Search

Estimating the displacements of intensity patterns between sequential frames is a very well-studied problem, which is usually referred to as optical flow estimation. The first assumption among many of the methods in the field is the brightness constancy during movements of pixels between frames. This assumption is proven to be not true in general, and therefore, the use of photometric invariant constraints has been studied in the past. One other solution can be sought by use of structural descriptors rather than pixels for estimating the optical flow. Unlike sparse feature detection/description techniques and since the problem of optical flow estimation tries to find a dense flow field, a dense structural representation of individual pixels and their neighbors is computed and then used for matching and optical flow estimation. Here, a comparative study is carried out by extending the framework of SIFT-flow to include more dense descriptors, and comprehensive comparisons are given. Overall, the work can be considered as a baseline for stimulating more interest in the use of dense descriptors for optical flow estimation. [ABSTRACT FROM AUTHOR]

Detecting the presence of contact lenses and their type helps increasing the reliability of iris-based authentication systems. We propose a machine-learning approach for this task, based on expressive local image descriptors. The image is first segmented to extract the iris and sclera regions, then scale-invariant local descriptors (SID) are computed densely on both areas, and summarized through the Bag-of-Features paradigm. Classification is based on a properly trained linear SVM. The major contributions of our proposal concern the segmentation algorithm, the use of information drawn from the sclera, and the use of non-rectified data to preserve local structures. A number of variants of the proposed method are investigated, working on different areas of the image, with alternative local descriptors, and with different encoding techniques. Eventually, results are compared with the state-of-the-art in the field. The experimental analysis, carried out on several publicly available datasets, shows that the proposed classification method based on a dense scale invariant descriptor outperforms all the reference techniques. [ABSTRACT FROM AUTHOR]

Estimating the displacements of intensity patterns between sequential frames is a very well-studied problem, which is usually referred to as optical flow estimation. The first assumption among many of the methods in the field is the brightness constancy during movements of pixels between frames. This assumption is proven to be not true in general, and therefore, the use of photometric invariant constraints has been studied in the past. One other solution can be sought by use of structural descriptors rather than pixels for estimating the optical flow. Unlike sparse feature detection/description techniques and since the problem of optical flow estimation tries to find a dense flow field, a dense structural representation of individual pixels and their neighbors is computed and then used for matching and optical flow estimation. Here, a comparative study is carried out by extending the framework of SIFT-flow to include more dense descriptors, and comprehensive comparisons are given. Overall, the work can be considered as a baseline for stimulating more interest in the use of dense descriptors for optical flow estimation.

One of the major challenges in the field of digital image watermarking is to authenticate the presence of watermark in the watermarked image even after it has been transformed intentionally or unintentionally. Transformation can be geometric-like rotation, scaling, and translation of image or may be due to any signal processing attack like noise corruption, compression, and cropping of image. There may also be some photometric changes, for example change in the brightness of watermarked image during transmission, due to which it becomes difficult to validate whether received image is watermarked or not. Illumination invariance property of Weber's descriptor has engrossed to use it in the proposed watermark authentication technique. Weber's descriptor is a descriptor based on two parameters of a pixel, differential excitation and orientation. These parameters are computed using the relative intensity value of neighbor pixels and current pixel. This descriptor remains the same even after intensity changes due to the contribution of all neighbor pixel's intensity in its computation. It is also known to be robust to scaling and rotation. Experimental results show that the proposed watermarking technique is able to authenticate the presence of watermark in the watermarked image even when it is distorted due to geometric and photometric attacks. In addition to this, it is found to be robust against noise, cropping, and compression attacks. [ABSTRACT FROM AUTHOR]

This paper presents a robust two-stage face recognition approach that combines the traits of global features in first stage and the local features in second stage. The global features are extracted from Zernike moments (ZMs) method that encompasses the useful characteristics of being invariant to image rotation, scale, and noise. The local features are obtained from the histogram-based Weber Law Descriptor (WLD) having tremendous qualities like invariance to scale, change in image intensities, rotation, and noise. The novelty of this paper is twofold: (1) an efficient approach is used for combining the global and local features which is based on the human psychology to trace and memorize the known persons, i.e., locate some similar faces from the overall appearance of different persons and later identify from this the specific individual on the basis of their interior differences like shape of eyes, nose, etc.; (2) a method is used for providing the weights to individual face patches in extraction of local features, which is based on the averaged discrimination competence of features within a patch. The performance of proposed two-stage face recognition approach is analyzed against some major hurdles of this system, i.e., illumination, expression, scale, pose, occlusion, and noise variations. The proposed method achieves the highest recognition rate of 98.0% and 94.1% on ORL and Yale databases, respectively. The experimental results on these well-known face databases demonstrate that the proposed method is highly robust to illumination variation and also generates superior results against other variations. [ABSTRACT FROM AUTHOR]