RegBenchmark

RegBenchmark: Quantitative Benchmarking Framework for 3D Histology Reconstruction Algorithms

RegBenchmark evaluates and quantitatively compares tissue reconstruction algorithms for three-dimensional histology using whole slide image datasets to assess reconstruction accuracy.

Key Features:

• Algorithm Benchmarking: Performs quantitative accuracy assessment of free and commercial 3D histology reconstruction algorithms.
• Whole Slide Image Evaluation: Uses two whole slide image datasets to measure reconstruction performance.
• Deformation Sensitivity Analysis: Identifies performance differences between algorithms that compensate for local tissue deformation and simpler reconstruction approaches.
• Standardized Evaluation Framework: Provides objective, reproducible comparison metrics for 3D reconstruction methods.

Scientific Applications:

• 3D Histology Reconstruction Assessment: Evaluates algorithmic accuracy for reconstructing serial tissue sections into three-dimensional tissue models.
• Spatial Omics Integration Support: Facilitates validation of reconstruction methods used for integrating cellular morphology mapping with spatially resolved omics data in 3D tissue contexts.
• Digital Pathology Method Development: Guides optimization and selection of reconstruction algorithms for diagnostic and research applications.

Methodology:

RegBenchmark applies multiple 3D histology reconstruction algorithms to two whole slide image datasets and quantifies reconstruction accuracy. Comparative analysis emphasizes the impact of local tissue deformation compensation on algorithmic performance.