Training Dataset for Body Condition Score Classification in Sows (BCS Grades C2–C4)

This dataset contains the images used to train and validate the body condition score (BCS) classification model for breeding sows, corresponding to grades C2, C3, and C4 of the 5-point scale proposed by García Flores (2022). Images were captured in the field using the body scanner module of a portable electronic device developed for the simultaneous monitoring of weight, body size, and body condition in pigs (Yorkshire and Duroc breeds), at the Pig Experimental Unit, Faculty of Animal Science, Universidad Nacional Agraria La Molina (UNALM), Lima, Perú. The dataset was used to train a ResNet-based classifier, achieving over 94% accuracy in BCS classification.