Hi, The picture is purely trained to detect people, but the training is done in grayscale, because only .tflite is needed. Can you use yolov5 instead of FOMO?
Thanks!