This paper presents a local learning framework in which the local classifiers can be pre-learned and the support size of each classifier can be selected to minimize the error bound. The proposed algorithm is compared with the conventional support vector machine (SVM). Experimental results show that our scheme using the user-defined parameters C and σ is more accurate and less sensitive than the conventional SVM.

In this paper, we present a design of MPEG-4 video codec chip to reduce the power consumption using frame level clock gating, macro block level and motion estimation skip scheme. It performs 30 frames/s of codec (encoding and decoding) mode with quarter-common intermediate format (QCIF) at 27 MHz. Power consumption is 290 mW at 27 MHz operation, which is achieving 35% power saving compared to a conventional CMOS. Motion Estimation skip method is employed to reduce 32% computation load. This chip performs MPEG-4 Simple Profile Level 2 (Simple@L2) and H. 263 base mode. Its contains 388,885 gates, 662 k bits memory, and the chip size was 9.7 mm9.7 mm which was fabricated using 0.35 micron 3-layers metal CMOS technology.

In this letter, we present a novel timing offset estimation method for chirp-based communication systems which is robust against frequency offset. For robust timing offset estimation, we propose a partial cross-correlation and differential multiplication method using up and down chirp symbols. The performances of the proposed estimator in indoor multipath channel model provided by IEEE 802.15.4a standard are presented in terms of mean-square error (MSE) obtained by computer simulation. The simulation results show that the proposed estimator has a significantly smaller MSE than the conventional estimators.