Circular particle detection using sectored ring mask for synchrotron PCXI images

Cystic Fibrosis (CF) is a genetic disorder that compromises the respiratory function and the ability of the mucociliary transit (MCT) system. One of the most recent researches introduced a noble method to investigate the progress of the treatment, in which small particles with mostly circular shape injected to the respiratory system and the images were taken using Synchrotron X-ray beam. Since the small particles flow through the respiratory system of the body, the direct observation of MCT measurement will help to understand the progress of the treatment. Identifying the particle is the critical step towards the automatic analysis of the image. However, the objects of interests are usually very small, not perfect circular shape and slightly overlapped from each other with lots of noise due to radiation. This paper proposes a robust and effective detection method of such particles using sectored ring mask (SRM) with gradient descent method. The proposed method extracts strong edges of the particles and the edge line gradients and circle fitting algorithm will filter out invalid edges, resulting in clear particle edge detection. The proposed method has validated through experimental study and presented robust detection rates of 91.9% precision and 89.0% recall.