Detection of pathological condition in distal lung images

Recently, the in vivo imaging of pulmonary alveoli was made possible thanks to confocal microscopy. For these new images, we wish to aid the clinician by developing a computer-aided diagnosis system, able to detect a pathological state in these images. An original approach that combines a texture-based characterization of the images and uses a boosted cascade of classifiers to detect a pathological condition is presented in this paper. We propose and compare two state-of-the-art texture descriptors: cooccurence matrices and local binary patterns (LBP). Recognition rates with LBP reach up to 86.3% and 95.1% for the non-smoking and smoking groups, respectively. Even though tests on extended databases are needed, these preliminary results are encouraging for this challenging task of image classification.