Face Recognition across Non-Uniform Motion Blur, Illumination, and Pose
Existing methods for performing face recognition in the presence of blur are based on the convolution model and cannot handle non-uniform blurring situations that frequently arise from tilts and rotations in hand-held cameras. In this paper, we propose a methodology for face recognition in the presence of space-varying motion blur comprising of arbitrarily-shaped kernels. We model the blurred face as a convex combination of geometrically transformed instances of the focused gallery face, and show that the set of all images obtained by non-uniformly blurring a given image forms a convex set. We first propose a non-uniform blur-robust algorithm by making use of the assumption of a sparse camera trajectory in the camera motion space to build an energy function with l1-norm constraint on the camera motion. The framework is then extended to handle illumination variations by exploiting the fact that the set of all images obtained from a face image by non-uniform blurring and changing the illumination forms a bi-convex set. Finally, we propose an elegant extension to also account for variations in pose.
In common, blurring due to camera shake is modelled as convolution with single blur kernel and the blur is uniform across the image this case is considered as space variant blur frequently in hand held cameras. Restoration of non-uniform blur is based local space invariant approximation and a recent methods for image restoration is motion-blurred image as an average of projectively transformed images.
Approaches to face recognition from blurred images can be broadly classified into four categories. (i) Deblurring-based in which the probe image is first deblurred and then used for recognition. However, deblurring artifacts are a major source of error especially for moderate to heavy blurs. (ii) Joint deblurring and recognition, the flip-side of which is computational complexity. (iii) Deriving blur-invariant features for recognition. But these are effective only for mild blurs. (iv) The direct recognition approach in which reblurred versions from the gallery are compared with the blurred probe image.
Patel et al.have proposed a dictionary-based approach to recognizing faces across illumination and pose.
DISADVANTAGES OF EXISTING SYSTEM:
Deblurring artifacts are a major source of error especially for moderate to heavy blurs.
The flipside of which is computational complexity in joint and recognition
In deriving blur-invariant features is only effective for mild blurs.
Although in subspace learning approach it is difficult to solve the problems like blur,pose,illumination etc..
A dictionary based approach, sparse minimization technique for recognizing faces with similar principles and offers robustness to alignment and pose. But these works do not deal with blurred images.
In this paper we are propose a face recognition that is robust to non-uniform i.e space varying motion blur arising from relative motion between the camera and the subject.
We will assume that only a single gallery image is available. The camera transformations can range from in-plane translations and rotations to out-of-plane translations, out-of-plane rotations and even general 6D motion. Observe that the blur on the faces can be significantly non-uniform.
The simple yet restrictive convolution model fails to explain this blur and a space-varying formulation becomes necessary.
We showed that the set of all images using the TSF model is a convex set given by the convex hull of warped versions of the image.
We develop our basic non-uniform motion blur (NU-MOB)-robust face recognition algorithm based on the TSF (Transformation Spread Function) model.
ADVANTAGES OF PROPOSED SYSTEM:
This proposed method of recognition allows us to circumvent the challenging and ill-posed problem of single image blind-deblurring.
It efficiently deals with blurred images.
This is the first attempt to systematically address face recognition under (i) non-uniform motion blur and (ii) the combined effects of blur, illumination and pose.
We prove that the set of all images obtained by non-uniformly blurring a given image forms a convex set. We also show that the set of all images obtained from a face image by non-uniform blurring and change of illumination forms a bi-convex set.
We extend our method to non-frontal situations by transforming the gallery to a new pose.
We propose a multi-scale implementation that is efficient both in terms of computation as well as memory usage
System : Pentium IV 2.4 GHz.
Hard Disk : 40 GB.
Floppy Drive : 44 Mb.
Monitor : 15 VGA Colour.
Ram : 512 Mb.
Operating system : Windows XP/7.
Coding Language : MATLAB
Tool : MATLAB R2013A
Abhijith Punnappurath, Ambasamudram Narayanan Rajagopalan, Senior Member, IEEE, Sima Taheri, Student Member, IEEE, Rama Chellappa, Fellow, IEEE, and Guna Seetharaman, Fellow, IEEE, “Face Recognition Across Non-Uniform Motion Blur, Illumination, and Pose”, IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 24, NO. 7, JULY 2015.