Generation of Multiview Images Using Stereo and Time-of-Flight Depth Cameras

We propose a multi-view image generation method using a low-resolution time-of-flight (TOF) depth camera and a stereoscopic camera. The TOF depth camera obtains the scene's range information in real-time using the emitted light signal. Then, it quantizes the measure range and provides depth images of the scene. However, the TOF depth camera has not only too small resolution but also noise and lens distortion. In order to utilize the TOF depth image with the high-resolution color images, we have to reduce noise and distortion, and also overcome the resolution difference. The proposed method generates disparity maps for a stereo color image using the TOF depth data. We reduce the noise and lens distortion in the TOF depth image and then warp the TOF depth values to the one color image position. The color image is segmented, and the warped TOF depth values are called the segmented regions as the disparity. After generating the disparity map for the other view, we obtain the multi-view image by synthesis of intermediate view images. Experimental results show the generated disparity maps and reconstructed virtual view images verify that the proposed method efficiently generate the disparity maps and the multi-view image. KeywordsDepth generation, Multi-view, ToF depth camera, 3DTV