摘要
Positron emission tomography (PET) is a functional medical imaging modality which is widely used in clinical applications for diagnosis of disease. Since PET images have high level of noise and low spatial resolution, the exploration of its denoising methods is of great and practical significance. Many denoising methods have been used for PET images, but they may not have good edges and lesions preserving property while denoising. In this paper we propose a graph wavelet transform (GWT)-based denoising method for dynamic PET images with multi-feature information. Combining PET composite images, the gradient information of PET composite images and magnetic resonance (MR) images as the prior information are used to construct the graph Laplacian matrix, and then the GWT method is applied to denoise dynamic PET images. The rational quadratic kernel function is used to construct adjacency matrix. Experimental results show that the proposed method can provide better denoising effect, and outperforms other GWT-based PET images denoising methods.
学科与关键词
- Dynamic PET images denoising using spectral graph wavelet transformYi, Liqun; Sheng, Yuxia*; Chai, Li; Zhang, JingxinMedical and Biological Engineering and Computing, 2023, 61(1): 97-107.被引 1
- Dynamic PET Image Denoising With Deep Learning-Based Joint FilteringHe, Yuru; Cao, Shuangliang; Zhang, Hongyan; Sun, Hao; Wang, Fanghu; Zhu, Huobiao; Lv, Wenbing; Lu, Lijun*IEEE Access, 2021, 9: 41998-42012.被引 8
- Dynamic PET Image Denoising Using Deep Image Prior Combined With Regularization by DenoisingSun, Hao; Peng, Lihong; Zhang, Hongyan; He, Yuru; Cao, Shuangliang; Lu, Lijun*IEEE Access, 2021, 9: 52378-52392.被引 20
