Fuente: Wipo "digitalization" Searching and identifying digital images based on semantic and spatial information, comprising: receiving user input query area 110, 112 (indicating target region for content) and query term 110a, 112a (indicating target content); generating query feature set 108 from query area and query term, using neural network; identifying images 106a portraying target content within region by comparing query feature set to feature sets learned from plurality of images within neural network. Searching and identifying images based on semantic and spatial information, comprising: receiving user input query area and query term; determining query feature set from query area and query term by generating representations of query term and query area, and providing representations to neural network; identifying images portraying target content within region by comparing query feature set to feature sets learned from plurality of images within neural network. Identifying images based on semantic and spatial information, comprising: plurality of feature sets for plurality of images, extracted from image neural network that preserves semantic and spatial information; neural network trained to generate query feature sets; server for storing instructions for: generating representations of query area and term; generating query feature set from representations; identifying images by comparing with query feature set.