An automated 3D annotation method for breast ultrasound imaging

Image spatial annotation is one of the most important steps in breast ultrasound scanning examinations because follow-up evaluation, diagnosis and treatment may be performed based on the annotation. The conventional annotation method is manual and highly dependent on the operator's experience. Therefore, it is time-consuming and prone to errors. This paper introduced a novel annotation method for breast ultrasound imaging. A spatial sensor was attached to the ultrasound transducer to obtain the real-time image location information. 3D virtual models of breasts and probe were shown on screen together with ultrasound images to display the spatial data intuitively. A program was developed to acquire ultrasound images and spatial signals, compute the image locations relative to the breast, and display images together with their corresponding 3D annotations. It was demonstrated that this novel annotation method could provide automated and accurate annotations in phantom experiments. This method also showed its potential to continuously annotate breast ultrasound images in relative short time in the preliminary in-vivo experiments.Department of Health Technology and InformaticsRefereed conference pape