Zhe Jing published ablation study results for RPINet on 2026-05-22. RPINet is a Remote-Projection and Intelligent Network integrating 2D remote sensing images with 3D point cloud data for semantic segmentation of urban scenes. The dataset likely contains performance metrics from experiments conducted on the SensatUrban dataset.
Use Cases
- Benchmarking semantic segmentation models based on the reported mean IoU of 66.5%
- Analyzing the impact of hybrid attention-based fusion modules on segmentation accuracy
- Studying the performance of adaptive sampling strategies in point cloud processing
- Evaluating the generalization of models on unseen urban datasets
Strengths
- Dataset is openly licensed under CC-BY-4.0
- Results are based on experiments on the challenging SensatUrban dataset
- Model achieved a reported mean IoU of 66.5%, outperforming existing methods
Limitations
- Dataset size is 5.5 KB, indicating a very limited scope likely containing only summary results
- Column-level documentation is absent; field semantics must be inferred after download
- Row count is unknown, which may limit suitability assessment
Provenance
- Source
- Zhe Jing
- Collection Method
- Likely contains experimental results from ablation studies of the RPINet model.
- Freshness
- Last updated 2026-05-22 17:38:42