Benchmarking Image Analysis Pipelines for Organelle Quantification in Human Cells

A dataset comparing automated, interactive machine learning, and deep learning pipelines for quantifying lipid droplets and mitochondria in live human osteosarcoma cells. The data was created by Chloé Daul and last updated on 2026-05 08. It includes images from fluorescence microscopy and label-free holotomography, with standardized downstream feature extraction applied across different analysis workflows.

Use Cases

• Benchmarking segmentation algorithms based on the comparison of Fiji, Python, CellProfiler, ilastik, and pretrained deep learning models described.
• Evaluating modality-specific performance for organelle quantification based on the analysis of fluorescence microscopy and holotomography images.
• Selecting robust analysis strategies for metabolic perturbation studies based on the illustrated analysis of lipid droplets and mitochondria.
• Improving reproducibility in microscopy studies by understanding pipeline sensitivity to image-processing choices as highlighted in the description.

Strengths

• Systematic cross-pipeline comparison of multiple analysis approaches (Fiji, Python, CellProfiler, ilastik, pretrained deep learning).
• Includes images from two distinct imaging modalities: fluorescence microscopy and label-free holotomography.
• Focuses on two specific organelles with distinct morphologies: lipid droplets and mitochondria.
• Provides practical guidance for pipeline selection derived from benchmark results.

Limitations

• Row count is unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• Dataset size is 1.2 MB, indicating a relatively small-scale benchmark study.

Provenance

Source

figshare

Collection Method

Images of live human osteosarcoma cells acquired via fluorescence microscopy and holotomography, processed with standardized downstream feature extraction.

Time Range

null

Freshness

Last updated 2026-05-08 12:05:52; freshness should be verified.

Geography

null