Deep Learning on field photography reveals the morphometric diversity of Colombian Freshwater Fish

This dataset contains the workflow, code, and data products used to automate the analysis of Colombian freshwater fish through field photography and deep learning. It integrates four main modules: (1) segmentation of fish images using Grounding-DINO and Segment Anything, (2) morphometric validation through IoU, Dice coefficient, and error metrics, (3) extraction of morphometric descriptors including area, perimeter, Hu moments, and Zernike moments, and (4) analysis of morphometric diversity via PCA and morphospace visualization. The dataset provides image-derived measurements, validation results, descriptors, and scripts that support the quantification of morphometric variation across Colombian freshwater fish species. It is designed for reproducible research in ecology, ichthyology, biodiversity monitoring, and computer vision.