Detecting and tracking of ASD using new block flow in ultrasound (a novel approach)

Medical ultrasound imaging is widely used to diagnose cardiac disorders and diseases. Atrial Septal Defect (ASD) is one of the known congenital heart defect that can be detected using ultrasound (Echocardiogram). There are many technique that have been developed for ASD detection (Chest X-ray, Electrocardiogram (ECG), Magnetic resonance imaging (MRI) and many other techniques). But sometimes these techniques are not accurate enough and need more tests to ensure the presence of ASD, in addition consuming a lot of time. In this work, a technique was developed called Dynamic Multi Shape Search (DMSS). This technique depends on detecting ASD and tracking blood flow; which combines the advantages of block matching and optical flow techniques. It uses a search window that divides the ultrasound image into blocks and finds the minimum difference between the test points of the test image with the standard one of ASD using Sum Absolute Difference (SAD) criteria, then applied three steps with semi diamond, square and diamond shapes to find the best match. The propos method compared with the baseline methods; which are less accurate, consume a lot of time and some of them painful for patient. The results showed validity for our method as competitive results quality-wise, and a significant gain speed-wise.