Development of an ultrasound image-based platform for guiding intravenous injection

Abstract

To increase the accuracy of intravenous injection and blood sampling medical personnel typically use a rubber band tied down the arm to make the vessel swell before inserting needle However the positional accuracy of inserting needle mainly depends on the experience of medical personnel If the subject has such situations as atrophicans vasculare small vessel thick subcutaneous fat etc the position of vessel could be difficultly observed to affect the accuracy for inserting needle Therefore this study developed an ultrasound-guided needle system for assistive intravenous injection which combined a needle assistive device and vessel detecting from ultrasound image during intravenous injection to improve the injection accuracy First this system automatically detected the position of maximum vessel size by a series of ultrasound image processing included automatic ROI selection of image contrast enhancement filtering thresholding smoothing initial edge extraction edge deciding and calculation of vascular size Subsequently the appropriate angle of needle insertion was calculated by trigonometric functions to assist needle inserting In addition the wavelet thresholding method was used to reduce the artifact from needle to enhance the recognition of needle on ultrasound image The results showed the percentage error is depend on the frequency (or resolution) of the transducer If use higher frequency (or resolution) of the transducer then the percentage error can be reduced; in the part of artifact reduction the result image will change with brightness prominent threshold value By comparing with the contrast value we select the value 0 7 to be the brightness prominent threshold This study demonstrated that the ultrasound-guided needle system could immediately accurately safely and conveniently find the central position of vessel Furthermore the components of the system mainly consist of only a needle and transducer holder and the developed software Therefore it tends to be cost-effective and may be applied to improve the accuracy of needle injection for resource-scarce communities

Date of Award	2015 Jan 22
Original language	English
Supervisor	Shyh-Hau Wang (Supervisor)