A rule-based approach for recognition of chemical structure diagrams

摘要: In chemical literature much information is given in the form of diagrams depicting structures. order to access this electronically, have be recognized and translated into a processable format. Although number approaches been proposed for recognition molecule literature, they traditionally employ procedural methods with limited flexibility extensibility. This thesis presents novel approach that models principal steps strictly rule based system. We develop framework enables definition set rules different bond types arrangements as well resolving possible ambiguities. allows us view diagram problem process rewriting an initial geometric artefacts graph representation without need adhere rigid procedure. demonstrate by extending it capture new compositions. experimental evaluation we can show implementation our outperforms currently available leading open source Finally, discuss how could applied other automatic tasks.