Emphysema is a characteristic component of chronic obstructive pulmonary disease (COPD), which has been pointed out as one of the main causes of mortality for the next years. Animal models of emphysema are employed to study the evolution of this disease as well as the effect of treatments. In this context, measures such as the mean linear intercept [Formula: see text] and the equivalent diameter [Formula: see text] have been proposed to quantify the airspace enlargement associated with emphysematous lesions in histological sections. The parameter [Formula: see text], which relates the second and the third moments of the variable [Formula: see text], has recently shown to be a robust descriptor of airspace enlargement. However, the value of [Formula: see text] does not provide a direct evaluation of emphysema severity. In our research, we suggest a Bayesian approach to map [Formula: see text] onto a novel emphysema severity index (SI) reflecting the probability for a lung area to be emphysematous. Additionally, an image segmentation procedure was developed to compute the severity map of a lung section using the SI function. Severity maps corresponding to 54 lung sections from control mice, mice induced with mild emphysema and mice induced with severe emphysema were computed, revealing differences between the distribution of SI in the three groups. The proposed methodology could then assist in the quantification of emphysema severity in animal models of pulmonary disease.