Protein crystallizing state discrimination based on image processing

In recent years, the research related to automation of the crystallization process has been examined in several projects. In usual crystallization process, the researchers observe the crystallization states in the protein solution samples and record the state based on one's visual impression. This evaluation is very important but requires a great deal of time. It should be done efficiently, therefore we started to discuss a method to classify protein crystallization states using image processing. We develop a state discrimination method and examine its performance.