Female Breast cancer is a serious metabolic disease that has one of the highest mortality rates worldwide. Research is being conducted into improving the accuracy of breast cancer detection, especially in women with dense breast tissues, by supplementing two different imaging modalities. However, the volumes for each modality are often acquired in different environments, thus making a surface reconstruction an essential requirement for performing an accurate fusion of the volumes.
In this study, we propose an algorithm for surface reconstruction of the breast surface from autonomously acquired Ultrasound (US) images to further enable fusion of a pre-operative Magnetic Resonance (MR) volume with an intra-operative US volume. The images are acquired with a 7 degree-of-freedom robotic manipulator that has a linear US probe attached to its end-effector. Several experiments are performed on a realistic breast phantom. The surface reconstruction of the breast was calculated by determining the transition points between the instances of contact and no contact between the breast surface. A minimal amount of contact between the surfaces with the US probe in horizontal orientation, as indicated in the results, has the highest accuracy with a mean error and standard deviation of 3.87 mm and 1.71 mm respectively.
To join the presentation via Microsoft Teams click here