Experimental data and analysis code for "A quantitative approach for high-throughput mapping of collagen fibril orientation using polarization-dependent Second Harmonic Generation"

This record contains the core dataset and analysis code for the study "A quantitative approach for high-throughput mapping of collagen fibril orientation using polarization-dependent Second Harmonic Generation" by Zanrè et al. DATA The data folder structure is described in the accompanying README file. Raw SHG image stacks are provided as .npy files for the following samples: MTLT_0.7deg - Transversely cut mineralized turkey leg tendon (MTLT). Used for P-SHG calibration of the d₃₃/d₃₁ ratio and for qPSHG model validation. MTLT_22.8deg - MTLT section at a known tilt angle. Used for qPSHG model validation. MTLT_43.1deg - MTLT section at a known tilt angle. Used for qPSHG model validation. MTLT_68.9deg - MTLT section at a known tilt angle. Used for qPSHG model validation. MTLT_89.5deg - MTLT section at a known tilt angle. Used for qPSHG model validation. Bone_osteon1 - Human osteonal bone. Used to demonstrate qPSHG application to retrieve the out-of-plane orientation of mineralized collagen fibrils. Bone_osteon2 - Human osteonal bone. Used to demonstrate qPSHG application to retrieve the out-of-plane orientation of mineralized collagen fibrils. CODE OVERVIEW The repository includes six Python modules. All processing is driven by two pipeline scripts: Calibration_PSHG_pipeline.pyApplied to MTLT_0.7deg to calibrate the d₃₃/d₃₁ ratio using the standard P-SHG model.Runs the full preprocessing chain (frame registration → spatial binning → thresholding) and then calls Calibration_PSHG_fitting.py for pixel-wise fitting. qPSHG_pipeline.py Applies the quantitative P-SHG model to all MTLT samples. Runs the full preprocessing chain (frame registration → spatial binning → thresholding) and then calls qPSHG_fitting.py for pixel-wise fitting. After fitting, runs an agnostic FFT-based quality control (agnostic_qc.py) and produces both raw and QC-filtered maps of θ (out-of-plane tilt), δ (in-plane orientation), and C₂. Supporting modules: Calibration_PSHG_fitting.py Implements the P-SHG model I(α) = K·(|sin²(α−δ) + d₃₃/d₃₁·cos²(α−δ)|² + |sin(2(α−δ))|²) and fits K, d₃₃/d₃₁, δ pixel-wise. qPSHG_fitting.py Implements the constrained qPSHG model I(α) = K·(C₁·cos⁴(α−δ) + C₂·cos²(α−δ) + 1) with C₁ = −0.163·C₂², fitting K, δ, C₂ pixel-wise. agnostic_qc.py Model-agnostic quality control via FFT decomposition of the polar intensity trace. Flags pixels where the harmonic content exceeds that expected from the fitted qPSHG model. image_processing.py Frame registration, spatial binning, and thresholding. visualization.py Plotting utilities: polar plots, parameter maps, histograms, direction field overlays. io_utils.py Path management and array I/O For questions or support regarding the code and data, please contact [e-mail] .