Skriv en iPhone-applikasjon

Write a library that can analyze a picture or video stream for face, smile and eyes for each person in the image. The algorithm should support faces which in tilted or rolled positions too. The output of the library provides the following information for each face in the image: - Mouth position and corner values - Right eye position and corner values - Left eye position and corner values - Face position and width These should be provided both inside the original image (so it can be drawn on top of the image), and as a normalised (vertically aligned) image. With these points, we should be able to detect the width of the mouth, the width of the face, and see if the corner of the mouth will match the corner of the eyes. The library should be possible to use in applications written in Obj-C and Swift. The library will be used in a commercial application without any licensing issues. The library should perform well, and be possible to use on both still images and in videos (e.g. while using the camera). Earlier investigations has shown that default facial detections using OpenCV and iOS Facial features API is not sufficient. Dlib and Landmark detection seems very promising...