This　paper　presents　a　new　method　for　detecting　the　edge　of　intravascular　ultrasound　images,　based　on　geometric　active　contours.　The　medical　image　segmentation　method　is　based　on　the　combination　of　traditional　Level　Set　method　and　B-Spline　models.　The　new　method　only　needs　a　few　control　points　to　achieve　the　effect　of　smoothing　curves　with　drastic　changes.　Compared　with　the　traditional　level　set　function,　this　method　has　many　advantages,　such　as　no　re-initializing　to　signed　distance　function,　a　higher　curve　smoothness　and　simple　evolution　and　so　on.　Experimental　results　show　that　the　algorithm　can　obtain　segmentation　results　of　medical　image　fast　and　accurately,　and　the　method　is　reproducible　and　robust　in　image　processing.