Hello.
I am new to OpenCV, and having already tried with success the examples, and tweaking to find face, eyes, mouth and nose; I would like to make a detector for ears with frontal face. The main purpose of the project I have in hands is to make ear detection with frontal face, but the information I collected hints me that there is no classifier for frontal face ear detection, so I am going to make one. And it is for this that I need advice.
I already have fotos to prepare the images for positive training, and am starting to process them so to leave only the ears, so I ask for your advice: - Should I leave only the ears, or do you suggest I leave a bit of skin from the face and the ear? - Should I create each positive with both ears, and on the creation of samples state both object instances, or is it better and easier to make one image for each ear?
Thank you for any help.