Resize and crop an image based on OpenCV detected faces

faces_example.py

import sys

from PIL import Image, ImageDraw

try:
    import cv
except ImportError:
    print 'Could not import cv, trying opencv'
    import opencv.cv as cv


def main():
    source_image = Image.open('example.jpg')
    source_width, source_height = source_image.size
    print 'Image is {}x{}'.format(source_width, source_height)

    target_width = 1000
    target_height = 200

    # Make image a reasonable size to work with. Using the source_height will
    # make sure it's just resized to the target_width
    source_image.thumbnail((target_width, source_height), Image.ANTIALIAS)

    # Find the faces and show us where they are
    faces = faces_from_pil_image(source_image)
    faces_found_image = draw_faces(source_image, faces)
    faces_found_image.show()

    # Get details about where the faces are so we can crop
    top_of_faces = top_face_top(faces)
    bottom_of_faces = bottom_face_bottom(faces)

    all_faces_height = bottom_of_faces - top_of_faces
    print 'Faces are {} pixels high'.format(all_faces_height)

    if all_faces_height >= target_width:
        print 'Faces take up more than the final image, you need better logic'
        exit_code = 1
    else:
        # Figure out where to crop and show the results
        face_buffer = 0.5 * (target_height - all_faces_height)
        top_of_crop = int(top_of_faces - face_buffer)
        coords = (0, top_of_crop, target_width, top_of_crop + target_height)
        print 'Cropping to', coords
        final_image = source_image.crop(coords)
        final_image.show()
        exit_code = 0

    return exit_code


def faces_from_pil_image(pil_image):
    "Return a list of (x,y,h,w) tuples for faces detected in the PIL image"
    storage = cv.CreateMemStorage(0)
    facial_features = cv.Load('haarcascade_frontalface_alt.xml', storage=storage)
    cv_im = cv.CreateImageHeader(pil_image.size, cv.IPL_DEPTH_8U, 3)
    cv.SetData(cv_im, pil_image.tostring())
    faces = cv.HaarDetectObjects(cv_im, facial_features, storage)
    # faces includes a `neighbors` field that we aren't going to use here
    return [f[0] for f in faces]


def top_face_top(faces):
    coords = [f[1] for f in faces]
    # Top left corner is 0,0 so we need the min for highest face
    return min(coords)


def bottom_face_bottom(faces):
    # Top left corner is 0,0 so we need the max for lowest face. Also add the
    # height of the faces so that we get the bottom of it
    coords = [f[1] + f[3] for f in faces]
    return max(coords)


def draw_faces(image_, faces):
    "Draw a rectangle around each face discovered"
    image = image_.copy()
    drawable = ImageDraw.Draw(image)

    for x, y, w, h in faces:
        absolute_coords = (x, y, x + w, y + h)

        drawable.rectangle(absolute_coords)
    return image


if __name__ == '__main__':
    sys.exit(main())