This uses a pre-trained, remotely located CocoSsd model to detect objects in an image which is randomly selected from images from the 2022 Festival catalog. (Reload the page for a new image). This model can detect 80 classes of objects - quite a limited number, and like most pre-trained models, culturally specific. The 'objects' in the Festival images tend mostly to be people (since popoli = people). Object detection differs from image classification in that this model can recognise multiple objects in one image, whereas the other one classifies the whole image as a single 'object'.

Find the code here.