Computer assisted diagnosis using ultrasonography

本研究旨在探讨增强超声影像的方法，以协助医学专业人士解读其影像。图像增强算法的设计将使不同组织之间的差异更加明显，并增加正常组织和疾病组织之间的外观差异。这种增强后的图像被称为虚拟组织学图像，源于传统的组织学方法，即在显微镜下染色和观察组织，病理医生使用这种方法来进行诊断。“虚拟组织学”图像将为医学专业人员提供一种新的信息，将提高诊断的准确性和快速性。有可能使用“虚拟组织学”图像来减少手术活检以确定恶性程度的需要。在某些情况下，侵入性的活组织检查可以被无害的超声扫描所取代，这种扫描是通过“虚拟组织学”算法增强的。在未来，“虚拟组织学”图像可以在检查时由超声波机自动生成，从而为患者提供更快的诊断时间，并允许低成本的早期病理检测。

The proposed research is to investigate ways of enhancing ultrasound images so as to aid their interpretation by a medical professional. Image enhancement algorithms will be designed to make differences between different kinds of tissues more visible, and to increase the difference in appearance between normal tissues and diseased tissues. The enhanced images are called "virtual histology" images after the classical histology methods of staining and viewing tissues under a microscope which are used by pathologists to make diagnoses. "Virtual histology" images will offer a new kind of information to medical professionals which will improve the accuracy and rapidity of diagnoses. It may be possible to use "virtual histology" images to reduce the need for surgical biopsies to determine malignancy. Invasive biopsy procedures could be replaced in some cases by a harmless ultrasound scan which is enhanced by "virtual histology" algorithms. In the future, "virtual histology" images could be generated automatically by ultrasound machines at the time of examination leading to faster time to diagnosis for patients and allowing for low-cost early detection of pathologies.