资源列表
image
- 基于PSO 的图像分割,运用粒子群优化算法在vc++6.0上对图像进行边缘分割- U57FA u4E8EPSO u7684 u56FE u50CF u5206 u5272 uFF0C u8FD0 u7528 u7C92 u5B50 u7FA4 u4F18 u5316 u7B97 u6CD5 u5728vc++6.0 u4E0A u5BF9 u56FE u50CF u8FDB u884C u8FB9 u7F18 u5206 u5272
reade-image
- 行人检测的程序,用于检测行人,并将行人进行简单的分割出来,便于后续的处理。-Pedestrian detection procedure for detecting pedestrians and pedestrians simple segmented to facilitate subsequent processing.
a-new-color-constancy-algorithm
- 在一般的拍摄情况下,光照是平行均匀地照射到物体上的,往往关注光源颜色的变化而忽略了亮度信息,导致校正效果不理想,于是提出了基于亮度的改进算法。该算法的基本思想是,将图像均匀分块求取中心像素并对亮度进行加权,然后求取近似于整幅图像R、G、B三分量的平均灰度值,改进GrayWorld算法时用亮度加权取代了平均灰度值,最后将它与White-patch算法整合到一起,建立方程组解出调节参数,并根据参数进行调整。-In normal shooting situations, the light is un
Self_Adaption
- 一幅图像的动态范围一般是指亮度的最大值和最小值的比值。人眼对物体的分辨率是非常高的,它亮度的动态范围可达到1000000000:1,而一般的成像设备的范围比人眼看到的要小得多,达不到人眼亮度动态范围的千万分之一。如果要让设备上的显示能尽可能地接近人眼,显示出更多的信息,因此需要进行动态范围压缩处理。-The dynamic range of an image generally refers to the ratio of the maximum and minimum brightness.
Gray_World
- Gray world认为R、G、B三个分量对光源的感知是等价的,该算法的基本思想是:通过利用公式求取三个基本颜色红、绿、蓝的像素平均值来确定出整个图像的整体平均值,再以整体平均值为基础调整每个像素的红、绿、蓝分量值,以达到校正色彩偏移的效果。物体表面的平均反射没有色度差。-Gray world that the R, G, B three components of light perception is equivalent to the basic idea of the algorithm
double-ellipses-skin-color-model
- 在复杂背景下双椭圆模型结合在CbCr和CgCr平面内体现的肤色聚集性和集中性的检测效果要比单高斯模型的效果好。在YCbCr和YCgCr颜色空间中,虽然Y分量和Cb、Cr、Cg分量表示的性质看似是相互独立的,但它们都是建立在红、绿、蓝三个基本颜色分量值的基础上算出来,因此它们之间仍然存在着千丝万缕的关系。根据肤色在两种色彩空间中体现出的聚集性提出CbCr-CgCr双椭圆肤色模型-Good background in complex dual oval model incorporated with
Homomorphic-filtering
- 图像增强的同态滤波算法matlab程序,能很好的适用于夜间等场景,并附测试图片-Image enhancement homomorphic filtering algorithm matlab program, can be well applied to other scenes at night, along with test pictures
BPshouxieshuzishibie
- 基于BP神经网络的手写数字识别,总共有三个程序,分别是建立样本,进行训练,最后是检验测试-Based on the BP neural network handwritten numeral recognition, a total of three procedures, namely, the establishment of samples, training, and finally test
watermark
- 基于脆弱性的水印嵌入算法,包括水印嵌入,水印攻击等-Based on the vulnerability of the watermark embedding algorithm, including watermark embedding, watermark attacks and so on
stable_watermaker
- 此算法是基于鲁棒性的水印嵌入法,通过对水印图像进行噪声处理,可以保持稳定性,同时可以定位图像的篡改位置-This algorithm is based on robust watermark embedding method, through the watermark image noise treatment, can maintain stability, and can locate the image of the tampering position
libfacedetection-master
- 于仕祺老师开源的人脸检测源码,能够检测多张人脸,效果比opencv自带人脸检测好-Shi Qi teacher in open source face detection, capable of detecting multiple faces, better than opencv face detection comes good
SkinColorDetection-master
- 皮肤检测算法源码,c++实现,采用四种不同策略提取皮肤部分。-Skin detection algorithm source code, c++ implementation, using four different strategies to extract part of the skin.