Skounakis Emmanouil, Sakkalis Vangelis, Marias Konstantinos, Banitsas Konstantinos, Graf Norbert
Institute of Computer Science at FORTH, Vassilika Vouton, Heraklion, Crete, Greece.
Annu Int Conf IEEE Eng Med Biol Soc. 2009;2009:3759-62. doi: 10.1109/IEMBS.2009.5334479.
This paper describes a flexible and easy-to-use annotation platform (GUI) for quick and precise identification and delineation of tumors in medical images. The design of the platform is clinically driven in order to ensure that the clinician can efficiently and intuitively annotate large number of 3D tomographic datasets. Both manual and well-known semiautomatic segmentation techniques are available in the platform allowing clinician to annotate multiple regions of interest at the same session. Additionally, it includes contour drawing, refinement and labeling tools that can effectively assist in the delineation of tumors. Furthermore, segmented tumor regions can be annotated, labeled, deleted, added and redefined. The platform has been tested over several MRI datasets to assess usability, extensibility and robustness with promising results.
本文描述了一个灵活且易于使用的注释平台（图形用户界面），用于在医学图像中快速精确地识别和勾勒肿瘤。该平台的设计以临床需求为导向，以确保临床医生能够高效、直观地注释大量三维断层数据集。平台提供了手动和知名的半自动分割技术，使临床医生能够在同一会话中注释多个感兴趣区域。此外，它还包括轮廓绘制、细化和标记工具，可有效辅助肿瘤的勾勒。此外，分割出的肿瘤区域可以进行注释、标记、删除、添加和重新定义。该平台已在多个MRI数据集上进行测试，以评估其可用性、可扩展性和稳健性，结果令人满意。
