Common carotid artery condition recognition technology using waveform features extracted from ultrasound spectrum images

Medical image recognition algorithms have been widely applied to help with the diagnoses of various diseases, reducing human resource investment while enhancing diagnostic accuracy. This paper proposes a new scheme that specifies in the reading of ultrasound spectrum images of common carotid artery blood flow. The proposed scheme automatically extracts effective waveform features from the images for diagnostic purposes by using five criteria, which are ratio of waveform region, waveform region area target under horizontal baseline, waveform region area under horizontal baseline, highest point of waveform region, and lowest point of waveform region. Traditionally used by physicians to differentiate between normal blood flow patterns and five abnormal blood flow types, these five criteria are now employed by the new scheme to digitally diagnose vascular disorders at an accuracy rate as high as 0.97.

► An automatic classification method by utilizing effective waveform features is proposed.
► The main criteria are area above baseline, area under baseline, the highest point, and the lowest point.
► The proposed result works well on analysis of blood flow diseases.