Face Detection and Verification Using Genetic Search

Abstract

We consider the problem of searching for the face of a particular individual in a twodimensional intensity image. This problem has many potential applications such as locating a person in a crowd using images obtained by surveillance cameras. There are two steps in solving this problem: first, face regions must be extracted from the image(s) (face detection) and second, candidate faces must be compared against the face of interest (face verification). Without any a-priori knowledge about the location and size of a face in an image, every possible image location and face size should be considered, leading to a very large search space. In this paper, we propose using Genetic Algorithms (GAs) for searching the image efficiently. Specifically, we use GAs to find image sub-windows that contain faces and in particular, the face of interest. Each sub-window is evaluated using a fitness function containing two terms: the first term favors sub-windows containing faces while the second term favors sub-windows containing faces similar to the face of interest. Both terms have been derived using the theory of eigenspaces. A set of increasingly complex scenes demonstrate the performance of the proposed genetic-search approach.

Front and Rear Vehicle Detection Using Hypothesis Generation and Verification

“…In the past, it has been used to solve different problems such as object detection [40], face recognition [55,56], vehicle detection [4], image annotation [57], gender classification [58] and target recognition [59].…”

Face recognition by fusing thermal infrared and visible imagery

“…They were designed to efficiently search large, non-linear, poorly understood search spaces. In the past, GAs have been used in target recognition [37], object recognition [38], face detection/verification [39,40], and feature selection [41][42][43][44]. Goldberg [45] provides a nice introduction to GAs and the reader is referred to this source for further information.…”

Object detection using feature subset selection

“…In the past, they have been used to solve various problems including target recognition [44], object recognition [45,46], face recognition [47], and face detection/veriÿcation [48]. This section contains a brief summary of the fundamentals of GAs.…”

“…Each subimage in the training and test sets was scaled to 32 × 32 and preprocessed to account for di erent lighting conditions and contrast followed the method suggested in Ref. [48].…”

“…For testing, we used a ÿxed set of 268 face and non-face subimages which were also extracted from disjoint set of images from the CMU face detection data set. Each subimage in the training and test sets was scaled to 32 × 32 and preprocessed to account for di erent lighting conditions and contrast [48]. To evaluate the performance of the proposed approach, we used a three-fold cross-validation procedure, splitting the training dataset randomly three times by keeping 84% of the face subimages and 84% of the non-face subimages (i.e., 516 vehicle subimages and 516 non-face subimages) for training.…”