmark
Share