In this paper, an application specific processor architecture is proposed as an IDCT (Inverse Discrete Cosine Transform) engine for MPEG-1[1-3] video stream decoding. The engine executes an efficient implementation of...In this paper, an application specific processor architecture is proposed as an IDCT (Inverse Discrete Cosine Transform) engine for MPEG-1[1-3] video stream decoding. The engine executes an efficient implementation ofthe Feig algorithm. Performance evaluation concludes that the proposed architecture can adequately deal with real bineMPEG-1 IDCT requirement together with achievable cost reduction when compared with dedicated hardware approach[4]. In addition, it can be observed that the proposed architecture can also be utilized to deal with Other functionalities such as variable length MPEG-1 bitstream decoding, inverse one dimensional (1D) DCT (Discrete CosineTransform) audio decoding.展开更多
由于传感误差、传感噪声、传输错误等因素的影响,同一个传感区域内多个传感器节点的传感数据具有一定程度的差异,这种差异导致的区域不确定性传感数据给查询、预测等后续深层次的数据处理提出了严峻挑战.针对这类传感数据的预测问题,提...由于传感误差、传感噪声、传输错误等因素的影响,同一个传感区域内多个传感器节点的传感数据具有一定程度的差异,这种差异导致的区域不确定性传感数据给查询、预测等后续深层次的数据处理提出了严峻挑战.针对这类传感数据的预测问题,提出一种基于多变量主元分析(multiple variable principal component analysis,MVPCA)的不确定性传感数据预测方法.通过MVPCA的特征提取这一预处理手段获得不确定性传感数据的本质特征,然后采用基于相关分析的多元回归方法对这些数据进行建模和预测.实际传感数据的实验结果表明,该方法能有效解决不确定性传感数据的预测问题.展开更多
文摘In this paper, an application specific processor architecture is proposed as an IDCT (Inverse Discrete Cosine Transform) engine for MPEG-1[1-3] video stream decoding. The engine executes an efficient implementation ofthe Feig algorithm. Performance evaluation concludes that the proposed architecture can adequately deal with real bineMPEG-1 IDCT requirement together with achievable cost reduction when compared with dedicated hardware approach[4]. In addition, it can be observed that the proposed architecture can also be utilized to deal with Other functionalities such as variable length MPEG-1 bitstream decoding, inverse one dimensional (1D) DCT (Discrete CosineTransform) audio decoding.
文摘由于传感误差、传感噪声、传输错误等因素的影响,同一个传感区域内多个传感器节点的传感数据具有一定程度的差异,这种差异导致的区域不确定性传感数据给查询、预测等后续深层次的数据处理提出了严峻挑战.针对这类传感数据的预测问题,提出一种基于多变量主元分析(multiple variable principal component analysis,MVPCA)的不确定性传感数据预测方法.通过MVPCA的特征提取这一预处理手段获得不确定性传感数据的本质特征,然后采用基于相关分析的多元回归方法对这些数据进行建模和预测.实际传感数据的实验结果表明,该方法能有效解决不确定性传感数据的预测问题.
