Research Of CT Image Compression In Remote Medical Transmission System

With the development of digital medical technology, medical imaging is playing a more important role in the medical diagnostic process. And the medical imaging resources are fully utilized in intelligent medical system. However, it is a challenge to transmit the huge amount of medical imaging data under the bandwidth-limited network, which affects the quality seriously. Therefore, the research of medical image compression technology is very significant under bandwidth-limited conditions.Firstly, in order to solve the problems occurred in the current remote medical image transmission system, a new remote medical image transmission system is designed in this thesis, which can reduce the duration of transmission and improve the work efficiency. The theory of image compression and ROI coding related to the system is deeply researched, combined with the CT image characteristics. Then a multiple ROI coding scheme under the condition of the low bandwidth is proposed. The simulation result shows that the scheme is feasible and accurate.For medical images, the background area will have a certain impact on the accuracy of doctor’s diagnosis, so it needs a certain quality of recovery in many cases, In the past, the algorithm of the balance between ROI region and background region is not efficient. In order to solve this problem, this thesis builds a neural networks model to predict PSNR by training a large number of kidney CT images. After that, a new algorithm is proposed based on the prediction of PSNR. This algorithm has lower complexity and can meet the design requirements through the simulation compared with other algorithms. So this algorithm has a practical value and can apply to medical systems.