RGB-D Depth Inpainting with Color Guide Inverse Distance Weight
Yossawee Keaomanee, Prakarnkiat Youngkong · 2019
This paper proposed a method to increase the precision of the depth inpainting method by using depth and color information from the RGB-D system in the same image frame. The procedure consists of 2 steps. The first part is color similarity checking, this part tries to select the same color in a window to collected neighbor depth information. The second step is weight calculation, for this step, The color weights of all selection points in a window are calculated and normalized. The predicted value of depth in a missing pixel is calculated by the relation of all color weight and their depth data. The experimental results showed that the proposed method improved the accuracy of depth inpainting.