-
py-faster-rcnn-master
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.)
- 2020-12-11 15:39:18下载
- 积分：1
-
icme14_deblur_code_orginal
代码实现2014ICML文章Motion blur kernel estimation via salient edges and low rank prior，基于显著边缘和低秩先验的运动模糊核估计。(The code implementation of Motion blur kernel estimation via salient edges and low rank prior)
- 2018-11-12 23:48:03下载
- 积分：1
-
ImageRetrieval
毕业设计，基于内容的图像检索，支持的检索特征包括 sift,颜色直方图，灰度矩阵，HU不变矩，边缘方向直方图，检索方法使用K-means和K-D树两种，需要OPENCV支持，运行时请先选定一个文件夹来生成特征库，特征库用access数据库保存，只支持JPG文件(Graduate design, content-based image retrieval, search features, including support sift, color histogram, gray matrix, HU moment invariants, edge direction histogram, retrieval method using the K-means and KD trees are two kinds of needs OPENCV support Please select a runtime folder to generate the feature library, feature library with access database save, only supports JPG files)
- 2010-09-14 17:31:44下载
- 积分：1
-
main
利用形态学，霍夫（Hough）变换，实现畸变QR码的4个角点的定位。(To correct the distortion of the QR code,use morphology operation, Hough transform, to locat the positioning of the four corner points)
- 2012-03-14 15:17:57下载
- 积分：1
-
IHS-PCA-fusion--Matlab
IHS,PCA加权图像融合三种算法的Matlab源代码(IHS,PCA Matlab )
- 2021-04-25 17:08:46下载
- 积分：1
-
balloon
本人自编的气球SNAKE模型代码，效果还算不错。(I own balloons SNAKE model code, the effect is still good.)
- 2008-12-18 17:01:38下载
- 积分：1
-
lypp
通过邻域搜索法对一个相机拍摄的序列图像中圆形非编码标志点进行匹配(Neighborhood search method by camera on a circular non-coding sequence of images to match landmarks )
- 2020-12-18 21:19:10下载
- 积分：1
-
DespeckleFilters
一个去除斑点噪声的滤波器的，特别适用于医学超声图像(A filter to remove speckle noise , especially for medical ultrasound image
)
- 2014-11-16 16:26:11下载
- 积分：1
-
第四次作业
说明： 1.使用medfilt2函数对‘lena.bmp’进行不同模板大小的中值滤波，并观察效果
2.使用fspecial函数构造‘gaussian’, ‘sobel’, ‘prewitt’, ‘laplacian’, ‘average’等不同滤波算子，使用filter2函数对‘cameraman.tif ’进行滤波，观察不同滤波器的效果
3.设计频域理想低通和理想高通滤波器，对‘phantom.bmp’图像进行频域滤波，显示滤波结果。然后改变截止频率，观察对结果的影响。
4.对‘homo.bmp’进行同态滤波，其中频域同态滤波器自行设计。要求在程序注释中说明你设计该滤波器的思想。(1' lena.bmp 'The median filtering with different template sizes was carried out and the effect was observed2. Fsspecial function is used to construct different filter operators such as' Gaussian ',' Sobel ',' Prewitt ',' Laplacian ',' average ', etc' cameraman.tif 'Filter and observe the effect of different filters3. Design ideal low pass and ideal high pass filters in frequency domain' phantom.bmp 'The image is filtered in frequency domain and the filtering result is displayed. Then change the cut-off frequency and observe the effect on the results.4. Yes' homo.bmp 'The homomorphic filter is designed by ourselves. The idea of designing the filter should be explained in the program notes.)
- 2020-12-11 21:53:25下载
- 积分：1
-
小波分析 编程
小波分析的各种尺度分解包括单尺度分解，多尺度分解以及小波包分解及重构，小波分解重构精简(All kinds of scale decomposition of wavelet analysis include single scale decomposition, multi-scale decomposition, wavelet packet decomposition and reconstruction, and wavelet decomposition and reconstruction.)
- 2018-04-11 09:10:51下载
- 积分：1
