3D Image And Video Coding Based On Feature Matching

The3D video (3DV) can provide the audience a more vivid word and real perception, thus it attracts widely concerns. It is widely used in3DTV, video conference, remote video monitoring, education, medical treatment, military and many other areas. Compared to the traditional two-dimensional image/video, the data of3DV is tremendous, thus, the compression of the original data is a quite challenging task. Therefore, efficient coding technology is very important. Currently, a special team on3DV coding was established by the International Coding organization to study on3D image/video coding. Meanwhile, a lot of references also focus on this research area. In this context, this paper mainly concerns about the3D video coding, and contributes the following achievement:1. According to the characteristics of3D image/video in the correlation, we propose a coding idea. Calculate the transformation matrix between views by feature points and then the transformation matrix can be applied for the three-dimensional image/video coding. In the experiment, we use different feature algorithms to extract the feature points and then compare the final quality of the reconstructed image to determine the best feature extraction algorithm and transform matrix estimation algorithm.2. In this paper, a novel scheme of two-stage3D image quality classification coding is proposed based on SIFT feature. In the first stage coding, we make full use of SIFT and RANSAC algorithm to obtain the correlation matrix between images of different views and the reference image for3D image coding. In the second stage coding, by using the residual information, the quality of the decoded image can be improved. The experimental results have shown that at high compression ratio, the proposed scheme can obtain better rate-distortion performance than H.264intra coding. Furthermore, with the change of the compression ratio, the proposed scheme can achieve more stable image reconstruction quality. For the image, human visual system (HVS, Human Visual System) pay more attention to the salient regions, while less attention to other areas. Therefore, this paper proposes a coding scheme based on the saliency, in which H.264intra coding is applied for salient regions of the image to get a high quality; while the other regions, a coding algorithm based on feature is adopted to obtain high compression ratio.3. In this paper, we propose a method of feature matching based asymmetric three-dimensional (3D) image coding with hierarchical reconstruction quality. At the encoder, for the main view the standard intra coding can be applied to obtain high reconstruction quality while for its neighboring views extracted feature descriptors can be utilized to calculate transformation matrix between views. The parameters of transformation matrix can be transmitted by very low bit rate and achieve the preliminary reconstruction. Furthermore, the residues can be exploited to improve the performance. The experimental results have shown that the proposed scheme can reach a very high compression ratio.