ME-CPT

April 30, 2026 · View on GitHub

#Repository URL Updated https://github.com/WHU-USI3DV/ME-CPT

ME-CPT

ME-CPT implements a 3D semantic change detection network based on multi-temporal point clouds. The proposed method is evaluated on four 3D change detection datasets. image The proposed network is built upon the Pointcept framework. The additional change detection code and implementation details will be provided after the article is published online.

New York City Semantic Change Detection Dataset (NYC-SCD)

We propose a more challenging multi-class 3D semantic change detection dataset. The new republic 3D semantic change detection dataset is sourced from high-quality ALS data collected in 2014 and 2017, released under the Open Geospatial Data Program by the New York City government. image The total area covered by these scenes is 22.5 km²。 image The dataset examples are as follows: image