Figure 7: Segment, Measure, and AutoQC the midsagittal CC (SMACC) pipeline -.

The midsagittal slice from a participant registered to MNI space with 6 degrees of freedom serves as an input to the UNet architecture used for the midsagittal corpus callosum segmentation. The Witelson atlas was used for segmenting the CC into five different regions. Global and subregion metrics (thickness and area-shown in green) were extracted from the segmentation. The thickness (black arrow) is defined as the distance in the inferior-superior direction between the top and bottom of the contour, after reorientation to standard space, at every point along the length of the segment, then average across the region of interest. These metrics serve as input for the ensemble machine learning model used for labelin CC segmentations as having passed or failed quality control (QC). Abbreviations: Montreal Neurological Institute - MNI, CC - corpus callosum, ML - Machine Learning, KNN - K Nearest Neighbors, SVC - Support Vector Classifier