Medical Image Compression

点击翻译按钮获取中文摘要

Prompt access to clinical information is crucial for improved healthcare decisions. The Electronic Health Record (EHR) promises better care by providing centralized access to the different sources of data that make up the patient's medical record. EHR is expected to help achieve a more efficient health-care system by enhancing productivity through timely access to information; it is also expected to help achieve a more effective system by reducing the duplication of tests. Medical images are very important information in the patient's health record. With Canada Health Infoway's investment in diagnostic imaging projects, medical images are centrally archived for easier long-term management. Images are also being shared and accessed by a very large number of healthcare professionals: by clinicians as part of a patient's follow-up and treatment; by radiologists to provide second opinions or first opinions when no radiologists are on site for trauma or remote-area cases; or by specialized physicians, such as surgeons or orthopedists, in planning surgical operations or procedures. The size of medical images is continuously growing. Computed Tomography (CT), for example, can generate a series of thousands of images whose size exceeds the gigabyte. Transferring large image sets on a communication network requires an enormous bandwidth, resulting in an inefficient and inadequate system. Images can be compressed with no information loss, reducing their size by 30%. Further compressing is needed, but currently available methods introduce artifacts and distortions that may impact diagnostic accuracy. In this research program, we propose to quantitatively assess the impact of compression methods on the diagnostic value of medical images. We also propose to develop novel compression schemes specially tailored to preserve that diagnostic value. Therefore, we expect to reduce the bandwidth needed to transfer medical images while preserving their diagnostic value.