Research on Endpoint Information Extraction for Chemical Molecular Structure Images

摘要

The methods of automatic extracting information of static chemical molecular structure images are introduced in this paper. Endpoint information extraction is the key research contents, and it mainly includes the following two parts: the endpoint character recognition based on BP neural network; the postprocessing operation, such as the character combination of endpoint, the endpoint information correction based on dictionary. This is the base for automatically extracting chemical molecular structure images.