As one of the most important advances in the field of medical imaging in the 20th century, MRI magnetic resonance imaging technology has been widely used in medical clinical diagnosis, so it is of great practical significance to study magnetic resonance imaging and its image processing methods.
In this paper, several main aspects of magnetic resonance medical imaging and image processing methods are studied. It mainly involves three sub-topics: the study of extended two-point Dixon water-fat separation algorithm based on chemical shift, including both specific imaging pulse sequence design and image post-processing;
Research on image enhancement, denoising and high-resolution image reconstruction algorithm based on nonlinear filtering: Research on progressive lossless compression algorithm of medical image based on integer wavelet transform and improved zerotree coding.
In this paper, the basic physical principles of magnetic resonance imaging are systematically reviewed, and on this basis, the extended two-point Dixon algorithm for water-lipid separation based on chemical shift is studied, and a low-pass filter is proposed to replace polynomial fitting iteration for two-dimensional phase unwrapping. The improved algorithm can reduce the computational complexity of separation treatment and improve the separation effect of water and fat.
In order to improve the quality of MRI medical images, this paper studies and analyzes linear enhancement algorithm and nonlinear filtering extrapolation image enhancement algorithm, and points out the reasons that lead to Matthew effect when enhancing the whole image.
Furthermore, a new clipping strategy envelope threshold clipping strategy is proposed to improve the nonlinear filtering algorithm, which makes the improved algorithm obviously superior to the original algorithm in extrapolating new high-frequency components for image enhancement. Using improved nonlinear filtering algorithm combined with low-pass filtering to denoise medical images can effectively eliminate high-frequency noise and keep useful high-frequency signals as much as possible.
Finally, the improved nonlinear filtering method is applied to high-resolution image reconstruction, and a better high-resolution reconstructed image is obtained than linear interpolation.
This paper briefly reviews integer wavelet transform and EZW zerotree coding algorithm, studies the shortcomings of EZW zerotree coding strategy in lossless image compression, and proposes a progressive lossless compression framework for medical images based on integer wavelet transform and improved zerotree coding.
The lossless compression experiment of medical images has achieved a high compression ratio. When the lossy progressive decoding is resumed, the lower bit rate achieves a better image signal-to-noise ratio, and at the same time, it has good progressive decoding characteristics, which can meet the needs of image decompression applications based on channel transmission, such as telemedicine.