Classification of benign/malignant microcalcification clusters is a major diagnostic challenge for radiologists. Clinical studies have revealed that the shape of the cluster, and the spatial distribution of individual microcalcifications within it, are important indicators of its malignancy. However, mammographic images of clustered microcalcifications confound their three-dimensional (3-D) distribution with image projection and breast compression. This paper presents a novel model-based method for reconstructing microcalcification clusters in 3-D from two mammographic views (cranio-caudal and medio-lateral oblique--"shoulder to the opposite hip" or lateral-medio). We develop a 3-D breast representation and a parameterised breast compression model which constraints geometrically the possible 3-D positions of a calcification in a two-dimensional image. Corresponding calcifications in the two views are matched using an estimate of the calcification volume. Both the geometric constraint and the matching criterion are utilized in the final reconstruction step to build the 3-D reconstructed clusters. Validation experiments are described using 30 clusters to verify the individual steps of the model, and results consistent with known ground truth are obtained. Some of the approximations in the model and future work are discussed in the concluding section.

No abstract

Abstract. A quick method to obtain the 3D transformation of a 3D freeform shape model from its 2D projection data is proposed. This method has been developed for the real-time registration of a 3D model of a cerebral vessel tree, obtained from pre-operative data (eg. MR Angiogram), to a X-ray image of the vessel (eg.Digital Subtraction Angiogram) taken during an operation. First, the skeleton of the vessel in a 2D image is automatically extracted in a model-based way using a 2D projection of a 3D model skeleton at the initial state (up to +20 degree difference in rotation). Corresponding pairs of points on the 3D skeleton and points on the 2D skeleton are determined based on the 2D Euclidean distance between the projection of the model skeleton and the observed skeleton, In the process, an adaptive search region for each model point, which is determined according to the projected shape, effectively removes incorrect correspondences. Based on a good ratio of correct pairs, linearization of a rotation matrix can be used to rapidly calculate the 3D transformation of the model which produces the 2D observed projection. Experiments using real data show the practical usefulness of the method.

No abstract

No abstract

No abstract

No abstract

No abstract