Segmentation and registration

Collaboration with European excellences has led to the development of an automatic segmentation method that uses information extracted simultaneously from two different imaging modalities (CT and US).

Tecnologie Avanzate is actively working on the development of algorithms and tools for the deformable registration of multimodal 3D images, in order to integrate these algorithms into a flow of evaluation and verification of the quality of RT treatment plans.

Two master thesis projects of the Politecnico of Turin and the University of Trieste led to the development of a CAD for the identification and contouring of breast cancer lesions in magnetic resonance images. The algorithm integrates contour validation metrics and geometric features combined with Texture Analysis parameters.

TA supported a multicenter study for the evaluation of the robustness and reliability of a segmentation algorithm on PET images, conceived and developed by a group of excellence in the nuclear medicine sector.

The constant interest and commitment of Tecnologie Avanzate on projects dedicated to breast imaging has allowed us to develop a library that allows to view tomosynthesis mammographic images with optimized modalities and provides tools for the automatic synchronization of positions between different breast images, in order to facilitate simultaneous navigation.

We are working on a deep learning system for the automatic segmentation of 3D lesions on PET images.