Tomographic processing method with a small number of projections of a contrasted object

A method for processing a sequence of a plurality of 1D or 2D projection images of an object of interest, wherein the projection images of the object of interest are acquired by a medical imaging system comprising a source of rays, the source being configured to move around the object in order to acquire the plurality of 1D or 2D projection images of the object of interest according to a plurality of angulations, the number of 1D or 2D projections and/or the angular coverage leading to the occurrence of specific sub-sampling artefacts, the method being recursive and comprising:

• defining, a priori, a sparse image and a series of models for breaking down the object of interest as a sum of a sparse component and of a complementary non-sparse component;
• initializing a sparse 2D or 3D image depending on the sparse image defined a priori and initializing the series of models for breaking down the object of interest;
• reconstructing a 2D or 3D image of the sparse component of the decomposition model of the object of interest from acquired 1D or 2D projection images and from the initialized sparse 2D or 3D image;
• and
• updating the sparse 2D or 3D image as being equal to the reconstruction of the 2D or 3D image of the sparse component and updating the decomposition model of the object of interest into a new sparse component and a new complementary non-sparse component, and returning to reconstruction of the 2D or 3D image of the sparse component so that, during the iterations, the reconstruction of the 2D or 3D image of the sparse component gradually reintroduces the complementary component into the sparse 2D or 3D image, in order to obtain a complete 2D or 3D image of the non-sparse object of interest.