You have chosen to sponsor your bid up to a maximum amount of .
I've a face detection code that reads image and detects faces. The code is based on skin training where covariance and mean is calculated and skin likelihood equation is implemented on the image.
My code needs the following to be optimized and enhanced:
1- Skin likelihood threshold. I used a fixed threshold which can not detect all skin areas. To optimize I suggest using histogram on the likelihood image and calculate the threshold.
2- Region growing need to be optimized as it includes non skin areas. RGB histogram on the original should be used for region growing optimization with my existing code that check the 8-connected pixels distance.
3- Eliminates non face blobs.
4- Draw Ellipse on the center of the face blob and fill it using fillholes Aforge library which it already included.
The image data set can be downloaded from
[url removed, login to view]
In the one face folder, for example, [url removed, login to view], [url removed, login to view], [url removed, login to view], [url removed, login to view] you ca
The code is C# 2008.
Any other optimization is welcomed