Population release 5.0

After one year of work, a new version is available. More information soon.

Posted in Uncategorized

Android and Automatic Number Plate Recognition (ANPR)

Vincent Tariel has developed an optical character recognition with Population library for the society alphanumeric-vision. For the ANPR market in mobile devices, during her internship, Sonia Adamah has integrated this code with the android-NDK and swig (to bind the C++ code). You can test the lite application .

androidANPR

Tagged with:
Posted in Uncategorized

Articles

I am quite happy ! Some external people begin to use Population and they have nice results with modules Processing and Analysis. See the publications:

J.-F. Bruchon, J.-M. Pereira, M. Vandamme, N. Lenoir, P. Delage, M. Bornert, Full 3D investigation and characterisation of capillary collapse of a loose unsaturated sand using X-ray CT. Granular Matter, 2013
SegSand
J.-F Bruchon., Pereira J.-M., M., Vandamme, N. Lenoir, P. Delage and M. Bornert X-ray microtomography characterisation of the changes in statistical homogeneity of an unsaturated sand during imbibition Géotechnique letter, 2013

Tagged with:
Posted in Uncategorized

Art

The pipeline of the germ grain model is the construction of: a germ, a collection of points, a grain, a collection of geometrical figures dressing the points, a colour, a collection of colours dressing the geometrical figures, a model, a structure defining by an interaction principle between the coloured grains.

In the following realization of random germ grain models, the pipelines are the same except the model.

Initial Image:
art

Dead leave model:
art_aborigene

Transparant model:
art_aborigene

Spot noise model:
art_aborigene

I am almost finished to write the article related to.

Tagged with:
Posted in Uncategorized

Panorama with SIFT algorithm

Now, I work for a vision company. I should use OpenCV because this fantastic open-source library is dedicated to this field. But I am quite productive with Population, so I test it! Oke, it is not optimized as OpenCV (no platform primitive optimization). However, I get some very nice results. For I project, I have to create a panorama as in this article and from these images

photo2
photo3

photo4
I get this result:
panorama

#include"Population.h"//Single header
using namespace pop;//Population namespace
int main(){
        std::string path= "D:/Users/vtariel/Downloads/";
        Mat2RGBUI8 img3;
        img3.load(path+"Photo1.jpg");
        Mat2RGBUI8 img4;
        img4.load(path+"Photo2.jpg");
        Mat2RGBUI8 img5;
        img5.load(path+"Photo3.jpg");

        Mat2RGBUI8 img6;
        img6.load(path+"Photo4.jpg");
        Vec<Mat2RGBUI8> vv;

        vv.push_back(img4);
        vv.push_back(img5);
        vv.push_back(img6);

        Mat2RGBUI8 panoimg =Feature::panoramic(vv);
        panoimg.display();    
}
Posted in Uncategorized

Python binding

Population has a built in python interface that utilizes SWIG for the language connection. In this interpreted language, a practitioner can learn rapidly in test-error method because time compilation. The notation are almost the same between C++ and Python. So it is easy to switch. If something does not work or if you have any suggestions, please leave a comment.

Posted in Uncategorized