Jan Erik Solem - Programming Computer Vision with Python [+Code] [2012, PDF, ENG]

Preface 7
Prerequisites and Overview 8
Introduction to Computer Vision 9
Python and NumPy 10
Notation and Conventions 10
Acknowledgments 11
1 Basic Image Handling and Processing 13
1.1 PIL – the Python Imaging Library 13
1.2 Matplotlib 16
1.3 NumPy 20
1.4 SciPy 31
1.5 Advanced example: Image de-noising 39
2 Local Image Descriptors 45
2.1 Harris corner detector 45
2.2 SIFT - Scale-Invariant Feature Transform 52
2.3 Matching Geotagged Images 63
3 Image to Image Mappings 73
3.1 Homographies 73
3.2 Warping images 78
3.3 Creating Panoramas 91
4 Camera Models and Augmented Reality 103
4.1 The Pin-hole Camera Model 103
4.2 Camera Calibration 109
4.3 Pose Estimation from Planes and Markers 110
4.4 Augmented Reality 114
5 Multiple View Geometry 127
5.1 Epipolar Geometry 127
5.2 Computing with Cameras and 3D Structure 136
5.3 Multiple View Reconstruction 144
5.4 Stereo Images 152
6 Clustering Images 161
6.1 K-means Clustering 161
6.2 Hierarchical Clustering 169
6.3 Spectral Clustering 175
7 Searching Images 185
7.1 Content-based Image Retrieval 185
7.2 Visual Words 186
7.3 Indexing Images 190
7.4 Searching the Database for Images 194
7.5 Ranking Results using Geometry 199
7.6 Building Demos and Web Applications 202
8 Classifying Image Content 209
8.1 K-Nearest Neighbors 209
8.2 Bayes Classifier 218
8.3 Support Vector Machines 223
8.4 Optical Character Recognition 228
9 Image Segmentation 237
9.1 Graph Cuts 237
9.2 Segmentation using Clustering 248
9.3 Variational Methods 252
10 OpenCV 257
10.1 The OpenCV Python Interface 257
10.2 OpenCV Basics 258
10.3 Processing Video 262
10.4 Tracking 265
10.5 More Examples 273
A Installing Packages 279
A.1 NumPy and SciPy 279
A.2 Matplotlib 280
A.3 PIL 280
4 CONTENTS
A.4 LibSVM 281
A.5 OpenCV 281
A.6 VLFeat 282
A.7 PyGame 282
A.8 PyOpenGL 283
A.9 Pydot 283
A.10 Python-graph 283
A.11 Simplejson 284
A.12 PySQLite 284
A.13 CherryPy 285
B Image Datasets 287
B.1 Flickr 287
B.2 Panoramio 288
B.3 Oxford Visual Geometry Group 289
B.4 University of Kentucky Recognition Benchmark Images 289
B.5 Other 290
C Image Credits 291

The final pre-production draft of the book (as of March 18, 2012).