Among the diagnostic imaging technologies, tomographic imaging modalities are indispensable to get a clear 3D image of a volume. Positron emission tomography (PET) and computed tomography (CT) are two such modalities. Accurate interpretation of PET and CT images is crucial for effective use of these imaging technologies in clinical practice. Artifacts, which are unwanted effects that hinder the detection of the structures in the body, can prevent this by degrading the image quality from acquired measurements.
In her thesis Defne Us proposes practical solutions for reducing the artifacts caused by high-density materials in dental CT images and missing measurements in partial ring PET scanners used in cancer screening and treatment.
“The artifacts are reduced largely by means of new image reconstruction methods,” she says.
Image reconstruction algorithms transform measurements from scanners into anatomically relevant images. The lower radiation dose requirements due to increased concern for patient safety introduce additional difficulties for reconstruction methods, requiring different approaches than earlier high-dose reconstruction methods. Another trend in imaging is to combine imaging technologies together, as each imaging modality has its strengths and weaknesses. Although combining these technologies can produce superior images, it can also present new challenges for the algorithms, such as the use of PET in proton therapy. In her work Defne Us also considered such challenges for the reconstruction methods.
“The new methods involve incorporating our knowledge on the measurement system and noise into the reconstruction methods in a robust way. The results point out the potential benefit of solving application-specific problems efficiently when designing new algorithms and other improvements in tomographic imaging methods,” Us says.
“With artifact-free images enabled by the use of these new reconstruction methods, the patients would need fewer repetitions of scans, leading to lower overall radiation dose. Without the artifacts from limited-angle measurements, diagnostically acceptable images from fewer angles can be achieved in the future. Such images would also make it easier for radiologists and physicians to decide on a course of action based on reliable and good-quality information.”
The doctoral dissertation of MSc (Tech) Defne Us in the field of signal processing titled Reduction of limited angle artifacts in medical tomography via image reconstruction will be publicly examined at the Faculty of Information Technology and Communication Sciences at Tampere University at 12:00 on weekday 08.02.2019 in auditorium TB109 in Tietotalo building, Hervanta Campus (Korkeakoulunkatu 1, Tampere). The Opponents will be Associate Professor Martin S. Andersen from Technical University of Denmark and Chief Physicist Antti Sohlberg from Joint Authority for Päijät-Häme Social and Healthcare. The Custos will be Professor Ulla Ruotsalainen from Faculty of Information Technology and Communication Sciences.
The dissertation is available online at http://urn.fi/URN:ISBN:978-952-03-0964-0.