Prostate MRI Lesion Segmentation Method Performance Comparison

Experimental results compare the proposed LM-UNet method against seven other segmentation methods on the PROMISE12 dataset. The analysis includes metrics for parameter count, computational memory requirements, and segmentation accuracy of lesion margins. The dataset is provided by author Kuncai Xu and was last updated in March 2026.

Use Cases

• Compare segmentation method performance on metrics like parameter count and computational memory requirements.
• Analyze the impact of edge feature extraction (EFE) and edge feature fusion (EFF) on segmentation accuracy.
• Benchmark the LM-UNet method's lesion margin segmentation against seven other methods.

Strengths

• Direct comparison of eight distinct segmentation methods, including the novel LM-UNet.
• Includes specific performance metrics like parameter count and computational memory requirements.
• Results are benchmarked on the established PROMISE12 dataset for prostate MRI.

Limitations

• The dataset is very small at 9.5 KB, suggesting it contains summary results rather than raw image data.
• No raw image data or segmentation masks are provided, limiting use to methodological comparison.
• The temporal coverage and sample size for the underlying PROMISE12 data are not specified.

Provenance

Source

figshare, author Kuncai Xu.

Collection Method

Experimental results from comparing segmentation methods on the PROMISE12 dataset.

Freshness

Last updated March 2026.