scAnt

scAnt generates automated detailed, textured 3D models of arthropods and small objects to enable quantitative morphometrics, synthetic data generation for machine learning, and specimen digitization.

Key Features:

• Automated Image Generation: scAnt automates the generation of Extended Depth Of Field (EDOF) images from multiple perspectives using an integrated scanner.
• Advanced Image Processing: The platform's automatic image masking routine combines random forest-based edge-detection, adaptive thresholding, and connected component labeling for precise segmentation.
• Photogrammetry Integration: Processed images can be exported to photogrammetry software, including the open-source Meshroom, for 3D reconstruction.
• Model Rigging and Realism: scAnt supports rigging for realistic specimen posing and implements a method for semi-realistic representation of planar and transparent structures such as wings.

Scientific Applications:

• Machine Learning-Driven Behavioral Studies: Synthetic data generated from 3D models supports development and testing of machine learning methods for arthropod behavior analysis.
• Comparative Morphometric Studies: Digital models enable precise area and volume measurements to expand morphometric parameter spaces.
• Outreach and Education: Digital 3D specimens produced by scAnt can be used for scientific outreach and public engagement with natural history collections.
• Digitization Efforts: The platform aids digitization initiatives by producing digital archives of specimens for natural history collections.

Methodology:

Automated generation of Extended Depth Of Field (EDOF) images; automatic image masking using random forest-based edge-detection, adaptive thresholding, and connected component labeling; export of processed images for photogrammetry (e.g., Meshroom); model rigging and methods for representing planar and transparent structures.