Deep learning technique to detect craniofacial anatomical abnormalities concentrated on middle and anterior of face in patients with sleep apnea

•Deep learning-based model using craniofacial photographs facilitates the detection of OSA.•Important craniofacial features that are primarily concentrated in the middle and anterior regions of the face.•The prediction efficiency can be improved by increasing the weight of important regional features.•The different severities of OSA become more prevalent as the number of the important craniofacial regions increase.