histeq

通过统计直方图并作相应的处理完成了红外图像非均匀性校正，特别是针对红外图像中常见的条纹非均匀性(Through statistical histogram and processed accordingly infrared image non-uniformity correction, especially for the common striped infrared image non-uniformity)

一种快速的算术编码程序. 也是JPEG和JPEG2000的部分核心算法.(A fast arithmetic coding procedures. JPEG and JPEG2000 is also part of the core algorithm.)

说明：  matlab 实现小波图像去噪，自编程序非调用库函数(matlab implementation of wavelet image denoising, non-self-compiled programs call library functions)

基于核函数的SAR图像目标识别,利用KPCA对SAR图像进行特征提取(SAR image target recognition based on kernel function，Use KPCA SAR image feature extraction)

K均值图像分割，读入彩色图片，输出区域分割后的图片(K-means image segmentation, read full-color photographs, the output image after the region segmentation)

有效地MR图像重建算法，主要用于在MR图像的成像过程中对图像进行压缩(Efficient MR Image Reconstruction for Compressed MR Imaging)

基于DCT的jsteg图像隐写及分析算法，MATLAB实现。 JSteg的算法主要思想是将秘密消息嵌入在量化后的DCT系数的最低比特位上，但对原始值为O、1的DCT系数不进行嵌入。提取秘密消息时，只需将载密图像中不等于0、l的量化DcT系数的LSB取出即可。(The main idea of the JSteg algorithm is to embed secret messages in the lowest bit of the quantized DCT coefficients, but the DCT coefficients of the original values are O and 1 are not embedded. When extracting secret messages, only the LSB of the quantized DCT coefficients not equal to 0 and l in the encrypted image can be taken out.)

E:

说明：  利用IDL语言编写程序 计算差值植被指数，包括波段的读取，波段数据的分块和指数计算(Using IDL language program to calculate the difference vegetation index, including the band read the sub-band data block and index calculation)

图像检测的算法，Faster R-CNN算法，先对整张图像进行卷积计算，然后通过感兴趣区域池化层(RoI Pooling Layer)将选择性搜索算法推荐出来的候选区域和卷积网络计算出的特征映射图进行融合，得到候选区域对应的特征矢量，这种共享卷积计算的操作极大地减少了卷积计算的次数。而且这些特征矢量的维度统一，方便后续的分类工作。通过感兴趣区域池化层处理卷积特征，并将得到的特征送往两个并行计算任务进行训练，分类和定位回归。通过这些方法和改进的框架，Fast R-CNN 用更短的训练和测试时长，取得了比 R-CNN 更好的效果(Faster R-CNN algorithm first convolutes the whole image, then fuses the candidate regions recommended by the selective search algorithm and the feature mapping maps calculated by the convolution network through the RoI Pooling Layer to get the corresponding feature vectors of the candidate regions, which greatly reduces the number of convolution calculations. Moreover, the dimension of these feature vectors is unified, which facilitates the subsequent classification work. The convolution feature is processed by the pooling layer of the region of interest, and the obtained feature is sent to two parallel computing tasks for training, classification and positioning regression. Through these methods and improved framework, Fast R-CNN uses shorter training and testing time and achieves better results than R-CNN.)