<![CDATA[The development of network technology and multimedia technology has promoted the development of the integration, digitalization and individualization of communication technology, and has realized multiple functions based on wireless channel network, including voice, data and image. On the basis of classification by Support Vector Machine (SVM), this paper has proposed a low bit-rate coding method for segmenting the Regions of Interest (ROI) of objects through reasonably selecting initial class center and dynamically adjusting the number of categories, and designs its 3-dimension model method. In low bit-rate coding, the ROI is coded first using SPIHT algorithm, and then the background or the regions insensitive are coded using H.264 code, analyzing the experimental data and designing the structure, we can get the relation among related parameters under conditions of different permanent-magnet thickness and air gaps. It is proved that this method can achieve a better recognition effect, greatly improve the coding efficiency and satisfy the visual effects of human eyes, which can provide criteria in theory and application for designing on new driving principle, control system. DOI: http://dx.doi.org/10.11591/telkomnika.v11i10.3463 ]]>
