Extracting Corresponding Point Based on Texture Synthesis for Nearly Flat Textureless Object Surface

Figure 3

The texture adding mechanism using human model. First row, from left to right: (a) detecting the largest region inside the textureless object with different sizes of slide windows. (b) Getting points overlap the slide window. (c) Getting the corresponding point. The second row, from left to right: (a), (b) computing centre position of the textureless object on each image and estimating the affine transform between two images. (c) Adding synthetic texture to the object surface in two images.

(a)

(b)

(c)