目前,大多数数据分析的结果就像是动画中的一帧:某一时间点上的快照。
Today, the result of most data analysis resembles a single frame from a motion picture: a snapshot of one point in time.
并且通过对关键问题的分析,逐步实现了图像头和条带解码,宏块解码、帧间预测等几个核心模块。
Through the analysis of key problems, we gradually finished several core modules like image head and straps decoding, macro block decoding, code streaming parsing and cross-frame forecast.
从理论上分析了帧内变换、预测编码算法,得出其优点,并通过实验验证了帧内变换、帧内预测、帧间预测的混合编码方法优越性。
Many newly technique and algorithms are emphatically analysed. Theoretical analysis educes the advantage of latter method, and experiment result validates the superiority of such coding method.
在对帧间差分算法进行分析的基础上,建立了基于帧间差分算法的目标探测识别系统,提出了针对该算法的脉冲光干扰方法。
The moving target detection method based on the three frame difference and background difference method was proposed, which is simply and fast to realize.
算法通过分析帧间的SURF特征点匹配数目随时间变化的性质来检测镜头边界。
The algorithm detects the shot boundary by analyzing the nature that the matching number of SURF feature points of frames change over time.
算法通过分析帧间的SURF特征点匹配数目随时间变化的性质来检测镜头边界。
The algorithm detects the shot boundary by analyzing the nature that the matching number of SURF feature points of frames change over time.
应用推荐