Simon Shi的小站

人工智能,机器学习 学习记录


  • Home

  • About

  • Tags

  • Categories

  • Archives

Math -- Matrix

Posted on 2020-09-02 | In Math
矩阵协方差矩阵正交矩阵正定矩阵https://zhuanlan.zhihu.com/p/44860862 【定义1】给定一个大小为 $n \times n$ 的实对称矩阵 $A$ ,若对于任意长度为$n$ 的非零向量 $x$,有$x^TAx > 0$ 恒成立,则矩阵 $A$ 是一个正定矩阵。 负定矩阵对角矩阵相似矩阵可逆矩阵初等矩阵等价矩阵施密特正交化实对称矩阵斜对称矩阵反对称矩阵旋转矩阵黑塞矩阵稀疏矩阵分解
Read more »

Scipy (Python 的科学计算包)

Posted on 2020-08-26 | In dev , python , Scipy
[TOC] sparsekron1234567891011121314>>> from scipy import sparse>>> A = sparse.csr_matrix(np.array([[0, 2], [5, 0]]))>>> B = sparse.csr_matrix(np.array([[1, 2], [3, 4]]))>>> sparse.kron(A, B).toarray()array([[ 0, 0, 2, 4], [ 0, 0, 6, 8], [ 5, 10, 0, 0], [15, 20, 0, 0]])>>> sparse.kron(A, [[1, 2], [3, 4]]).toarray()array([[ 0, 0, 2, 4], [ 0, 0, 6, 8], [ 5, 10, 0, 0], [15, 20, 0, 0]]) $A \times B$ ...
Read more »

Graphices Point Cloud 点云数据处理方法-刚性篇

Posted on 2020-08-17 | In Graphices , PointCloud
[TOC] 点云数据处理方法概述 ICP点云配准就是我们非常熟悉的点云处理算法之一。实际上点云数据在形状检测和分类、立体视觉、运动恢复结构、多视图重建中都有广泛的使用。点云的存储、压缩、渲染等问题也是研究的热点。随着点云采集设备的普及、双目立体视觉技术、VR和AR的发展,点云数据处理技术正成为最有前景的技术之一。PCL是三维点云数据处理领域必备的工具和基本技能,这篇文章也将粗略介绍。 三维点云数据处理技术(刚性变换篇)1. 点云滤波(数据预处理)点云滤波,顾名思义,就是滤掉噪声。原始采集的点云数据往往包含大量散列点、孤立点,比如下图为滤波前后的点云效果对比。 点云滤波的主要方法有:双边滤波、高斯滤波、条件滤波、直通滤波、随机采样一致滤波、VoxelGrid滤波等,这些算法都被封装在了PCL点云库中。 2. 点云关键点(提取算法)我们都知道在二维图像上,有Harris、SIFT、SURF、KAZE这样的关键点提取算法,这种特征点的思想可以推广到三维空间。从技术上来说,关键点的数量相比于原始点云或图像的数据量减小很多,与局部特征描述子结合在一起,组成关键点描述子常用来形成原始数据的表示 ...
Read more »

3D 点云可視化

Posted on 2020-07-23 | In Graphices , PointCloud
3D visiual[TOC] Pyrender123456789import trimeshimport pyrenderfuze_trimesh = trimesh.load('hello.obj')mesh = pyrender.Mesh.from_trimesh(fuze_trimesh)scene = pyrender.Scene()scene.add(mesh)pyrender.Viewer(scene, use_raymond_lighting=True) psbodyMesh 1234from psbody.mesh import Mesh, MeshViewerscan = Mesh(filename=join(path, 'smpl_registered.obj'))scan.set_texture_image(tex_file='***.jpg')scan.show() MeshViewer 12345678from psbody.mesh import Mesh, MeshViewerscan = ...
Read more »

Algorithm Impl Render

Posted on 2020-07-23 | In Algorithm , CG
[TOC] Render impl todo
Read more »

Algorithm Impl CV/CG Transform

Posted on 2020-07-23 | In Algorithm , CG
Transform[TOC] angle2matrix1234567891011121314151617181920212223242526def angle2matrix(angles): ''' get rotation matrix from three rotation angles(degree). right-handed. Args: angles: [3,]. x, y, z angles x: pitch. positive for looking down. y: yaw. positive for looking left. z: roll. positive for tilting head right. Returns: R: [3, 3]. rotation matrix. ''' x, y, z = np.deg2rad(angles[0]), np.deg2rad(angles[1]), np.deg2rad(angles[2]) # x ...
Read more »

Algorithm Survey

Posted on 2020-07-23 | In Algorithm , CG
[TOC] 算法类型图形学算法S,R,T Scala 旋转 位移 对齐 (mtcnn align 算法) 仿射变换 (CV-transformations.md) 线性插值 CG Algorithm (Algorithm-CG.md) /home/simon/data/blogs/source/_posts/CV-3D-Transform.md LBS算法 (CV-3D-BuildModel-CMPL.md) 点云对齐算法CV-3D-Base.md 统计机器学习算法 k-means 数据降维 PCA 数据降维基础 奇异值分解SVD 特征值分解EVD 游戏树算法 minimax pn-search alpha-beta剪枝 增强学习算法 MCTS UCT CFR
Read more »

Algorithm CG

Posted on 2020-07-23 | In Algorithm , CG
Read more »

Python Lib OpenCV

Posted on 2020-07-16 | In dev , OpenCV , python
[TOC] OpenCV -API (python) APi Tutorialsimgproc 模块. 图像处理 OpenCV 2.3.2 documentation 图像平滑(模糊)处理 blur /GaussianBlur / MedianBlur / BilaterFilter 腐蚀-膨胀 形态变换 一、Image I/Oimreadimsaveimshow()resizecv2.resize的第二个参数dim是(W, H) 1cv2.resize(array, (W, H)) copyMakeBordercv2.copyMakeBorder的第二个到第五个参数是top, bottom, left, right,是先H后W opencv中以左上角为原点,W方向为x,H方向为y 1def copyMakeBorder(src, top, bottom, left, right, borderType, dst=None, value=None): # real signature unknown; restored from __doc__ 扩充src的边缘 ...
Read more »

PyTorch Doc Detail

Posted on 2020-07-15 | In DNN_platform , pytorch
APIs [TOC] Tensorpytorch12345678910111213141516171819202122#1 list->tensordata = [[1,2],[3,4]]x_data = torch.tensor(data, dtype=torch.float)### torch.FloatTensor, torch.DoubleTensorfloat_tensor = torch.FloatTensor([4,5,6])#2 np.array -> tensornp_array = np.array(data)x_np_data = torch.from_numpy(data)#3 tensor -> tensorx_ones = torch.ones_like(x_data)x_rand = torch.rand_like(x_data, dtype=torch.float)#4 new empty tensorshape = (2,3)x_ones = torch.ones(shape)x_zeros = torch.zeros(shape)x ...
Read more »
1…192021…29

Simon Shi

286 posts
132 categories
243 tags
RSS
© 2024 Simon Shi
Powered by Hexo
|
Theme — NexT.Muse v5.1.4