id	author	title	date	pages	extension	mime	words	sentences	flesch	summary	cache	txt
work_xt5lrac64vgajg4ssffkqsiucu	Marina Ivasic-Kos	A knowledge-based multi-layered image annotation system	2015	15	.pdf	application/pdf	14263	2008	64	with uncertain and ambiguous knowledge and can annotate images with concepts on different levels of abstraction that is more human-like. of the classified image segments, automatic scene recognition, and the inference of generalized and derived The key components of the proposed multi-layered image annoation system are the KRFPN scheme based on Fuzzy Petri Net foralism and integrated fuzzy inference engine. mages in the Fig. 1, one can use concepts that are related to the obects that appear in the image (sand, sea, sky, snow), concepts that repesent the scene (beach, coast, coastline, shore, seashore), more general For instance, the proposed multi-layered image annotation related to Fig. 1c is EC = {snow, polar bear}; SC = {Scene-Polarbear}; GC defined according to expert knowledge; a set � includes a relationship occurs_with between elementary classes that models the common occurrence of elementary classes in the image and its negation	./cache/work_xt5lrac64vgajg4ssffkqsiucu.pdf	./txt/work_xt5lrac64vgajg4ssffkqsiucu.txt