3D IOU metrics for arkit and mlcore object detection

I am using object detection to classify an object based on a arframe passed to coreml. I know in 2d space you can calculate an IOU score, is their a way to get a IOU in arkit? It would require the x, y, z dimensions. I would like to test the IOU in three space in order to reduce the number of overlaps in already exisiting objects.