EE Seminar: Denoising of medical images acquired with radiation constraint
Speaker: Michael Green
Ph.D. student under the supervision of Prof. Nahum Kiryati and Dr. Arnaldo Mayer
Sunday, January 19th, 2020 at 15:00
Room 011, Kitot Bldg., Faculty of Engineering
Denoising of medical images acquired with radiation constraint
Abstract
Computed Tomography (CT) has become an invaluable tool in diagnostic imaging, providing fast, cost effective, and high-quality imaging in countless indications.
Mammography is the only imaging modality cleared by the FDA for breast cancer screening, saving each year million of lives around the word.
Both imaging modalities share the same disadvantage, the exposure of the patient to a significant dose of x ray radiation which may increase the risk of developing cancer.
Decreasing radiation dose will progressively degrade the image quality as the amount of noise increases, until the diagnostic value of the image is lost.
In this research, novel image denoising techniques have been developed to enhance CT and mammography images acquired at ultra-low radiation doses, corresponding to a few percent of the normal doses used in clinical routine. The resulting denoised images are visually similar to normal dose images and have recovered their diagnostic value.
During this talk, I will focus on different directions we used for developing a good and stable denoising algorithm, for removing the noise while maintaining small and important details.
