一种应用于实时FRC中的淡入淡出检测算法

以硬件代价优化为目的,提出并实现了一种基于图像像素差法的淡入淡出检测方法,并通过运动补偿及综合考虑淡入、淡出线性特性等方法予以优化。通过实验证明,该方法相比于基于灰度直方图特征的视频淡入淡出检测算法准确率有所提高,而硬件面积和功耗在不同的综合时钟频率下则分别节省84%和72%以上。凭借其良好的硬件性能,该方法及结构非常适合应用于帧率提升IP核的设计中。

A Fade-in and Fade-out Detection Method for real-time FRC

In order to lower cost in hardware implementation, this paper proposes and implements a fade-in and fade-out detecting method based on image pixel difference method and optimized it with motion compensation and the general consideration of linearity in fade-in and fade-out scenes. Experiment results show that compared with the fade-in and fade-out algorithm based on histogram features, the proposed method performs better in accuracy and consumes 84% and 72% less in term of chip area and power respectively. With its excellent hardware performance, the proposed method and architecture are very suitable for use in the design of Frame Rate Up-Conversion IP core.