X-ray image segmentation based on genetic optimization and watershed transformation

X-ray image is important foundation in clinical medical diagnosis, but its exact segmentation is difficult to obtain because of the mixture of noise and artifact. Watershed transformation has the virtues of accurate positioning and the achievement of continuous closed borders, however, the accurate results of segmentation depends on correct choose of structure element. This article proposes a method that combines the genetic optimization and watershed transformation, and this method has solved the problem that the choose of structure element in morphology operation depends on the researcher´s experience, which also realizes the auto select of structure element. This method is proved feasible by simulation result.