spedtime Hi, I was looking to do an art installation with low light pose detection. I saw this last question: https://discuss.luxonis.com/d/728-night-depth-vision-w-body-skeleton-tracking/2 As a person who is new to NN, how hard is it to retrain the pose models for grayscale? Apologies I couldn't find much more info in the discussions.
jakaskerl spedtime Maybe you don't need to retrain it and can just pass in 3x grayscale to mock the 3-channel RGB input.
