AdamPolak I am using the face detection model from this experiment: https://github.com/luxonis/depthai-experiments/blob/30e2460557a3209770eb8943db41bc997a423212/gen2-face-detection/main.py#LL30C4-L30C4 The NN node needs post-processing before getting the final detection boxes. Once I do the final post-processing is it possible to send this data to a SpatialLocationCalculator on device to find X,Y,Z coordinates? I would need to pass through the nn data sequence number but if I package it the right way, can the SpatialLocationCalculator node do the calculation with a depth frame?
erik AdamPolak I'd suggest sending back the depth frames as well, and sending both corodinates + depth to the SLC. Or perhaps just calculate the XYZ on the host (example here) after decoding the NN result.
AdamPolak erik I cannot send the depth to host because of bandwidth so I need to keep it on device. Is there a format I can use to send from a script node and "create" a ImgDetection message to SpatialLocationCalculator on the device?
erik Hi AdamPolak , Yep, you can do that inside Script node. Example here: https://github.com/geaxgx/depthai_hand_tracker/blob/main/template_manager_script_solo.py#L176-L187
