Performance Comparison on SensatUrban Dataset for Semantic Segmentation

RPINet achieved a mean IoU of 66.5% on the SensatUrban dataset, outperforming existing methods. This 5.5 KB Excel file, authored by Zhe Jing and last updated in May 2026, contains performance comparison data for a novel model integrating 2D images with 3D point clouds. The dataset is licensed under CC-BY-4.0 and hosted on figshare.

Use Cases

• Benchmarking new semantic segmentation models against the reported 66.5% mean IoU.
• Analyzing the performance impact of fusing 2D remote sensing images with 3D point cloud data.
• Studying the effectiveness of hybrid attention-based fusion modules for multimodal urban data.
• Evaluating adaptive sampling strategies and multi-objective loss functions for geometric consistency.

Strengths

• Includes a specific, state-of-the-art performance metric (66.5% mean IoU) for benchmarking.
• Data is associated with a published method (RPINet) that demonstrated strong generalization on unseen datasets.
• Licensed under the permissive CC-BY-4.0 license for open use.

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• The 5.5 KB size suggests the dataset likely contains summary results, not the raw training or validation data.

Provenance

Source

figshare

Collection Method

Likely contains experimental results from the RPINet paper.

Time Range

Dataset last updated in 2026; temporal coverage of the underlying SensatUrban data is unknown.

Freshness

Last updated 2026-05-22 17:38:40; freshness should be verified.

Geography

Based on the SensatUrban dataset, likely covers an urban area; specific geography is unknown.