Fig. 9From: Automatic classification of 3D positional relationship between mandibular third molar and inferior alveolar canal using a distance-aware networkVisualization of Grad-CAM results from the 3D distance-aware networks by SegNet, simple U-Net, Attention U-Net, and Dense121 U-Net with inputs of the CBCT image and segmentation mask (left), the SDM (middle), and the CBCT image and SDM (right). Ground truth (on CBCT image) and predicted classes are shown on each image, where B and L indicate buccal and lingual relationships, respectivelyBack to article page