Abstract
In this thesis, we propose a new technique for compensating radial distortion of photos acquired with wide-angle lens by using facial features extracted from images. Our method utilizes a statistical facial feature model to recover radial distortion and the facial features are further used for adaptive cylindrical projection which will reduce perspective distortion near the image boundary. Our method has several advantages over the traditional methods. First, traditional calibration patterns, like man-made straight buildings, chessboards, or calibration cubes, are not required in our method. Secondly, the faces in the corrected photos by using the proposed method are better corrected than those by using the traditional methods. In addition, the proposed method produces photos with less perspective distortion compared to the traditional methods because our method includes an adaptive cylindrical projection. Our distortion correction system is composed of four components: offline training of the statistical face feature model, feature point extraction from distorted faces, estimation of radial distortion parameters and distortion compensation, and adaptive cylindrical projection. In order to estimate the distortion parameters, we propose an energy considering the fitness between the undistorted coordinates of the facial feature points extracted from the input image and the learned statistical facial feature model. Given the distortion parameters, the fitness is calculated by solving a linear least squares system. The distortion parameters that minimize the fitness function are searched in a hierarchical manner. Experimental results are shown to demonstrate the distortion reduction in the corrected images by using the proposed method.