Edge detection with BP neural networks
暂无分享,去创建一个
A new edge detection technique is proposed which makes use of a backpropagation (BP) neural network. We classify the edge patterns in binary images into 18 categories. After training on the pre-defined edge patterns, the neural network is applied to classify any type of edge into one of the 18 categories. For different edge patterns, corresponding algorithms are employed to process the image to extract edges. Compared with the existing edge detection techniques, our method is more flexible to the variation of edges. The results indicate that the new technique provides better results than traditional edge detection techniques.
[1] Mahmood Fathy,et al. A combined edge detection and background differencing image processing approach for real-time traffic analysis , 1995 .
[2] Josef Kittler,et al. On threshold selection using clustering criteria , 1985, IEEE Transactions on Systems, Man, and Cybernetics.
[3] Ling Guan,et al. Neural network approach to adaptive image enhancement , 1992, Electronic Imaging.
[4] Mahmood Fathy,et al. Neural-vision based approach for real-time road traffic applications , 1997 .