Deep learning on field photography reveals the morphometric diversity of Colombian freshwater fish

Abstract Neotropical freshwater fish are among the most morphologically diverse vertebrates; however, their study has long depended on preserved specimens, which limits our understanding of their natural body shapes due to preservation-induced distortions. Field photography provides a powerful, noninvasive alternative to capture the morphology of fish as it occurs in nature. However, automatically extracting accurate shape information from these images remains a major challenge, especially for highly diverse taxa. Here, we present an AI-based workflow that integrates Segment Anything and Grounding DINO to automate fish segmentation and shape extraction from field photographs. This approach enables a broad sampled analysis of the morphological diversity spectrum of Colombian freshwater fish. We applied this workflow to CavFish-Colombia, a curated dataset of 1749 images representing 78% of orders ( n = 11), 82% of families ( n = 48), 41% of genera ( n = 189), and 25% of species/morphospecies ( n = 428) of Colombian freshwater fish species, obtained using the PhotaFish standardized imaging system. Achieving more than 97% segmentation accuracy, our workflow enables precise and consistent extraction of natural fish body shapes. We provide the first structured morphospace of Colombian freshwater fish based on natural body shapes, quantified through descriptors such as area, perimeter, and invariant moments. This morphospace reveals distinct gradients in body size and outline related to locomotion and habitat use, spanning from large species with rounded shapes to small, elongate species. Most Colombian freshwater fishes share a predominant morphotype: small-bodied, laterally compressed. Deviations are order-specific, including Synbranchiformes that are elongate and slender; Characiformes, spanning deep-bodied and streamlined elongate types; and Siluriformes including small-bodied, streamlined, or dorsoventrally flattened armored shapes. Our results demonstrate that AI-driven field photograph analysis can allow for large-scale morphological studies, delivering accurate, rapid, and scalable data for biodiversity evaluations, functional trait analyses, and ecological research. This noninvasive morphological monitoring, directly from field images, opens new opportunities to assess fish morphology and analyze shape variation as it naturally occurs, capturing more accurate representations of living specimens.