CAJ | 학술논문

在多媒体图像数据库构建中，需要设计一种有效的多媒体图像资源递阶索引调度算法，提高对图像的实时处理和图像数据挖掘能力，传统方法中采用簇内图像数据寻优算法进行多媒体资源调度，性能受到了图像分布间隔的非均匀性的限制，图像索引调度性能不好。提出一种多媒体图像数据库资源递阶控制的图像索引算法，提高多媒体资源调度性能。设计多媒体图像空间位置放置算法，采用STARMA模型对多媒体图像资源进行高斯核平滑处理，得到STARMA模型展开后的图像特征模型，定义任务相邻边，计算新到达的任务与已放置任务相邻的可重构单元边，得到多媒体资源递阶索引的任务相邻边，进行图像预处理，实现多媒体图像空间位置放置。对每一个元独立地服从一个高斯随机数相位生成像点，构建最大空闲矩形矩阵，存放以当前单元为基点的最大空闲矩形包含的单元个数，得到多媒体资源的递阶控制模型，实现算法改进。仿真结果表明，算法能提高多媒体图像资源的索引精度，调度性能较好，图像调度稳健性提高。
재다매체도상수거고구건중，수요설계일충유효적다매체도상자원체계색인조도산법，제고대도상적실시처리화도상수거알굴능력，전통방법중채용족내도상수거심우산법진행다매체자원조도，성능수도료도상분포간격적비균균성적한제，도상색인조도성능불호。제출일충다매체도상수거고자원체계공제적도상색인산법，제고다매체자원조도성능。설계다매체도상공간위치방치산법，채용STARMA모형대다매체도상자원진행고사핵평활처리，득도STARMA모형전개후적도상특정모형，정의임무상린변，계산신도체적임무여이방치임무상린적가중구단원변，득도다매체자원체계색인적임무상린변，진행도상예처리，실현다매체도상공간위치방치。대매일개원독입지복종일개고사수궤수상위생성상점，구건최대공한구형구진，존방이당전단원위기점적최대공한구형포함적단원개수，득도다매체자원적체계공제모형，실현산법개진。방진결과표명，산법능제고다매체도상자원적색인정도，조도성능교호，도상조도은건성제고。
In the multimedia image database construction, the need to design a multimedia image resources hierarchical in?dex effective scheduling algorithm, improve the real-time processing and image data on image mining capacity. A multime?dia image database resources hierarchical control algorithm is proposed to improve the image indexing, multimedia re?source scheduling performance. Placement algorithm design of multimedia image space position, using the STARMA model of multimedia image resources for the Gauss kernel smoothing, image features of STARMA model after unfolded, tasks de?fined adjacent edges, calculate the new arrival task and the task placed adjacent to the reconfigurable cell edge, get the mul?timedia resources hierarchical indexing task adjacent edges, image pretreatment, realize placing multimedia image space position. On the formation of image point each element independently subject to a Gauss random number phase, construct?ing a maximal empty rectangle matrix, stored in the current unit as the unit number maximal empty rectangle point contains the hierarchical control model of multimedia resources. The improved algorithm, the simulation results show that the algo?rithm can improve the accuracy of image indexing multimedia resources, better scheduling performance, improve the robust?ness of image scheduling.