The three screenshots are the final project's results. My main idea was to render a 3D cube based on perspective using eye detection and facial positioning, and as you can see, it works pretty well. I used OpenCV for the facial and eye detection, and the rest of the code I made myself. The camera first makes an assumption that your face is 5 inches wide, and then it determines how far away your face it using pixel measurements. It then calculates the offset of the face pixel-wise, and then it converts it to inches and once again. It projects the face's offset using the distance, and voila! it is able to detect where your face is offset from the camera with an accuracy of ±2 inches! In the screenshots, the left window is an OpenGL rendering of the face relative to the webcam. The transform of the face is the transform of the camera and the cube represents the webcam. If you want me to post the code, you can just ask me.