Automatic identification and extraction of bone contours from x-ray images is an essential first step task for further medical image analysis. In this paper we propose a 3D statistical model based framework for the proximal femur contour extraction from calibrated x-ray images. The automatic initialization is solved by an Estimation of Bayesian Network Algorithm to fit a multiple component geometrical model to the x-ray data. The contour extraction is accomplished by a non-rigid 2D/3D registration between a 3D statistical model and the x-ray images, in which bone contours are extracted by a graphical model based Bayesian inference. Preliminary experiments on clinical data sets verified its validity. [10] 3D statistical models are used for 2D segmentation and 3D reconstruction from calibrated 2D fluoroscopic images (location and orientation of the fluoroscopic source w.r.t. the image acquisition planes are known). Compared with 2D statistical modes, 3D statistical model usually only contains shape information but not the intensity information on the 2D images. In principle it can be used for segmenting an image taken from an arbitary view direction. 3D statistical model also need an initialization, which is usually manually defined [7][9]. Due to the dense mesh of the 3D statistical model [16], fully automated solutions based on evolutionary algorithm is computational expensive [17].
MotivationIn this paper we propose a 3D statistical model based fully automatic segmentation framework for calibrated fluoroscopic images. In our approach, the initialization is accomplished by an Estimation of Bayesian Network Algorithm on a simplified multiple component model instead of the triangulated surface mesh of the 3D model, which reduces the computational complexity. The statistical model based fine shape extraction is achieved by a Bayesian inference on a Bayesian network, which encodes the shape and texture information of the model and therefore enhances the robustness and accuracy of the contour extraction.
Related WorkBayesian network based approach [18][19][20] is used to identify or track object such as human body, which is composed with multiple subparts and among the subparts structral or kinematic constrains exist. The Bayesian network embeds the subparts constraints in a graphical model associated with image observations. Bayesian network is also exploited for finding deformable shapes