No abstract

The most frequently employed measure for performance characterization of local feature detectors is repeatability, but it has been observed that this does not necessarily mirror actual performance. In this letter, improved repeatability formulations are presented which correlate much better with the true performance of feature detectors. Comparative results for several state-of-the-art feature detectors are presented using these measures; it is found that Hessian-based detectors are generally superior at identifying features when images are subject to various geometric and photometric transformations. Introduction:The extraction of image features that are reasonably independent of scale, orientation and photometrical changes has long been an aim of the vision research community. The last decade has seen the development of a number of such operators, the best-known of which is SIFT [1-2] that incorporates both a detector and a descriptor. The evaluation of the performances of these detectors

When matching images for applications such as mosaicking and homography estimation, the distribution of features across the overlap region affects the accuracy of the result. This paper uses the spatial statistics of these features, measured by Ripley's K-function, to assess whether feature matches are clustered together or spread around the overlap region. A comparison of the performances of a dozen state-of-the-art feature detectors is then carried out using analysis of variance and a large image database. Results show that SFOP introduces significantly less aggregation than the other detectors tested. When the detectors are rank-ordered by this performance measure, the order is broadly similar to those obtained by other means, suggesting that the ordering reflects genuine performance differences. Experiments on stitching images into mosaics confirm that better coverage values yield better quality outputs.

IntroductionAugmented Reality (AR) combines real world imagery with synthetic content generated by a computer. The first comprehensive review of AR [1] identified a broad range of applications of this technology, including medicine, manufacturing and robot path planning. Subsequently, AR has been applied in cultural heritage [2] such as the reconstruction of ancient Olympia in Greece [3,4]. Although it is reported that AR enriches human perception [5] in general, the principal reason for performing AR reconstructions in cultural heritage is that the owners of sites are usually reticent to permit physical reconstructions in situ so that the archaeology remains undisturbed for future generations [3].Developments in multimedia technology facilitate the learning experience in cultural heritage [6] with the aid of improved user interaction methods. Developed models or virtual tours in reconstructions of archaeological sites (e.g. [7,8]) provide entertaining means of learning. However, ex situ reconstructions such as models and movies are difficult to visualize in the context of the archaeological remains. AR reconstructions can be produced in situ with minimal physical disturbance, an attractive property, even though they may take a significant time to develop [9].There are several forms that AR reconstructions may take, and the work reported here is directed towards a kind of 'historical mirror, ' in which virtual buildings are built around flat surfaces visible in the real world and human participants are clothed appropriately for the historical period. The literature presents examples of using Kinect for cultural heritage [10][11][12]. It was used as a 3D scanner in the work given in [10] and as a Abstract This paper explores the use of data from the Kinect sensor for performing augmented reality, with emphasis on cultural heritage applications. It is shown that the combination of depth and image correspondences from the Kinect can yield a reliable estimate of the location and pose of the camera, though noise from the depth sensor introduces an unpleasant jittering of the rendered view. Kalman filtering of the camera position was found to yield a much more stable view. Results show that the system is accurate enough for in situ augmented reality applications. Skeleton tracking using Kinect data allows the appearance of participants to be augmented, and together these facilitate the development of cultural heritage applications. Bostanci et al. Hum. Cent. Comput. Inf. Sci. (2015) Hum. Cent. Comput. Inf. Sci. (2015) 5:20 motion tracker to navigate in virtual reality reconstructions in [11,12]. User experience for large screens with Kinect in a cultural heritage setting, such as in an exhibition, has been demonstrated in [13]. This paper uses Kinect to establish 3D world and 2D image correspondences and, from them, determine camera pose. It then finds planar objects within the real world and augments them, the aim being to render the appearance in antiquity in front of real-world features. The advantage of thi...

The integral image, an intermediate image representation, has found extensive use in multi-scale local feature detection algorithms, such as Speeded-Up Robust Features (SURF), allowing fast computation of rectangular features at constant speed, independent of filter size. For resource-constrained real-time embedded vision systems, computation and storage of integral image presents several design challenges due to strict timing and hardware limitations. Although calculation of the integral image only consists of simple addition operations, the total number of operations is large owing to the generally large size of image data. Recursive equations allow substantial decrease in the number of operations but require calculation in a serial fashion. This paper presents two new hardware algorithms that are based on the decomposition of these recursive equations, allowing calculation of up to four integral image values in a row-parallel way without significantly increasing the number of operations. An efficient design strategy is also proposed for a parallel integral image computation unit to reduce the size of the required internal memory (nearly 35% for common HD video). Addressing the storage problem of integral image in embedded vision systems, the paper presents two algorithms which allow substantial decrease (at least 44.44%) in the memory requirements. Finally, the paper provides a case study that highlights the utility of the proposed architectures in embedded vision systems.