Positron emission tomography (PET) is a nuclear medical imaging technology that produces 3D images of tissue metabolic activity in human body. PET has been used in various clinical applications, such as diagnosis of tumors and diffuse brain disorders. High quality PET image plays an essential role in diagnosing diseases/disorders and assessing the response to therapy. In practice, in order to obtain the high quality PET images, standard-dose radionuclide (tracer) needs to be used and injected into the living body. As a result, it will inevitably increase the risk of radiation. In this paper, we propose a regression forest (RF) based framework for predicting standard-dose PET images using low-dose PET and corresponding magnetic resonance imaging (MRI) images instead of injecting the standard-dose radionuclide into the body. The proposed approach has been evaluated on a dataset consisting of 7 subjects using leave-one-out cross-validation. Moreover, we compare the prediction performance between sparse representation (SR) based method and our proposed method. Both qualitative and quantitative results illustrate the practicability of our proposed method.
|Number of pages||9|
|Journal||Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)|
|Publication status||Published - 2014 Jan 1|
ASJC Scopus subject areas
- Computer Science(all)
- Theoretical Computer Science