| 山东科技大å¦ç¡•士å¦ä½è®ºæ–‡æ‘˜è¦Abstract The main advantages of ultrasonic diagnosis are non invasive, fast and safe, etc. Doctorscan't observe three dimensional tissue structures accurately by traditional two dimensionalultrasound images and they make diagnosis heavily depending on their medical experientialknowledge.Therefore, three dimensional reconstruction technology of medical ultrasonicimage plays an important role in clinical diagnosis and surgical planning and others, whichgradually become the research, focus of medical image processing. This paper preprocesses B-type ultrasound image firstly,this paper studies for B-typeultrasound image preprocessing and image segmentation method aspects and the algorithm,giving an improved Otsu method. Image enhancement and noise reduction methods have alsobeen improved to solve the traditional ultrasound images of dark, low contrast and image noiseand other problems exist. Traditional MC method has defects such as Inefficient and many triangular patches, thispaper propound an improved MC method based on the summary of former achievements andanalysis of application requirements. In this method the isosurface can be protracted directlybased on the function value distributing of the cube vertices. The process to look up theconfiguration style in the predefined table is avoided. Meanwhile a dissection method to solvethe ambiguity problem is developed. An example of three-dimensional reconstruction frommedical images is presented and the results show that this algorithm accelerates the speed oftraditional MC algorithm. In the end,three-dimensional reconstruction of B-type ultrasound images was achieved inVC++6.0 environment. Keywords:Ultrasound image, Three dimensional reconstruction, Maximum variance between clusters, Surface reconstruction,Improved MC method... |