کاهش پیچیدگی محاسباتی در جداسازی تصاویر پزشکی
محورهای موضوعی : مهندسی پزشکی- بیوالکتریک
رضا غفوری
1
,
علیرضا آل سعدی
2
,
بابک غلامی
3
1 - گروه مهندسی پزشکی، واحد کازرون، دانشگاه آزاد اسلامی، کازرون، ایران
2 - گروه مهندسی برق، واحد کازرون، دانشگاه آزاد اسلامی، کازرون، ایران
3 - گروه برق، واحد کازرون، دانشگاه آزاد اسلامی، کازرون، ایران
کلید واژه: جداسازی, تصاویر پزشکی, مجموعه های هم سطح, واریانس دودویی, پیچیدگی محاسباتی, مشتقات تصویر,
چکیده مقاله :
با توسعه روشهای جداسازی تصاویر پزشکی مبتنی بر هوش مصنوعی، شاهد استفاده روزافزون از این روشها در جداسازی بافتها و اندامها در انواع تصاویر پزشکی هستیم. جداسازی تصاویر پزشکی نقش بسیار مهمی در بررسی خودکار ساختار اندامها، به دست آوردن ابعاد و اندازه آنها و در برخی موارد، بررسی عملکردی آنها دارد. روشهای مبتنی بر هوش مصنوعی در کنار مزایایی چون دقت بالا و قابلیت یادگیری و بهبود عملکرد، به علت حجم بالای محاسبات و نیاز به داده های زیاد و متنوع در فرآیند آموزش، میتوانند سبب کاهش سرعت رسیدن به پاسخ شوند. این نکته در شبیه سازی روشهای پردازش موازی با کامپیوترهای دیجیتال بیشتر به چشم میخورد زیرا در هر پالس ساعت پردازنده، حداکثر یک سلول شبکه را میتوان به روزرسانی نمود. در این مقاله از یک روش تحلیلی برای جداسازی تصاویر پزشکی استفاده خواهیم کرد. سپس یک روش مناسب برای تقرب به پاسخ را انتخاب میکنیم و در نهایت با کاهش پیچیدگی محاسباتی مورد نیاز برای نیل به هدف جداسازی، سرعت آن را افزایش خواهیم داد. مقایسه سرعت الگوریتم با روش پیشنهادی در این پژوهش با حالت متعارف، نشان دهنده افزایش قابل قبول در سرعت پردازش و رسیدن به پاسخ نهایی بدون از دست دادن دقت جداسازی است
With the development of AI-based medical image segmentation methods, there is and increasing demand for using these methods in segmentation of tissues and organs in medical images [1]. Medical image segmentation plays a very important role in automatically examining the structure of organs, obtaining their dimensions and in some cases, examining their functionality. AI-based methods, along with advantages such as high accuracy and the ability to learn and improve their performance, can slow down the overall speed of image processing due to enormous number of calculations and the need for large and diverse data sets in the training process. This point is more noticeable in the simulation of parallel processing methods with digital computers because in best cases, only one artificial neuron can be updated in each processor clock pulse. In this article, we will use an analytical method for medical image segmentation and finally, we will increase the speed of image segmentation by using a novel method to decrease computational complexity of calculation of image derivatives. Comparing the speed of the algorithm with the proposed method in this study with the conventional method of segmentation shows an acceptable increase in processing speed and reaching the final answer without losing segmentation accuracy
[1] A. Eltanboly, M. Ghazal, H. Hajjdiab, A. Shalaby, A. Switala, A. Mahmoud, P. Sahoo, M. El-Azab, A. El-Baz, “A level sets-based image segmentation approach using statistical shape priors”, Applied Mathematics and Computation, vol. 340, pp. 164–179, 2019.
[2] C. Huang, L. Zeng, “Level set evolution model for image segmentation based on variable exponent p-Laplace equation”, Applied Mathematical Modelling, vol. 40, pp. 7739–7750, 2016.
[3] A. Joshi, M. S. Khan, K. N. Choi, “Medical image segmentation using combined level set and saliency analysis”, IEEE Access, vol. 12, pp. 102016–102026, 2024.
[4] Z. Li, S. Tang, Y. Zeng, S. Chai, W. Ye, F. Yang, K. Huang, “A level-set method with a multiplicative–additive constraint model for image segmentation and bias correction”, Knowledge-Based Systems, vol. 297, 2024.
[5] S. Wali, C. Li, M. Imran, A. Shakoor, A. Basit, “Level-set evolution for medical image segmentation with alternating direction method of multipliers”, Signal Processing, vol. 211, 2023.
[6] S. Osher, J. A. Sethian, “Fronts propagating with curvature-dependent speed: Algorithms based on Hamilton-Jacobi formulations”, International Journal of Computational Physics, vol. 79, no. 1, pp. 12–49, 1988.
[7] A. Tsai, A. Yezzi, W. Wells, C. Tempany, A. Fan, W. A. Grimson, A. Wilsky, “A shape-based approach to the segmentation of medical imagery using level sets”, IEEE Transactions on Medical Imaging, vol. 22, no. 2, pp. 137–154, Feb. 2003.
[8] M. Leventon, E. Grimson, O. Faugeras, “Statistical shape influence in geodesic active contours”, in Proc. IEEE Computer Vision and Pattern Recognition, Hilton Head, SC, USA, pp. 316–322, Jun. 2000.
[9] T. Chan, L. Vese, “Active contours without edges”, IEEE Transactions on Image Processing, vol. 10, no. 2, pp. 266–277, Feb. 2001.
[10] A. Yezzi, A. Tsai, A. Willsky, “A statistical approach to snakes for bimodal and trimodal imagery”, in Proc. Int. Conf. Computer Vision, Corfu, Greece, vol. 2, pp. 898–903, 1999.