Ablation Experiments for Prostate MRI Segmentation on PROMISE12 Dataset

Ablation experiments evaluating a Lightweight Mamba-UNet (LM-UNet) method for prostate lesion segmentation in MRI. The dataset, uploaded by Kuncai Xu in March 2026, compares model performance on parameters like parameter count, computational memory, and segmentation accuracy against seven other methods. The data is stored in an XLS file sized 9.5 KB.

Use Cases

• Benchmarking segmentation model performance based on reported metrics like parameter count and computational memory.
• Analyzing the impact of architectural components like Parallel Vision Mamba (PV-Mamba) and Efficient Multi-scale Attention (EMA) on segmentation accuracy.
• Studying feature fusion techniques for medical imaging based on described Edge Feature Extraction (EFE) and Edge Feature Fusion (EFF) modules.
• Investigating methods to reduce semantic discrepancies in encoder-decoder networks based on the described Multi-stage and Multi-level Skip Connections (MMSC).

Strengths

• Results are benchmarked against seven comparative segmentation methods, providing a relative performance context.
• The dataset is licensed under CC-BY-4.0, permitting broad reuse with attribution.
• File size is 9.5 KB, indicating a compact and likely fast-to-download resource.

Limitations

• Row count and specific column definitions are unknown, limiting suitability assessment.
• The dataset is very small (9.5 KB), suggesting it contains summary results rather than raw image data or extensive records.
• Column-level documentation is absent; field semantics must be inferred after download.

Provenance

Source

figshare, author Kuncai Xu.

Collection Method

Experimental results from testing the LM-UNet segmentation method.

Time Range

Experiments likely conducted prior to March 2026; specific temporal coverage of the underlying PROMISE12 data is not stated.

Freshness

Last updated 2026-03-23 17:33:26.

Geography

null