Image metadata provides useful information for applications such as image retrieval, content description and geolocation. However, it is relatively easy to tamper with this data using metadata manipulation tools. In this paper, we address the issue of authenticating the time of capture of an image based on its geolocation information. By utilizing a novel two-stage shadow detection process, we are able to estimate the azimuthal direction of the sun in an image and compare it against a calculated theoretical value to establish the authenticity of the time of capture. Results are provided which show that our technique is able to estimate the sun azimuth, and thereby the time of capture, with good accuracy. Additional results pertaining to date verification and camera direction estimation are also provided.

Image manipulation has become commonplace with growing easy access to powerful computing abilities. One of the most common types of image forgeries is the copy-paste forgery, wherein a region from an image is replaced with another region from the same image. Most prior approaches to finding identical regions suffer from their inability to detect the cloned region when it has been subjected to a geometric transformation. In this paper, we propose a novel technique based on transform-invariant features. These are obtained by using the features from the MPEG-7 image signature tools. Results are provided which show the efficacy of this technique in detecting copy-paste forgeries, with translation, scaling, rotation, flipping, lossy compression, noise addition and blurring. We obtain a feature matching accuracy in excess of 90% across postprocessing operations and are able to detect the cloned regions with a high true positive rate and lower false positive rate than the state of the art.

Image metadata provides useful information for applications such as image retrieval, content description and geolocation. However, it is relatively easy to tamper with this data using metadata manipulation tools. In this paper, we address the issue of authenticating the time of capture of an image based on its geolocation information. By utilizing a novel two-stage shadow detection process, we are able to estimate the azimuthal direction of the sun in an image and compare it against a calculated theoretical value to establish the authenticity of the time of capture. In case of ambiguous direction estimation from shadows, we improve an existing method using cues from vertical surfaces to resolve the ambiguity. The sky is used as a cue to perform verification in difficult scenarios such as the absence of shadows. Results are provided which show that our technique is able to estimate the sun azimuth with good accuracy. Additional results pertaining to date verification and camera direction estimation are also provided.

An efficient self-configurable systolic architecture is proposed in this paper for very large scale integration implementation of a face recognition system. The proposed system applies principal component neural network (PCNN) with generalized Hebbian learning for extracting eigenfaces from the face database. It demonstrates a recognition performance of more than 85% when evaluated on the benchmark Yale and FRGC databases containing images with varying illumination and expression. Unlike the existing face recognition systems, the proposed approach not only recognizes the faces using computed eigenfaces, but also updates eigenfaces automatically whenever the face database changes. The challenge, however, lies in hardware realization of the PCNN-based face recognition system. In the presence of computation-intensive steps of varying nature, it is not straightforward to map the overall computation to a single systolic architecture. A primary contribution of this paper from the architecture point of view is an optimized mapping of fine-grained systolized signal flow graphs (SFGs) for each individual step of the algorithm on to a single self-configurable linear systolic array by appropriate merging of the computations pertaining to different nodes of different SFGs. The architecture has the flexibility of processing face images and databases of any size and it is easily scalable with the number of eigenfaces to be computed. The proposed PCNN-based systolic face recognition system has been implemented and evaluated on a Xilinx ML403 evaluation platform with Virtex-4 XC4VFX12 FPGA. The FPGA-based design for a reasonably large-sized face database can process more than 400 faces in a video image frame which is fast enough for video surveillance in busy public places and sensitive locations.

The widespread availability of photo manipulation software has made it unprecedentedly easy to manipulate images for malicious purposes. Image splicing is one such form of tampering. In recent years, researchers have proposed various methods for detecting such splicing. In this paper, we present a novel method of detecting splicing in images, using discrepancies in motion blur. We use motion blur estimation through image gradients in order to detect inconsistencies between the spliced region and the rest of the image. We also develop a new measure to assist in inconsistent region segmentation in images that contain small amounts of motion blur. Experimental results show that our technique provides good segmentation of regions with inconsistent motion blur. We also provide quantitative comparisons with other existing blur-based techniques over a database of images. It is seen that our technique gives significantly better detection results.

Fake images have become ubiquitous today in society, as they can be created at the expense of a few clicks using photo manipulation software. The detection of splicing in images has received research attention in recent years. In this paper, we present a novel method of detecting splicing in images, using discrepancies in motion blur. We use motion blur estimation through spectral matting in order to detect inconsistencies between the spliced region and the rest of the image. Experimental results show the effectiveness of our technique. Comparison studies show that this technique performs better than existing blur-based techniques when applied to motion blur.

Human face recognition is considered to be one of the toughest problems in the domain of pattern recognition. The variations in face images due to differing expression, pose and illumination are some of the key issues to be addressed in developing a face recognition system. In this paper, a new measure called gray Hausdorff distance (denoted by Hpg) is proposed to compare the gray images of faces directly. An efficient algorithm for computation of the new measure is presented. The computation time is linear in the size of the image. The performance of this measure is evaluated on benchmark face databases. The face recognition system based on the new measure is found to be robust to pose and expression variations, as well as to slight variation in illumination. Comparison studies show that the proposed measure performs better than the existing ones in most cases

In this paper, we present a new algorithm to construct a discrete Voronoi diagram based on the Euclidean distance metric in a binary image. The algorithm has linear time complexity and is suited to very large-scale integration (VLSI) implementation due to the use of local neighborhood calculations on reduced bit-width data. A cellular architecture for construction of the diagram is proposed. The proposed architecture has been implemented in VLSI using 0.35 micron 2-poly 3-metal layer complementary metal-oxide-semiconductor technology, and the dimensions of the chip are 3.16 mm/spl times/3.16 mm, with the maximum frequency of operation being 50 MHz.

This paper presents a parallel algorithm for computing the Voronoi diagram for collision-free path planning of robots. The algorithm constructs the Voronoi diagram on a binary image of the workspace and is based on the Euclidean distance function. It has been implemented in ANSI-C on an HP workstation. The algorithm is suitable for VLSI implementation in a cellular architecture. Such an architecture is described. Comparisons with an existing algorithm are also presented.

This paper presents a new algorithm to extract the skeleton and its Euclidean distance values from a binary image. A VLSI implementation of the algorithm in a locally connected cellular array is also given. The algorithm runs in O(n) time for an image of size n/spl times/n. The extracted skeleton reconstructs the objects in the image exactly.