人体动作数据集的大容量快速存储算法仿真

Simulation of Large Capacity Fast Storage Algorithm for Human Motion Data Set

传统存储方法容易忽略大容量人体动作数据集的冲突问题,造成节点失效,且存储质量不高、存储容量低,提出一种人体动作数据集的大容量快速存储算法。采用链地址法建立哈希表,通过扩展哈希编码的方式扩展存储节点,在扩展节点条件下,通过人体动作数据集存储使用强度与节点传输概率完成人体动作数据节点传输匹配。根据匹配结果,将节点划分为不同的存储级别,按照人体动作数据存储系统中的硬件存储容量以及数据集具体情况,将不同级别存储节点分割成若干阈值级别,避免节点失效,完成大容量人体动作数据集存储。实验结果表明,所提方法存储速度快,整体性能强,在存储人体数据应用中具有较高可用性。

The traditional storage methods are easy to ignore the conflict problem of large capacity human action data set, resulting in node failure, and the storage quality is not high and the storage capacity is low. A large capacity fast storage algorithm for human action data set is proposed. The hash table is established by chain address method, and the storage node is extended by extending hash coding. Under the condition of extended node, the transmission matching of human action data node is completed through the storage intensity of human action data set and the transmission probability of node. According to the matching results, the nodes are divided into different storage levels. According to the hardware storage capacity in the human action data storage system and the specific situation of the data set, the nodes will be different. The level storage node is divided into several threshold levels to avoid node failure and complete the storage of large capacity human action data set. The Simulation results show that the proposed method has fast storage speed and strong overall performance. It is feasible to store human body data.