Chinese Journal of Lasers, Volume. 49, Issue 4, 0409002(2022)
An Image Encoding Method for Fusion Display of Virtual and Real Scenes in Holographic Stereogram
To achieve the fusion display effect of real and virtual scenes in a holographic stereogram, an image coding method based on instance segmentation and depth value determination is proposed. A theoretical analysis and the experimental results show that the proposed image coding method can effectively add some virtual 3D elements in the real scene to enhance the visual experience. The combination of virtual and real scenes fully considers the occlusion relationship, which is not a simple scene superposition. The key to fully considering the occlusion relationship is to assign and determine the pseudo depth after scene layering. There is still a significant gap between the virtual reality fusion method discussed in this paper and the latest work in the AR field. However, the proposed method provides a basic idea for further research into holographic stereogram AR displays. For example, by continuously improving the efficiency of instance segmentation, the scene is more accurately layered. The accurate depth calculation method is studied, and the accurate depth value is used as the basic processing data, which can better show the effect of virtual real scene fusion display in a holographic stereogram. This paper only considers the simple and small scenes used in an experiment and tends to assume ideal conditions. In theory, the proposed method is suitable for scenes with a clear occlusion relationship between different instances in the sampled image; however, it cannot effectively deal with the occlusion between the same instances. Therefore, further analysis of complex scenes is required.
Get Citation
Copy Citation Text
Yunpeng Liu, Xi Wang, Xinlei Liu, Tao Jing, Xiaoyu Jiang, Xingpeng Yan. An Image Encoding Method for Fusion Display of Virtual and Real Scenes in Holographic Stereogram[J]. Chinese Journal of Lasers, 2022, 49(4): 0409002
Category: holography and information processing
Received: Jun. 9, 2021
Accepted: Jul. 5, 2021
Published Online: Jan. 18, 2022
The Author Email: Yan Xingpeng (yanxp02@gmail.com)