customized facedetect script for OpenCV haarcascdes

I am looking for someone who will write me a customized facedetect script based on the OpenCV haarcascades. The script should be configurable through a config file (i attached a sample file, see below). The script should be able to: - process only the areas (x, y coordiantes, rectangle/circle) that are defined in the config file to speed up the process and keep false negatives away. - blur out the area that was detected by openCV or if no cascade defined (area2,area3), blur out the whole area. - only processes pictures (up to 15 Megapixel) !!! no stream or webcam. Attached you will find 2 pictures, the first shows the original picture with the areas that are defined in the config sample. The second one shows the expected result after processing. Furthermore openCV has some additional parameters that should be configurable for each area (for example minSize, maxSize, minNeighbors). You will find them together with the explanation in the sample config file. The "blur" should be customizable between 1 (less blur) and 10 (maximum blur factor). [login to view URL] [login to view URL] [login to view URL] The script can be written in c,c++ or python. it should be runable from my linux shell with args for config file, input file, output file (opencv, c compiler, python opencv support installed and everything running fine). So you are only asked to provide the script. PS: I am aware that in this example the face is from behind/side and recognition wont work properly - in fact this doesnt matter its just for your illustrtion, you may place your own face inside the area if you consider testing it. Furthermore i would be glad to establish a long time realtionship for further programming and extensions.