Abstract: A context-preserving volume rendering approach is proposed for accurate medical image visualization and analysis,to enhance the display of region of interest(ROI) and preserve the clear shape cues of background.By the approach,a specific iso-surface along the viewing ray is first located,and secondly for any sampling point,its gradient and its distance to the iso-surface are calculated.Then,an attenuation function is constructed using the distance and the gradient of the sampling point.Finally,the attenuation function is integrated into opacity transfer function to control the rendering.The proposed method is tested on the head,abdomen,and feet CT data of digital human.The experimental results show that the display of ROI is efficiently enhanced while the contours of the tissue are well preserved.