In the images with local strong direction, there is often noise in other directions. To elimi-nate the noise, we propose a novel directional filtering method which takes advantage of directional se-lection ability of ...In the images with local strong direction, there is often noise in other directions. To elimi-nate the noise, we propose a novel directional filtering method which takes advantage of directional se-lection ability of wavelet transform skillfully through rotating the image to be filtered and controls the times of rotation through direction quantization. This method employs only directional characteristics for denoising and there is no need to evaluate the frequency characteristics or statistical characteristics of signals and noise, which makes this method easy to implement. The experimental results on finger-print image enhancement show good ability of this method.展开更多
文摘In the images with local strong direction, there is often noise in other directions. To elimi-nate the noise, we propose a novel directional filtering method which takes advantage of directional se-lection ability of wavelet transform skillfully through rotating the image to be filtered and controls the times of rotation through direction quantization. This method employs only directional characteristics for denoising and there is no need to evaluate the frequency characteristics or statistical characteristics of signals and noise, which makes this method easy to implement. The experimental results on finger-print image enhancement show good ability of this method.
