Multi-Scale LiDAR Segmentation Model Performance Metrics

5.5 KB of tabular data presents performance metrics for the MSCNet semantic segmentation model. The dataset, authored by Xuewen Feng and uploaded in April 2026, contains results from experiments on the SemanticKITTI, SemanticPOSS, and Pandasets.

Use Cases

• Compare model parameter counts against segmentation accuracy metrics across different datasets.
• Analyze the relationship between running time and model performance for real-time application suitability.
• Benchmark the multi-scale dilated convolution residual block's effectiveness against other projection-based methods.

Strengths

• Model performance validated on three established datasets: SemanticKITTI, SemanticPOSS, and Pandasets.
• Data is small (5.5 KB) for quick access and analysis.

Limitations

• Extremely limited scope with only 5.5 KB, likely containing summary metrics rather than raw or training data.
• Unknown row count and specific column structure limits detailed reproducibility.

Provenance

Source

Xuewen Feng via figshare.

Collection Method

Experimental results from training and evaluating the MSCNet model.

Freshness

Last updated April 2026.