Separate modal analysis for tumor detection with DIET : traineeship report

Digital Image Elasto Tomography (DIET) is a non-invasive breast cancer screening modality that induces mechanical vibrations into a breast and images its surface motion with digital cameras. A new approach in software based diagnosis of this surface motion is presented, focussing on the second natural frequency of the breast. Separate modal analysis is used to estimate the modal parameters using imaging data from silicone phantoms. The proposed method is verified on a dataset consisting of 13 silicone phantoms with tumor sizes of 0, 5, 10 and 20 mm in diameter. The second natural frequency proves to be a reliable metric with the potential to clearly distinguish cancerous and healthy tissue as well as providing an approximate location for the tumor for tumors of 10 mm in diameter or larger (p-value less than 0.01). The proposed method shows promise for real-time, non-invasive breast cancer screening