Real and Simulated Hair Fiber Images for Cross-Sectional and Curvature Phenotyping

The Hair Phenotyping Methods Project dataset contains simulated and real images of hair fibers used in the paper 'High-throughput phenotyping methods for quantifying hair fiber morphology'. The data was collected by the Shriver Lab at Pennsylvania State University from 192 hair samples and includes cross-sectional micrographs and longitudinal curvature images. The dataset is designed for use with the fibermorph Python package.

Use Cases

• Developing image segmentation algorithms based on simulated hair cross-sections and arcs.
• Training models to measure hair curvature from longitudinal images of real fiber fragments.
• Benchmarking phenotyping tools using a dataset containing both simulated and real biological images.
• Validating morphological analysis pipelines on high-resolution micrographs of hair cross-sections.

Strengths

• Includes both simulated data (ellipses and arcs) and real data (micrographs and fragment images) for method validation.
• Real data comprises images from 192 hair samples collected with informed consent and ethical approval.
• Dataset is associated with a specific research paper and a dedicated Python package (fibermorph) for analysis.

Limitations

• Row count and total number of images are unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• Description notes that not all 192 samples have both cross-section and curvature images due to quality control.

Provenance

Source

Hair Phenotyping Methods Project, Shriver Lab at Pennsylvania State University

Collection Method

Real hair samples were embedded, sectioned, and imaged (cross-sections) or cut, washed, and imaged (curvature). Simulated data was generated via custom Python and R scripts.