离散余弦变换
计算机科学
图像压缩
记忆电阻器
数据压缩
电子工程
压缩(物理)
电路设计
点(几何)
算法
图像(数学)
人工智能
图像处理
数学
工程类
嵌入式系统
材料科学
几何学
复合材料
作者
Qinghui Hong,Bang He,Zedi Zhang,Pingdan Xiao,Sichun Du,Jiliang Zhang
标识
DOI:10.1109/jetcas.2023.3243569
摘要
Discrete Cosine Transformation(DCT) is an important method in image compression. With the advent of the era of big data, how to make efficient DCT calculation is the key to its application. Therefore, this paper proposes a one-step DCT calculation method from the point of view of analog circuit, which can greatly speed up the operation in the form of low energy consumption. Firstly, a new weight programming is proposed, and the negative weight problem is solved. On the basis of this analysis, One-dimensional DCT(1D-DCT) and One-dimensional IDCT(1D-IDCT) calculation circuit is proposed. After that, the One-dimensional circuit is designed into Two-dimensional circuit by using parallel processing. PSPICE simulation results show that the average calculation accuracy of the circuit is above 99%. Finally, the designed two-dimensional circuit is used in image compression, and it is found that the compression effect is significant in the case of high speed.
科研通智能强力驱动
Strongly Powered by AbleSci AI