Key Technique Study Of Computer Aided Ultrasound Follicular Volume Measurement Based On The GVF Snake Model

The volume of the ovarian follicle is an important indicator of oocyte maturity during the assisted reproductive treatment (ART). The quantitative monitoring and analysis of ovarian follicles can help the obstetricians know the formation of mature oocytes in advance and thus, improve the success rate of the assisted reproduction. In clinical, the monitoring of all stimulated ovulatory is often performed by using the three-dimensional ultrasound technology to measure the follicle volume. However, the most frequently used technologies of the follicle volume measurement (i.e., the VOCAL and SonoAVC, developed by the GE Healthcare) suffer from the speckle noise, weak edges, and the ringing effect in the ultrasound image and produce instable results due to their ineffective contour extraction algorithms.This thesis presents an automated follicle volume measurement method based on the active contour model, which needs to get all the tomographic data of the target follicle firstly. The core content of the proposed method is a novel contour extraction algorithm, which will be used to extract contours of all the tomographic data and subsequently calculate the follicle volume using a three-dimensional integral formula in the mathematic. The focus of this thesis includes:1. An unsupervised edge detection method based on the traditional Canny and multiplicative gradient operator was proposed, which had good suppression ability to the speckle noise.2. A novel GVF Snake model was proposed for the ultrasound image, containing two aspects. On the one hand, to improve the convergence accuracy and speed, an adaptive pressure was added to the external force of the generalized GVF Snake. On the other hand, the edge map was obtained by the proposed edge detection method.3. A volume measurement system was realized by the Matlab software, based on the proposed contour extraction method and the three-dimensional integral formula.4. The simulation and clinical experiments were designed to verify the feasibility and stability of the system. Statistical analysis tells us that the Pearson correlation coefficients of the proposed method with SonoAVC and VOCAL are all0.99, and the self-correlation coefficient of the proposed method is0.999, which is higher than SonoAVC and VOCAL, indicating that the proposed method has the great feasibility and highest stability to follicle monitoring.