Skip to content

valdecar/faceRepresentWithHoG

Repository files navigation

faceRepresentWithHoG

description

With this tool one can create classes (labels) for images (manually)
from database and then transform that images to vectors of features space
(using HoG - histogram of oriented gradients method).
This space will be divided at created class (using knn classifier).
In the end one can use this features space for classify some particular image.

Features space created using Scale-Invariant Feature Transform
(http://docs.opencv.org/3.1.0/da/df5/tutorial_py_sift_intro.html#gsc.tab=0)

Default imgDB use one of these:http://www.face-rec.org/databases/
You can also create own image database in same format like imgDB (see DB/createBD.py)

alt tag

requirements

python3.4,opencv2,numpy,scipy,skimage,sqlobject
download imgDB.db and prepImg.db from

imgDB: https://drive.google.com/open?id=0B5urs7cbv2vJRDFxNGNOaWVRSFE

prepImg: https://drive.google.com/file/d/0B5urs7cbv2vJcVBBWnE5b0czV2s/view?usp=sharing

add path to "img.db" in "faceRepresentWithHoG/DB/paths.txt" file (in dictionary form) 

also add path to "prepImg.db" in "faceRepresentWithHoG/DB/paths.txt" file (in dictionary form)

usage

python3 gui.py

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages