Cardiovascular diseases are presented as the first cause of death in western world in the last decades. Creating new technologies capable of extracting relevant information from the body structures are crucial to improve prevention and diagnosis Cardiovascular parameters such as blood pressure (BP) and pulse arrival time (PAT) are ...»»»»
Cardiovascular diseases are presented as the first cause of death in western world in the last decades. Creating new technologies capable of extracting relevant information from the body structures are crucial to improve prevention and diagnosis Cardiovascular parameters such as blood pressure (BP) and pulse arrival time (PAT) are strongly linked with current state of blood vessels.
The objective of this project is to develop an algorithm for direct detection and qualification of the common carotid artery (CCA) to enhance the interpretation of blood vessels current state. We propose a different technique to allocate the CCA as an input of an overall wall tracking system (WTS).
Matlab version R2018a was used to design and evaluate the algorithm. A dataset containing ultrasound data from 10 subjects was analyzed. The approach to detect the walls was based on the hypo-echogenic nature of the lumen. The denoising method used was The Meyer wavelet filter followed by a Hilbert transform. Further application of heuristics was needed in order to identify the wall positions and the corresponding quality indicators. The performance was evaluated in the context of wall tracking system (WTS) by comparing manually annotated references to spatial positions from a new, ECG-gated wall tracking system using a reset per heartbeat versus the conventional tracking system using only a single initial starting position prior to the recording. The results proved that this new technique can greatly detect the CCA and be an important module block in the novel WTS. The evaluation showed a significant increase in temporal efficiency (9.53 times faster than manual annotations), high spatial accuracy, and improved feature consistency. In addition, it allows for quality assessment in terms of dispersion and contrast criteria, making the algorithm dynamic and powerful. Nevertheless, problems concerning the anatomy of the vessels were encountered. Deeper lying vessels may need another localization approach such as dynamic time warping (DTW). Such that full automation remains to be implemented.^^^^
