节点文献
BP神经网络图像压缩算法乘累加单元的FPGA设计
FPGA Design of Multiply-accumulate Module in Image Compression Algorithm Based on BP Neural Network
【摘要】 提出一种基于三层前馈BP神经网络实现图像压缩算法的方案,该方案采用可重载IP核和VHDL代码相结合的设计方式。对方案中重要单元-乘累加单元进行了FPGA设计,该模块设计采用流水线处理方式,增大了数据吞吐量,减小了系统延时,提高了时钟频率,并完成了该单元的行为级功能仿真。仿真结果验证了FPGA设计的可行性。
【Abstract】 A realization scheme for image compression algorithm based on three layers Back Propagation(BP) neural network is presented,which adopts the method combining the universal IP core with the VHDL language.The FPGA design of multiply-accumulate module which is one of the important components is carried out,it is designed with pipeline,which could increase the data of processing,decrease the latency of system,raise the frequency of clock.The behavior of functional simulation for MAC is completed.The simulation results show that the design is practicable.
【基金】 国家重点实验室基金项目(9140C5305020706)
- 【文献出处】 现代电子技术 ,Modern Electronics Technique , 编辑部邮箱 ,2009年19期
- 【分类号】TP183;TN791
- 【被引频次】4
- 【下载频次】289