Presenter:

1
Boost CVPR
http//www.boost.org/

• Presenter? ? ?

2
Contents

• Boost Multidimensional Array Library
• by Ronald Garcia
• Open System Lab, Indiana University
• Boost Graph Library
• by Jeremy Siek
• Indiana University
• Boost Lamda Library
• by Jaakko Järvi
• Boost Computer Vision Pattern Recognition
  Library
• Natl. Lab. of Pattern Recognition, CASIA
• At current, contains contributions from Haifeng,
  Hongxia, Xiaoxing.
• Under developing, any kind of contributions are
  welcome.

3
Boost MultiArray Library

• Introduction
• Short Example
• Creating Views
• Facilities for CVPR

4
Introduction

• The Boost MultiArray library provides a class
  template specifying an interface for
  N-dimensional containers.
• It includes a general array class template and
  array adaptors.
• The MultiArray class supports idiomatic array
  operations and interoperate with C Standard
  Library containers and algorithms.

5
Short Example
Instantiate the template

• include "boost/multi_array.hpp"
• using namespace boost
• int main()
• typedef multi_arrayltfloat,3gt  MultiArray
• typedef MultiArrayindex  index
•  MultiArray   A(extents343)
• int values    0
• for(index  i    0 i  lt  3 i )
• for(index  j  0 j  lt  4 j )
• for(index  k    0 k  lt  3 k )
• Aijk    values
• return 0

A signed integral type used for indexing into A
Create a 3D array that is 3 x 4 x 3
Access elements
6
Creating Views

• The sub-view supports the similar interface as
  the original array.
• It can retain the same number of dimensions as
  the original.
• It can have less dimensions than the original.

7
Examples

• typedef  MultiArrayarray_viewlt3gt type 
  view3D
• typedef  MultiArrayarray_viewlt2gt type 
  view2D
• typedef  MultiArrayindex_range  r
•  
• view3D aView3  A indicesr(0,2)r(0,4,2)r() 
  
• view2D aView2  A indices0r()r()     

The view type with Dims dimension template
array_viewltDimsgt type
This type specifies a range of indices over some
dimension of a MultiArray.
from beginning to end
base, stride, bound) 0,2,4) 0, 2
base,bound) 0,2) 0, 1
8
Facilities for CVPR

• The MultiArray container perfectly suits
  3D-graphics and digital image.
• For 3D-grahpics, slices are gained effortlessly
  with the tool of creating views.
• For digital images, we can easily choose one
  color channel similarly.

9
More Facilities

• By creating a view type of the same dimensions of
  the original, we can easily sample on a multigrid
  lattice.

e.g. use MultiArray to contain a color image
(256256) MultiArray   A(extents2562563
) create such a view view3D sample  
A indicesr(0,256,4)r(0,256,4)r(
) 
10
More Facilities

• MultiArray provides a flexible process on the
  numbers of dimension.

• e.g.
• If we have a group of images for training,
  , with the size of m x n.
• The MultiArray of one more dimension than A
  can be created to contain this group. It is
  convenient to train all these images (2D
  (mn)xK), as well as to show the result of just
  one image (3D mxnxK).

• More facilities for CVPR are to be developed.

11
Boost Graph Library

• Generic Data Structures
• Generic Graph Algorithms

12
Generic Data Structures

• adjacency_list
• adjacency_matrix
• edge_list

• Suit directed, undirected, or bidirected graph
  when specially parameterized
• Provide containers for both node and edge

13
Generic Graph Algorithms

• Breadth First Search
• Depth First Search
• Uniform Cost Search

• Build blocks for constructing graph algorithms,
  currently including 13 in the BGL, such as
  connected components.

14
Boost Lambda Library

• Unnamed functions can be created easily with an
  intuitive syntax.
• e.g. for_each( a.begin(), a.end(), cout ltlt
  (1 _1))
• Most of the restrictions in argument binding are
  removed, arbitrary arguments of practically any
  C function can be bound.
• Separate function composition operations are not
  needed, as function composition is supported
  implicitly.

15
Boost CVPR Library

• Introduction
• Commonly Used Components

• Kernels Filters
• Color Space Transformation
• Color Transformation
• Array Utils
• IO Stream
• Read Write QT, ImageMagic
• Statistical Learning
• Plotting
• Poisson Image Editting

16
Introduction

• The Boost CVPR library is an specialized
  application of Boost MultiArray.
• The Boost CVPR library, containing a family of
  class and function templates, is intended to
  offer basic tools for image processing.

17
Kernels Filters
defined in boost_cvpr/filter.hpp

• Providing function templates to generate kernels,
  including gaussian, dog (a derivative of
  gaussian), disk, disko, sixneighbor.
• Providing function templates for filtering and
  resizing.
• filt
• resize

18
Examples
Generate a gaussian kernel
Output a multiarray by overloaded function
coutltlt

• multi_arrayltfloat,2gt gauKernel(extents55)
  float sigma  1.5
• gaussian(gauKernel,sigma) coutltlt"gauKernel  "lt
  ltendlltltgauKernelltlt""ltltendl multi_arrayltfloat, 
  3gt in multi_arrayltfloat, 3gt out
  load_image(in, "in.bmp")     
  out.resize(extentsin.shape()0
• in.shape()1in.shape()2)     
• for(index cc  0 cc lt in.shape()2 cc)
      view2D filt_in  in indicesr()r()cc 
      view2D filt_out  out indicesr()r()cc
        filt(filt_in,gauKernel,filt_out)
  save_image(out,"out.bmp")

Specify the extents of out
Read
Smooth filter
Write
19
Color Space Transformation
defined in boost_cvpr/color_space.hpp

• This head file provides function templates to
  transform RGB to LUV color space and in the
  reverse direction.

rgb2luv(rgbImageArray, luvImageArray) luv2rgb(luv
ImageArray, rgbImageArray)
MultiArray
20
Color Transformation
defined in boost_cvpr/color_trans.hpp

• This head file provides several components for
  color amendments.

nbin is the number of the color scales, and maxv
is the maximal value after process.

• Five useful functions
• saturate (in, lb, ub)
• histogram (in, hist, bin)
• histeq (in, out, nbin, maxv)
• level (in, out, bin, curv)
• autolevel (in, out, lb, ub)

p lt lb, let p lb p gt ub, let p ub.
curv bin are 1D multiarray of the same size,
guiding the leveling process.
hist bin are 1D multiarray of the same size.
histi store the statics data of the pixals
being the color of bini.
out can be omitted.
21
Array Utils
defined in boost_cvpr/array_utils.hpp

• Wrapper of multi-folded loops for multiarray
• multi_for_each(_const)
• multi_copy(_const)
• multi_arith(_const)

e.g.multi_arrayltfloat, 3gt  a(extents100100)
multi_arrayltfloat, 3gt  b(extents100100)
multi_arrayltfloat, 3gt  c(extents100100)
multi_for_each(a, _12) multi_copy(a,
b, _22_1) multi_arith(a, b, c, _32_1_2)
22
IO Stream
defined in boost_cvpr/cvpr_stream.hpp

• By overloading the operator ltlt, the objects of
  multi_array, array_view, and sub_array types can
  be output at one time.

• Using the sentences just like (both file and std.
  console)
• coutltlt A cingtgt A
• fcoutltlt A fingtgt A

23
Read Write
defined in boost_cvpr/cvpr_qt.hpp and
boost_cvpr/cvpr_magick.hpp

• Using QT to load and save images

• 3 functions are designed
• load_image() save_image() save_sequence()

• Components defined in cvpr_magick.hpp, using
  ImageMagick, have similar usages.

24
Statistical Learning

• ISOMAP
• k-nearest neighbor
• eigs
• PCA
• Connected Components
• For both graph and image lattice
• connect_components

25
Plotting - Open Scene Graph
26
Poisson Image Editing
27

• Any Questions?
• Thanks!

THE END