How to improve the image color and bightness quality using opencv

asked 2016-12-07 01:24:50 -0500

Huz gravatar image

updated 2016-12-07 03:27:01 -0500

Here is a link of what I want to implement using opencv, the human detection part is the hardest to replicate. http://www.howtogeek.com/69929/how-to... Basically I want to detect the part of the face that is not illuminated well automatically which is done using manual interfacing and then apply histogram equalization to the image. So I am stuck in the face detection part, how to decide if for an image we have such areas of dim illumination and then detecting the areas. For eg. the image in the above link or http://www.adorama.com/alc/alc_images...

edit retag flag offensive close merge delete

Comments

2

please try to explain what you want here , not with a lame link.

(let's re-open it, if you come up with an update !)

berak gravatar imageberak ( 2016-12-07 01:27:27 -0500 )edit

So I am stuck in the face detection part, how to decide if for an image will have such areas of dim illumination and then detecting the areas.

Huz gravatar imageHuz ( 2016-12-07 02:44:11 -0500 )edit

one of your "bad" example images would be extremely helpful.

berak gravatar imageberak ( 2016-12-07 02:58:27 -0500 )edit
1

Make an intensity histogram of the face region. If it is not flat (or averaged), and has excessive peaks, then you have a badly illuminated face! It is basically what histogram equalization tries to undo in face detectors in OpenCV

StevenPuttemans gravatar imageStevenPuttemans ( 2016-12-07 03:03:49 -0500 )edit

How to detect the face region using opencv as in a general image we do not have a clar cut demarkation and if we try to distinguish the face using colors there is an error because of the bad illumination which makes it difficult to detect the face region exactly

Huz gravatar imageHuz ( 2016-12-07 03:29:10 -0500 )edit

use the face detector?

StevenPuttemans gravatar imageStevenPuttemans ( 2016-12-07 03:51:19 -0500 )edit