A System for Real-Time Syringe Classification and Volume Measurement Using a Combination of Image Processing and Artificial Neural Networks
- 124 Downloads
The purpose of this research was to develop a system that can read and report the volume of liquid medication present in syringes.
The system is comprised of a digital webcam which is designed to communicate with a computer program developed using MATLAB. The system includes two functional modules, one supporting syringe classification, and another supporting volume measurement. Adaptive template matching was used to determine the best match point between target and template images. The connected component labeling method was used during volume measurement. An artificial neural network (ANN) model was developed using MATLAB to support the intended volume measurement functionality. The developed ANN was designed as a classifier which determines the plunger depth of the syringe and then leverages this result to calculate and derive the volume of medication inside the syringe as the final system output. Commercial Luer-lock syringes of sizes from 1 to 30 ml were used in conjunction with syringe tip caps of blue and yellow color. Tap water or aqueous dye solutions of yellow, red, and blue color simulated liquid medication in the syringe.
The developed syringe classification system successfully detected and categorized all tested syringes according to their size. The best accuracy of the system was found to be 99.95% with a 3-ml syringe, while the worst accuracy was 95.82% with a 5-ml syringe. It took approximately 6 s to perform the entire task demonstrating the utility of this system to report volumes in real time.
The developed system can be used across a variety of settings that routinely support measuring and handling liquids in syringes including hospitals, pharmacies, and the pharmaceutical industry.
KeywordsCompounded sterile preparations Real-time measurement Syringe volume Dose accuracy Image analysis Artificial neural networks
- 2.Aspden P, Wolcott JA, Bootman JL, Cronenwett LR. Preventing medication errors. National Acad. Press; 2007.Google Scholar
- 7.Sasich LD, Sukkari SR. Unknown risks of pharmacy-compounded drugs. The Journal of the American Osteopathic Association. 2008;108(2):86-.Google Scholar
- 8.ISMP list of high-alert medications in acute care settings. Institute for Safe Medication Practices 2014. http://www.ismp.org/tools/highalertmedications.pdf. Accessed 4 Aug 2017.
- 13.Forcinio H. Trends and best practices in visual inspection: using best practices for manual or automatic inspection can improve the inspection process. Pharm Technol. 2014.Google Scholar
- 15.Wichtl M. Herbal drugs and phytopharmaceuticals: a handbook for practice on a scientific basis: CRC press; 2004.Google Scholar
- 18.Gunatilake P, Siegel M, Jordan AG, Podnar GW, editors. Image understanding algorithms for remote visual inspection of aircraft surfaces. Electronic Imaging’97; 1997: International Society for Optics and Photonics.Google Scholar
- 19.Acquire Images from Webcams. Mathworks, MA, USA. 2016. http://www.mathworks.com/help/supportpkg/usbwebcams/ug/acquire-images-from-webcams.html. Accessed 12 Aug 2016.
- 20.Mohr D, Zachmann G, editors. FAST: Fast Adaptive Silhouette Area based Template Matching. BMVC; 2010.Google Scholar
- 23.Roy S, Rathod D. Real-time object tracking and learning using template matching.Google Scholar
- 25.Sharmila R, Uma R. A new approach to image contrast enhancement using weighted threshold histogram equalization with improved switching median filter. Int J Adv Eng Sci Technol. 2011;7:206–11.Google Scholar
- 26.Otsu N. A threshold selection method from gray-level histograms. Automatica. 1975;11(285–296):23–7.Google Scholar
- 28.Yapa RD, Harada K. A connected component labelling algorithm for greyscale mammography image processing as a pre-processing tool. Mach Graph Vision Int J. 2007;16(3):305–27.Google Scholar
- 29.Walczyk R, Armitage A, Binnie D. Comparative study on connected component labeling algorithms for embedded video processing systems. 2010.Google Scholar
- 30.Rudrapatna M, Sowmya A, editors. Feature weighted minimum distance classifier with multi-class confidence estimation. Australasian Joint Conference on Artificial Intelligence; 2006: Springer.Google Scholar
- 32.Gerstner W. Supervised learning for neural networks: a tutorial with Java Exercises. Recurso disponible on-line: http://diwww.epfl.ch/mantra/tutorial/english/supervised.pdf. 1998.
- 33.Fine S, Scheinberg K. Efficient SVM training using low-rank kernel representations. J Mach Learn Res. 2001;2(Dec):243–64.Google Scholar
- 34.Vlacic L. Learning and soft computing, support vector machines, neural networks, and fuzzy logic models. Vojislav Kecman; MIT Press, Cambridge, MA, 2001, ISBN 0-262-11255-8, 2001, pp. 578. Elsevier; 2002.Google Scholar
- 35.Yu H, Wilamowski BM. Levenberg-Marquardt training. Ind Electron Handb. 2011;5(12):1.Google Scholar