CITOBOT AI

The CITOBOT AI dataset contains anonymized cervical colposcopy images and associated screening labels collected in a real-world clinical setting as part of a public hospital–based cervical cancer screening program in Colombia. The dataset was created to support the development, validation, and evaluation of artificial intelligence models for cervical cancer risk stratification using image-based approaches. Images were obtained during routine colposcopy examinations following standard clinical protocols and curated to ensure sufficient quality for automated analysis. Screening labels were defined using colposcopy-guided biopsy as the reference standard and categorized for screening purposes. The dataset reflects the variability of routine clinical practice and is intended to facilitate research on scalable digital health solutions for cervical cancer screening, particularly in resource-limited settings.