论文部分内容阅读
医学X线图像受到探测器面积大小的限制,成像范围有限,对较大器官的扫描无法一次完成。在观察病变部位时,医生需要结合多幅图像来进行诊断或治疗,因此需要对多张影像进行拼接处理。作为图像拼接技术的核心,图像配准技术已被广泛应用于医学成像中,将那些从扫描中获得的多类型信息进行配准从而得到更详细的信息。首先,本文重点综述了目前面向X线图像的比较主流和新兴的配准技术,如基于互信息的配准法,基于特征的配准法和基于变换域的配准法。其次,指出了X线图像配准中存在的影像漂移问题、拍摄角度的限制、非刚性配准仍未成熟、没有绝对的配准评价标准等问题。最后,总结了基于FPGA等硬件的医学图像配准、采用超分辨率重建技术以获取更高质量的待配准图像从而提高图像配准的精度和速度等发展趋势与研究前景。
Medical X-ray images are limited by the size of the detector area, the imaging range is limited, the scan of larger organs can not be completed in one time. When looking at the lesion, doctors need to combine multiple images for diagnosis or treatment, so the need for multiple images stitching. As the core of image mosaic technology, image registration technology has been widely used in medical imaging, which will get more detailed information by registering many types of information obtained from scanning. First of all, this article focuses on the current mainstream and emerging registration technologies for X-ray images, such as registration based on mutual information, feature-based registration and transform domain based registration. Secondly, it points out the problems of image drift, limitation of shooting angle, non-rigid registration still not mature, and no absolute standard of registration evaluation in X-ray image registration. Finally, the development trends and research prospects of medical image registration based on hardware such as FPGA are summarized, and the super-resolution reconstruction technology is used to obtain higher quality image to be registered to improve the accuracy and speed of image registration.