Vision Lab Features Library (VLFeat)

Version

0.9.21

Author

The VLFeat Team

Copyright © 2012-14,18 The VLFeat Authors

Copyright © 2007-11 Andrea Vedaldi and Brian Fulkerson

The VLFeat C library implements common computer vision algorithms, with a special focus on visual features, as used in state-of-the-art object recognition and image matching applications.

VLFeat strives to be clutter-free, simple, portable, and well documented.

Contents

• Visual feature detectors and descriptors
  • Scale Invariant Feature Transform (SIFT)
  • Dense Scale Invariant Feature Transform (DSIFT)
  • Maximally Stable Extremal Regions (MSER)
  • Covariant feature detectors
  • Gaussian Scale Space (GSS)
  • Histogram of Oriented Gradients (HOG) features
  • Fisher Vector encoding (FV)
  • Vector of Locally Aggregated Descriptors (VLAD) encoding
  • Local Intensity Order Pattern (LIOP) descriptor
  • Local Binary Patterns (LBP) descriptor
• Clustering and indexing
  • K-means clustering
  • Integer K-means (IKM)
  • Hierarchical Integer K-means (HIKM)
  • Gaussian Mixture Models (GMM)
  • Agglomerative Information Bottleneck (AIB)
  • KD-trees and forests
• Segmentation
  • Simple Linear Iterative Clustering (SLIC)
  • Quick shift image segmentation
• Statistical methods
  • Agglomerative Information Bottleneck (AIB)
  • Homogeneous kernel map
  • Support Vector Machines (SVM)
• Utilities
  • Random number generator
  • Mathematical operations
  • String operations
  • Image operations
  • PGM image format
  • Generic heap object (priority queue)
  • Rodrigues formula
  • MATLAB MEX helper functions
  • Drop-in getopt_long replacement
• General information
  • Conventions
  • General support functionalities
  • Portability features
  • Memory and resource handling
  • Objects
  • Threading
  • MATLAB integration
  • Preprocessor metaprogramming
• Developing the library
• Glossary