Milde

MiLDe is a powerful environment (similar to MatLab) designed to develop applications that use machine learning algorithms.  MiLD3 also features an extensive image analysis and a full-featured numerical library.  MiLD3 brings these tools together using the Lua interpreted programming language and an integrated environment that includes a graphical user inerface with a full-featured script editor, debugging capabilities and interactive image manipulation.

Features:

Efficient SVM+ implementation

  • Multi-class problems      
  • Dense or Sparse vectors
  • Fast caching mechanisms

A parallel implementation of SVM+ for solving very large problems:

  • MMX/SSE on x86 processors
  • Support for BLAS
  • Multi-threading on dual/quad machines
  • Full parallelization of single optimization on cluster of processors (patent pending)
An extensive set of Standard Machine Learning Algorithms
  • Integrates the Torch module
  • Gradient descent-based algorithms, including multi-layered perceptrons, radial basis functions, mixtures of experts, convolutional networks and time-delay neural networks.
  • Ensemble models such as bagging or adaboost.
  • Non-parametric models such as K-nearest-neighbors, Parzen regression and Parzen density estimator.
  • Distributions stuff, like Kmeans, Gaussian mixture models, hidden Markov models, input-output hidden Markov models, and Bayes classifier.
Large library for image processing and analysis
  • color analysis (PCA, clustering, HSL mapping, etc.)
  • image processing (Gabor, edges, smoothing, sharpening, thinning, etc.)
  • interpolated image transforms (rotation, texture mapping)
  • image analysis (connected components, feature extraction, contour extraction (snakes, active contours), shape extraction, shape filtering, shape clustering, shape overlap, etc.)
  • i/o support for TIFF, JPEG, MPEG2, AVI, RAW
Complete Numerical Library
  • Basic vector – matrix operations
  • Complex Numbers
  • Popular linear algebra algorithms: Linear equation solver, Eigen-decomposition, PCA, LDA.
  • Kernel space functions: KPCA
  • BLAS and Lapack
The Lua Programming Language
  • Full-fledged interpreted programming language (www.lua.org)
  • Modular by design.
  • Acts as the ‘glue’ between the various libraries by handling objects passing between library calls.

A fully integrated graphical user interface

  • Full Text Editor for scripts with automatic coloring.
  • Graphical debugging facilities (error highlight, selective execution, etc.)
  • Interactive on-line Help and Reference for all library functions.
  • Display and interactive manipulation of images, 2D and 3D graphs, histograms and scatterplots.
  • Interaction with vectorized shapes obtained from image analysis functions (creation, edition, deletion, selection) with intuitive control of selective overlays.
  • Support for building end-user applications with dialogs for parameters-setting, interactive menus, etc.



Screenshot of MiLDe's integrated development environment

 

Downloads:


Collaborators:  Eric Cosatto, Iain Melvin, Hans Peter Graf, Igor Durdanovic, Christopher Malon, Pierre-Francois Laquerre