I collected a large number of photos from social media. I would like now to use KNimes image processing to detect if it is a photo (high # of colors) or a graphic (low # of colors). Any recommendations on how to easily (and fast) get this accomplished?
even if it sounds easy, I don't think its a straighforward thing. However, there are some things you can try: Extract features from your image (e.g. first order statistics using the "Image Feature Node") and then you can try simple clustering (with 2 clusters) or if you have marked data train a model (e.g. Decision Tree) to disinguish between charts and "real" images/photos.