Research On ROI Extraction And Compression Algorithms For Medical Pathological Images

With the advancement of science and the development of medical equipment,medical pathological image data is becoming larger and larger,and the storage and transmission of medical pathological images have put forward higher requirements for image coding.During the compression of medical pathological images,it is necessary to ensure the compression quality of the cell area and improve the compression efficiency of the overall image.If the bit rate allocation of the cell area is less,it will cause a large quality loss and affect the diagnosis result;if the bit rate allocation of the background area is more,it will increase the storage cost.The current core problem is how to obtain optimal coding performance under the conditions of limited bandwidth and allowed delay.The exploration and research on this issue is of great significance and application value.The current research direction is mainly in terms of time and space to reduce redundancy to improve coding performance.This topic mainly studies how to accurately control the coding quality of medical pathological images ROI and UROI,and minimize the spatial and temporal redundancy to maximize the coding rate.Aiming at this problem,this thesis proposes a medical pathology image coding method based on ROI.Generally,it can be divided into spatial coding redundancy improvement and temporal redundancy coding performance improvement.Enhancing performance coding in terms of spatial redundancy.As the image is compressed,more code rates are assigned to the region of interest(ROI)in the image,and less code rates can be assigned to the non-interest region(UROI).Improve coding efficiency.1)Taking into account the complexity of the medical pathological image and the attention of the human eye,the ROI region of the medical pathological image is extracted by a combination of edge detection and region growth.2)Under the X264 and X265 coding frameworks,consider the texture complexity and quantify the effect of step size on the quality of medical pathological image ROI and UROI coding,establish a medical pathological image ROI quality control model and UROI quality control model and use them Video encoding framework.Through the establishment of a good quality control model of ROI and UROI to control the coding quality of different regions of pathological images,and the experiment proves that the method can effectively control the coding quality of different regions in medical pathological images.Improve coding performance in terms of temporal redundancy.1)The buffer has been optimized to improve encoding efficiency.2)Based on the established quality control method,a multi-threaded parallel encoding method based on X264 is used to compress medical pathological images to develop an efficient parallel encoding and decoding system.The algorithm in this thesis can effectively detect the region of interest in the video and use thequality control model to control the encoding quality.The parallel video encoding and decoding system can effectively improve the encoding time of the image.