计算机科学
帧速率
帧(网络)
延迟(音频)
低延迟(资本市场)
简单(哲学)
转换为二维
计算机视觉
计算机图形学(图像)
人工智能
高分辨率
立体显示器
算法
图像(数学)
电信
遥感
认识论
计算机网络
地质学
哲学
作者
Kunio Yamada,Yasunari Suzuki
标识
DOI:10.1109/isce.2009.5156848
摘要
This paper describes an algorithm for creating pseudo stereo image generation (2D-to-3D conversion) and its real-time hardware implementation. The proposed method is developed to fulfill the growing need for high quality generic 3D contents. The algorithm intends to approximate 3D structures of natural scenes and to generate visually pleasing stereo images. The depth data is generated with simple models and by warm/cool color theories. The algorithm has advantages in picture quality, low numerical complexity, and no need of tuning of parameters for different types of pictures. The hardware implemented with the simple algorithm does not need any large memory such as frame memory and gives very low latency. This technology had attracted many people's attentions at demonstrations including CEATEC 2008.
科研通智能强力驱动
Strongly Powered by AbleSci AI