Automatic Recognition of Chemical Images

This is the authors' abstract. We don't add key points for this paper.

Images of chemical molecules can be produced, manipulated, simulated and analyzed using sophisticated chemical software. However, in the process of publishing such images into scientific literature, all their chemical significance is lost. Although images of chemical molecules can be easily analyzed by the human expert, they cannot be fed back into chemical software and loose much of their potential use. We have developed a system that can automatically reconstruct the chemical information associated to the images of chemical molecules thus rendering them computer readable. We have benchmarked our system against a commercially available product and have also tested it using chemical databases of several thousand images with very encouraging results.