Research And Implementation On De-noising And Edge Detection Algorithm For Medical Image Based On Wavelet Transform

Medical image has become an important basis for doctors in the diagnosis and treatment, it contains abundant information of human body. In order to show the lesions site of the human body to doctors in a more clear and intuitive way, it must be processed. Image de-noising and edge detection is an important step in medical image processing, so It is important to the subsequent processing. This paper mainly researches the image wavelet threshold de-noising and morphological edge detection, and apply to medical image prove the effectiveness of the proposed algorithm, the main contents are:1. Aiming at the disadvantages of soft and hard threshold function, this paper proposes an improved threshold function. This function has two thresholds, it retains the useful information as much as possible by gradually increasing the shrink of wavelet coefficients between the two thresholds, until the coefficients reduced to zero. This paper proves continuity and error analysis of the improvement function. Finally, we demonstrate and validate the de-noising effect of this new threshold function by simulation experiment and objective evaluation standards.2. Aiming at the characteristics of mixed noise, this paper presents a method of de-noising that median filtering and wavelet transform combined. Firstly, we use the median filtering. And then to remove noise by wavelet transform. Finally, we demonstrate and validate this method is better than traditional soft, hard threshold and median filtering by simulation experiment and objective evaluation standards.3. Aiming at the traditional edge detection is very sensitive to noise, this paper presents a method that wavelet threshold de-noising and morphological edge detection combined. Firstly, we de-composite the image by wavelet transform. And then to take the high frequency coefficients by wavelet threshold de-noising, a new designed double structure edge morphology operator is used in the treatment of low frequency coefficient. Finally, we use the new coefficients to reconstruct the image. The experimental results show that this method can remove the noise while preserve edge information. It is an effective edge detection method and have an important reference value for actual application.