Abstract: The segmentation of a specific object in a single frame image has been faced with the problem of low segmentation accuracy due to background complexity and illumination variation. In this paper, a shape prior local binary fitting (LBF) based on contour pre-positioning was proposed for segmentation of human upper-limb images. Firstly, the upper-limb contour template was selected and pre-positioned by a kind of shallow convolutional neural network, and the coarse contour was obtained. Then, the LBF algorithm based on a prior shape was used to evolve the coarse contour, and the precise contour was obtained. Experimental results show that the success rate of the algorithm is over 90%, which shows that the method has good effect on the segmentation of a specific object in a single frame image faced with background complexity and illumination variation.