This project aims at creating algorithms to simulate dose reduction in images corrupted by Poisson-Gaussian noise. We have performed validation studies using Digital Mammography, Digital Breast Tomosynthesis and Fluorescent Cells images. A pre-clinical trial study was conducted using clinical DBT patient data and human observers.
This project aims at improving image quality of x-ray images by developing appropriate image restoration techniques. The methods have been validated on Digital Mammography, Digital Breast Tomosynthesis and Molecular Breast images. A pre-clinical trial was conducted using clinical DBT patient data and human observers. A clinical trial is currently being conducted using MBI images and radiologist at a major clinical center in the USA.
This project aims at developing methods for assessing image quality from clinical images using a non-reference quality metric. The validation has been performed using a virtual clinical trial pipeline and a set of DBT images from patients. The next step of this project is to correlate the metric with the perception of quality reported by radiologists.
This project aims at improving reconstruction methods to minimize the out-of-plane artifacts generated by high contrast features of the image. Preliminary phantom studies have been performed on a DBT system. The next step of this project is to test the performance of the method on musculoskeletal tomosynthesis images.
