Hi there,
using depth_estimation_stereo on Raspberry Pi5 with Hat+ 2, I am having trouble to reproduce the expected result. The code builds and executes fine on my current setup, but the output looks odd. Using the reference images left.jpg and right.jpg as input, I get some distorted output. The same happens when using camera input from 2 camera3 modules. I am using the default net hailo10h/stereonet.hef in this setup.
Does this setup work for anybody out there?
Output image attached
