PCA

说明：  连续隐马尔可夫模型cdhmm,多用于语音识别中，在文本相关的语音识别中，识别效果很好。(cdhmm)

Matlab Code for: Removing the Blocking Artifacts of Block-Based DCT Compressed Images in IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 12, NO. 7, JULY 2003

描述了matlab基础功能和基础代码的编写，对信号在时域和频域的特点进行分析，并介绍了几种采样方法(Describe the basic function and matlab code to write, on the characteristics of signal in time domain and frequency domain analysis, and sampling method are introduced)

说明：  做实验的一些心得体会，愿大家有所受益很好，欢迎大家学习，交流！(verygood！let we see！verygood！let we see！verygood！let we see！verygood！let we see！verygood！let we see！)

计算MSE和PSNR的小程序，自己写的，比较简单实用(MSE and PSNR calculation of small programs, to write their own, relatively simple and practical)

input file of ode 45

功率谱估计是利用有限长的数据估计信号的功率谱，广泛应用于各个领域。功率谱估计主要分为经典谱估计与现代谱估计。常用的经典谱估计方法有周期图法，相关法，周期图的改进法，常用的现代谱估计方法有最大熵谱估计，AR模型，MA模型，ARMA模型。经典谱估计适用于长序列的信号，其主要缺陷是描述功率谱波动的数字特征方差性能差，频率分辨率低，现代谱估计适用于短序列的信号，旨在改善谱估计的分辨率，并将其应用于实际地震资料的谱分析。 (Power spectrum estimation is the use of a finite length data to estimate the power spectrum of the signal, widely used in various fields. Power spectrum estimation is divided into classic and modern spectral estimation spectral estimation. Commonly used methods of classical periodogram spectrum estimation method, the relevant law, the cycle graph Improvement Act, commonly used in modern spectral estimation methods have maximum entropy spectral estimation, AR model, MA model, ARMA model. Classical spectral estimation applied to signal a long sequence, its main drawback is the fluctuation power spectrum describing digital signature variance performance is poor, the frequency resolution is low, modern spectral estimation applied to signal short sequences designed to improve spectral estimation resolution, and The analysis applied to real seismic data spectrum.)

采用脉冲响应不变法和双线性变换法设计IIR数字滤波器。 (Using impulse response method and bilinear transform IIR digital filter design. )

说明：  完成从信源到目的端信息的传输与接收，信源信息经过QPSK调制，然后经过瑞利信道，在目的端仿真其译比特率（BER）。(BER simulation.)

一个盲去模糊算法，速度很快！用到了l1/l2范数和对卷积核的l1范数，对噪声有抑制作用！(A blind go fuzzy algorithm, the sppeed is very fast! l1/l2 norm and l1 norm are used for compution, the noise was inhibited!)