CrossSwap: Performance Metrics for Pose, ID, FID, and Expression

Hojun Seo's dataset provides performance metrics for the CrossSwap method and baseline methods on Pose, ID, FID, and Expression tasks. The data is stored in a 5.5 KB XLS file and was last updated on May 14, 2026. It is licensed under CC-BY-4.0.

Use Cases

• Benchmarking model performance on pose estimation tasks based on the Pose metric.
• Comparing identity preservation capabilities between methods based on the ID metric.
• Evaluating the realism of generated images based on the FID metric.
• Analyzing facial expression transfer or generation quality based on the Expression metric.

Strengths

• The dataset is small (5.5 KB), facilitating quick download and inspection.
• It provides a clear performance comparison between CrossSwap and baseline methods.
• The license (CC-BY-4.0) is permissive for reuse and redistribution.

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

Hojun Seo

Freshness

Last updated 2026-05-14 17:41:42; freshness should be verified.