Abstract: Removing the background in the cross-section slice images of a mouse is an important step for the reconstruction of a digital mouse model. According to the feature that the cross-section contour of a mouse in the successive slice images changes gradually for the most part, a method based on the template matching theory for removing the background is proposed in this study. More exactly, the newly detected edge of the mouse in a slice image is adopted as a template and referred to extract the contour in the adjacent slice image every time. Then the glitches and gaps in the extracted contour are removed by means of image processing such as the Lagrange interpolation to form a continuous contour. Finally, the refilled contour and the original image under go the logic operation to complete the background removal. The results indicate that this method performs well with the great advantages in reducing the manual workload and accelerates the process of 3D reconstruction of a mouse model.