This paper describes a totally new way to process Xray data in order to reconstruct the external and internal boudaries of objects, which do not involve Computed Tomography (CT). We show that the segmentation can be performed directrly with the raw data, the sinogram produced with the scanner, and that those segmented shapes can be geomtetrically transformed into reconstructed shapes in the usual space. Thus, if we are interested in only the boundaries of the objects, our method eliminates the computationally expensive step of Computed Tomography. Experimental results are presented for both synthetic and real data, leading to subpixel positioning fo the reconstructed boundaries. Our method gives its best results for sparse, high contrasted objects such as bones or blood vessels in angiograms. It can be adapted to any kind of scanner, including 3D scabbers. At last, we present an extension of our method which allows "on the fly" processing of the data and real time tracking of the objects boudnaries.