Extracting the contours of objects in intricate background is a problem which virtual image processing system must face, however, traditional contour-extrcacting method is liable to fail because of too much noise. The contours of SDC images processed in our paper immerse in noise familiar to actual contours, so which is difficult to extract. Using extracted contours of UBF images, we build string models and excellently extract those gastric contours of SCD images. The method has a valuable reference to other image processing in intricate background.%复杂背景下的物体轮廓提取是一个实际的图象处理系统必须面对的问题，而传统的轮廓提取方法在此情况下往往因为大量的噪声而失效。本文所处理的胃部的SDC图象中物体的边缘就处在与边缘的斜率和灰度梯度极为相近的噪声中，所以较难提取。我们利用UBF图象已提取的轮廓建立弹簧模型，根据两种图象中胃部的形变关系，较好的提取了SDC图象中的胃部轮廓。该方法对其它复杂背景下的图象处理有较好的参考价值。
展开▼
