Abstract: For the current workload of labeling image data for physicians in the process of radiomics,a compression method of multi-layer LVP image of single sequence based on value assessment is proposed.A filtering compression algorithm is used on the multi-layer LVP original image,the multi-layer LVP segmented image is subjected to a pooling compression algorithm.The compression algorithm is applied to the MRI glioma dataset to compare the labeling time of the original image before compression and the compressed single layer HVP image,and to test the effect of the compressed image on radiomics classification task.The experimental results show that the time to label a case after compression is significantly reduced,compared with that before compression,there is little difference between the result of AUC and that of the algorithm proposed in this paper.Thereby reduce the workload of data annotation significantly for physicians.