处理带有变形凸性物体灰度图像的CNN新模板(英文)

NEW TEMPLATES CNN FOR PROCESSING GRAY-SCALE IMAGE WITH DISTURBED CONVEX OBJECTS

细胞神经网络 ( CNN)是图像处理的有力工具 .它已用于人工视觉 ,录像压缩 ,图像融合、运动和图形识别等领域 .本文提出了一组 CNN新模板 ,用于恢复数字化灰度图像中扭曲凸物体像的凸性 .通过计算机模拟 ,利用凸性恢复 CNN处理了一幅带有高斯噪声的数字化理想圆灰度图像 .处理后图像中的理想圆比用识别算子处理的图像具有更好的凸性 .可以预期

The cellular neural network (CNN) is a powerful tool for image processing and has been used for artificial vision,video compression,image fusion,motion and pattern recognition.This paper presents a set of new CNN templates for recovering convexity of disturbed convex objects in digitized gray scale images.The computer simulation showed that comparing to the recognizing operator,the CNN can recover more effectively degraded convex circles by Gaussian noise.It is expected that the new CNN templates can be used for analyzing electronic diffraction patterns (EDPs) of crystal and star images in the Galaxy.