Image Alignment

Given a fleet of robots, automatic estimation of the relative poses between them could be inaccurate in specific environments. We propose a framework composed by the fleet of robots with embedded stereoscopic cameras providing 2D and 3D images of the scene, a human coordinator and a Human-Machine Interface. We suppose auto localising each robot through GPS or landmarks is not possible. 3D-images are used to automatically align them and deduce the relative position between robots. 2Dimages are used to reduce the alignment error in an interactive manner. A human visualises both 2D-images and the current automatic alignment, and imposes a new alignment through the Human-Machine Interface. Since the information is shared through the whole fleet, robots can deduce the position of other ones that do not visualise the same scene. Practical evaluation shows that in situations where there is a large difference between images, the interactive processes are crucial to achieve an acceptable result.

Preclinical PET/CT is a well-established noninvasive imaging tool for studying disease development/progression and the development of novel radiotracers and pharmaceuticals for clinical applications. Despite this pivotal role, standardization of preclinical PET/CT protocols, including CT absorbed dose guidelines, is essentially nonexistent. This study (1) quantitatively assesses the variability of current preclinical PET/CT acquisition and reconstruction protocols routinely used across multiple centers and scanners; and (2) proposes acquisition and reconstruction PET/CT protocols for standardization of multicenter data, optimized for routine scanning in the preclinical PET/CT laboratory. Methods: Five different commercial preclinical PET/CT scanners in Europe and the United States were enrolled. Seven different PET/CT phantoms were used for evaluating biases on default/general scanner protocols, followed by developing standardized protocols. PET, CT, and absorbed dose biases were as...

This paper introduces an algorithm for the registration of rotated and translated volumes using the three-dimensional (3-D) pseudopolar Fourier transform, which accurately computes the Fourier transform of the registered volumes on a near-spherical 3-D domain without using interpolation. We propose a three-step procedure. The first step estimates the rotation axis. The second step computes the planar rotation relative to the rotation axis. The third step recovers the translational displacement. The rotation estimation is based on Euler's theorem, which allows one to represent a 3-D rotation as a planar rotation around a 3-D rotation axis. This axis is accurately recovered by the 3-D pseudopolar Fourier transform using radial integrations. The residual planar rotation is computed by an extension of the angular difference function to cylindrical motion. Experimental results show that the algorithm is accurate and robust to noise

%0"1 "% "()%0% " 2 % " 30 4 2 "( ¨0)"© 53¨6 #"%7 8 2 #"" ' #0% ! #"%%7 ¨© & 0%%"( "# 0( © 1% ! #"7 ¨© $ ' 0"1 "%"0# "# ("%0"1 "% "()%0% ("%0"17 ¨© %" & "% % & $ ' ' ""$( © % "% " 2 )"% !" 07 9"% %" & $ ' #% "%1 "% @ 00 "(( ("%0"17 ¨© %" & #(" #("% ##% 5A336 % & $ ' ' ""$( "%1 ©)© ©"% B7CD "(( ("%0"17 ¨© ©)© # %" & "% % & $ ' ' ""$( !" @7E 00 "% E7F 00 # ' ( "% !" $ ' © (""( 0"( #% (7 ¨© (! ' ""$( % © 0%"% " 2 !" % © 0"( 0#©"%#"( " 2

Since the single-channel framework of image restoration possesses serious conceptual and numerical problems, in this paper we present a general concept of image signal recovering. We also propose a space-variant restoration method using sliding spectral transforms. To provide image processing in real time, fast recursive algorithm for computing the sliding sinusoidal transform is utilized. Computer simulation results using a real image are provided and discussed.

Non uniform filtering is important for many image processing algorithms. However, for large kernel sizes the filtering can become computationally expensive. We introduce cosine integral images (CII) which represent a large set of spatial and range filters, based on their frequency decomposition. The filtering requires a constant number of operations per image pixel, independent of filter size. We make use of CII to compute the Gabor filters, whose complexity is a constant O(1) operations per image pixel. We also improve previous constant time approximations of spatial Gaussian smoothing and bilateral filtering.

In this paper, a wavelet based method is proposed to estimate the blur in an image using information contained in the image itself. We look at the sharpness of the sharpest edges in the blurred image, which contain information about the blurring. Specifically, a smoothness measure, the Lipschitz exponent, is computed for these sharpest edges. When one assumes that the blur can be described by one parameter (eg, blur from a gaussian PSF can be described by the variance of the gaussian, defocus blur can be ...

This paper proposes an original approach, based on navigation data analysis, for underwater mosaic creation from videos recorded by an AUV (Autonomous Underwater Vehicle). Two methods are presented: one uses only the AUV navigation data for image merging and the other also uses some techniques of feature extraction and image matching. The process of image matching is described and some improvements are proposed. A lighting correction procedure is also implemented in both cases to improve blending quality. Besides the mosaicking methods, a technique for trajectory refinement is proposed to handle the inherent uncertainties of navigation data and enhance the mosaic reliability. This technique is based on the analysis of overlapping images and it uses the interval algebra to compare proprioceptive and exteroceptive data. The mosaicking methods were tested in two missions carried out in the Mediterranean Sea. All the mosaics created, some additional images and videos are available as multimedia supporting materials.

Recognizing human facial expressions continuously is useful since it has many potential applications. We have implemented a continuous facial expression recognition system using Active Appearance Model (AAM). AAM has been widely used in face tracking, face recognition, and object recognition tasks. In this study, we adopt an independent AAM using Inverse Compositional Image Alignment method. The evaluation of this system carried out with the standard Cohn-Kanade facial expression database. Result shows that it could useful for many applications.

This paper proposes a new method to perform real-time face pose estimation for ±90 • yaw rotations and under low light conditions. The algorithm works on the basis of a completely automatic and run-time incremental 3D face modelling. The model is initially made up upon a set of 3D points derived from stereo grey-scale images. As new areas of the subject face appear to the cameras, new 3D points are automatically added to complete the model. In this way, we can estimate the pose for a wide range of rotation angles, where typically 3D frontal points are occluded. We propose a new feature re-registering technique which combines views of both cameras of the stereo rig in a smart way in order to perform a fast and robust tracking for the full range of yaw rotations. The Levenberg-Marquardt algorithm is used to recover the pose and a RANSAC framework rejects incorrectly tracked points. The model is continuously optimised in a Bundle Adjustment process that reduces the accumulated error on the 3D reconstruction. The intended application of this work is estimating the focus of attention of drivers in a simulator, which imposes challenging requirements. We validate our method on sequences recorded in a naturalistic truck simulator, on driving exercises designed by a team of psychologists.

Image hash regimes have been widely used for authenticating content, recovery of images and digital forensics. In this article we propose a new algorithm for image haunting (SSL) with the most stable key points and regional features, strong against various manipulation of content conservation, including multiple combinatorial manipulations. In order to extract most stable keypoint, the proposed algorithm combines the Speed Up Robust Features (SURF) with Saliency detection. The keyboards and characteristics of the local area are then combined in a hash vector. There is also a sperate secret key that is randomly given for the hash vector to prevent an attacker from shaping the image and the new hash value. The proposed hacking algorithm shows that similar or initial images, which have been individually manipulated, combined and even multiple manipulated contents, can be visently identified by experimental result.The probability of collision between hacks of various images is almost nil. Furthermore, the keydependent security assessment shows the proposed regime safe to allow an attacker without knowing the secret key not to forge or estimate the right havoc value.

Alignment of multimodality images is the process that attempts to find the geometric transformation overlapping at best the common part of two images. The process requires the definition of a similarity measure and a search strategy. In the literature, several studies have shown the ability and effectiveness of entropy-based similarity measures to compare multimodality images. However, the employed search strategies are based on some optimization schemes which require a good initial guess. A combinatorial optimization method is critically needed to develop an effective search strategy. Artificial Immune Systems (AIS S ) have been proposed as a powerful addition to the canon of meta-heuristics. In this paper, we describe a framework which combines the use of an entropy-based measure with an AIS-based search strategy. We show how AIS S have been tailored to explore efficiently the space of transformations. Experimental results are very encouraging and show the feasibility and effectiveness of the proposed approach.

Flexible macromolecules pose special difficulties for structure determination by crystallography or NMR. Progress can be made by electron microscopy, but electron cryo-microscopy of unstained, hydrated specimens is limited to larger macromolecules because of the inherently low signal-to-noise ratio. For three-dimensional structure determination, the single particles must be invariant in structure. Here, we describe how we have used negative staining and single-particle image processing techniques to explore the structure and flexibility of single molecules of two motor proteins: myosin and dynein. Critical for the success of negative staining is a hydrophilic, thin carbon film, because it produces a low noise background around each molecule, and stabilises the molecule against damage by the stain. The strategy adopted for single-particle image processing exploits the flexibility available within the SPIDER software suite. We illustrate the benefits of successive rounds of image alignment and classification, and the use of whole molecule averages and movies to analyse and display both structure and flexibility within the dynein motor.

We present a new integrated tool, DERMA, which allows to measure and assess the time evolution of chronic wounds. A laser triangulation 3D scanner is used to acquire the wound geometry with high precision and to capture an RGB image aligned to the geometry. DERMA provides a single and uniform interface to: manage patient data, 3D scanning of the lesion region, and to perform different kinds of measurements and comparisons: geometric (on the 3D model) and colorimetric (on the image). All acquired data (3D geometries and ...

Imaging and Image sensors is a field that is continuously evolving. There are new products coming into the market every day. Some of these have very severe Size, Weight and Power constraints whereas other devices have to handle very high computational loads. Some require both these conditions to be met simultaneously. Current imaging architectures and digital image processing solutions will not be able to meet these ever increasing demands. There is a need to develop novel imaging architectures and image processing solutions to address these requirements. In this work we propose analog signal processing as a solution to this problem. The analog processor is not suggested as a replacement to a digital processor but it will be used as an augmentation device which works in parallel with the digital processor, making the system faster and more efficient. In order to show the merits of analog processing the highly computational Normalized Cross Correlation algorithm is implemented. We propose two novel modifications to the algorithm and a new imaging architecture which, significantly reduces the computation time.

The Argentera Massif (French-Italian Alps), with its uniform lithology, was selected to evaluate how known Plio-Pleistocene tectonics have conditioned the drainage network geometry. The drainage network was automatically derived and ordered from a 10 m-resolution DEM. On hillshade images, alignments of morphological features were identified. The Massif was subdivided into 22 domains of 50 km 2 within which the directions of every river channel segment and the direction of the aligned morphological features were compared and contrasted with the strike of tectonic structures measured in the field. Results suggest that the Argentera drainage system is variously controlled by recent tectonics, depending on the Massif sector taken into account. In the NW sector, the vertical uplift is less because the strain has been accommodated in an oblique direction along a lateral thrust. In the SE sector, strain in a predominantly vertical direction along a frontal thrust has resulted in a major vertical displacement. Accordingly, the NW sector is characterized by (i) a strong geometric relationship between the main tectonic structures and the directions of river channels, (ii) longitudinal main rivers bordering the Massif, and (iii) a general trellis pattern within the domains. In the SE sector, the prolonged uplift has forced an original longitudinal drainage system to develop as a transverse system. This change has occurred by means of fluvial captures that have been identified by the presence of windgaps, fluvial elbows and knickpoints. At the domain scale, intense uplift of the SE sector has prompted the drainage pattern to evolve as a dendritic type with no clear influence of structure in the channel orientations.

Parameterized Appearance Models (PAMs) (e.g. eigentracking, active appearance models, morphable models) use Principal Component Analysis (PCA) to model the shape and appearance of objects in images. Given a new image with an unknown appearance/shape configuration, PAMs can detect and track the object by optimizing the model's parameters that best match the image. While PAMs have numerous advantages for image registration relative to alternative approaches, they suffer from two major limitations: First, PCA cannot model non-linear structure in the data. Second, learning PAMs requires precise manually labeled training data. This paper proposes Parameterized Kernel Principal Component Analysis (PKPCA), an extension of PAMs that uses Kernel PCA (KPCA) for learning a non-linear appearance model invariant to rigid and/or non-rigid deformations. We demonstrate improved performance in supervised and unsupervised image registration, and present a novel application to improve the quality of manual landmarks in faces. In addition, we suggest a clean and effective matrix formulation for PKPCA.

Active Appearance Models (AAMs) have been extensively used for face alignment during the last 20 years. While AAMs have numerous advantages relative to alternate approaches, they suffer from two major drawbacks: (i) AAMs are especially prone to local minima in the fitting process; (ii) few if any of the local minima of the cost function correspond to acceptable solutions. To minimize these problems, this paper proposes a method to learn the fitting cost function that explicitly optimizes that the local minima occur at and only at the places corresponding to the correct fitting parameters. The paper explores two methods to parameterize the cost function: pixel weighting and subspace learning. Experiments on synthetic and real data show the effectiveness of our approach for face alignment.

Image alignment has been a long standing problem in computer vision. Parameterized Appearance Models (PAMs) such as the Lucas-Kanade method, Eigentracking, and Active Appearance Models are commonly used to align images with respect to a template or to a previously learned model. While PAMs have numerous advantages relative to alternate approaches, they have at least two drawbacks. First, they are especially prone to local minima in the registration process. Second, often few, if any, of the local minima of the cost function correspond to acceptable solutions. To overcome these problems, this paper proposes a method to learn a metric for PAMs that explicitly optimizes that local minima occur at and only at the places corresponding to the correct fitting parameters. To the best of our knowledge, this is the first paper to address the problem of learning a metric to explicitly model local properties of the PAMs' error surface. Synthetic and real examples show improvement in alignment performance in comparison with traditional approaches. In addition, we show how the proposed criteria for a good metric can be used to select good features to track.

Since the precise modeling of reflection is a difficult task, most feature points trackers assume that objects are lambertian and that no lighting change occurs. To some extent, a few approaches answer these issues by computing an affine photometric model or by achieving a photometric normalization. Through a study based on specular reflection models, we explain explicitly the assumptions on which these techniques are based. Then we propose a tracker that compensates for specular highlights and lighting variations more efficiently when small windows of interest are considered. Experimental results on image sequences prove the robustness and the accuracy of this technique in comparison with the existing trackers. Moreover, the computation time of the tracking is not significantly increased.

Since Roentgen's discovery of X-rays, just over a century ago, the vast majority of radiographs have been acquired and interpreted on the basis of absorption contrast. The recent development of X-ray phase contrast imaging techniques opens up a new opportunity especially for weakly-absorbing objects. Spatial and temporal coherence of X-ray synchrotron radiation permits to realize phase contrast imaging with ideal conditions. But it is possible to set up phase contrast imaging in laboratory. Laboratory phase contrast imaging can provide new solutions in an industrial context. The aim of this paper is to present the different solutions to perform phase contrast imaging in laboratory: − classical micro focus X-ray tube, − X-ray ultra microscope (XuM) developed by an Australian team − table top synchrotron developed by a Japanese team, − multilayer, − polycapillary.

We present a new approach for aligning families of 2D gels. Instead of choosing one of the gels as reference and performing pairwise alignment, we construct an ideal gel that is representative for the entire family and obtain a set of piecewise affine transformations that optimally align each gel of the family to the ideal gel. The coefficients defining the transformations as well as the ideal landmarks are obtained as the solution of a large-scale quadratic programming problem that can be solved efficiently by interiorpoint methods.

This paper presents a new computational framework for automatic foot classification from digital plantar pressure images. It classifies the foot as left or right and simultaneously calculates two well-known footprint indices: the Cavanagh's arch index and the modified arch index. The accuracy of the framework was evaluated using a set of plantar pressure images from two common pedobarographic devices. The results were outstanding, since all feet under analysis were correctly classified as left or right and no significant differences were observed between the footprint indices calculated using the computational solution and the traditional manual method. The robustness of the proposed framework to arbitrary foot orientations and to the acquisition device was also tested and confirmed.

Primarily author has already done one fundamental paper work on image
clustering and segmentation but here in this paper author has continued that same type of
work on clustered and segmented images as a mode of comparative study for author has
chosen three different parameters like mean square error, peak SNR and dimensions of
images (length, width, height). The author has all three parametric methods on one
particular to justify the comparison. So this paper is a cumulative case of a comparative
study for which author has chosen the above mentioned parameters to justify the best
results of the clustered and segmented images.

Detection of variability in agricultural fields depends on the spatial scale of the observed variable. Plant water status can be evaluated using thermal IR images that can provide valuable information on the water status, whereas visible RGB images can provide detailed information on the plants' color, which is not a good indicator of the water status. The informative mode (thermal IR images) has coarse resolution, as opposed to the excessive resolution of the less informative mode (visible RGB). In the present study, we present a method to enhance the information obtained from the thermal IR mode, by combining information from the visible RGB mode. We propose to un-mix the temperature of objects in the thermal images based on the information extracted from the high resolution RGB image.

Face recognition accuracy is affected by many factors. This paper studies one of the factors, and provides a reliable image alignment for face recognition. For this purpose, a performance metric is extracted from an analysis of face recognition similarity scores. The metric varies with face alignment, and has a relationship with the actual recognition accuracy. Our method adjusts face alignment online by selecting an alignment candidate corresponding to the largest performance metric. The experimental results show that the presented method can improve the accuracy and robustness of current face recognition systems.

Popular image registration techniques A relatively few techniques have been developed and widely accepted for image registration, with each tuned to work best for a different set observations. The most common algorithms are: • Crosscorrelation: This technique works best for aligning images containing primarily indistinct objects such as nebulae, galaxies, unresolved clusters and so on.

Imaging and Image sensors is a field that is continuously evolving. There are new products coming into the market every day. Some of these have very severe Size, Weight and Power constraints whereas other devices have to handle very high computational loads. Some require both these conditions to be met simultaneously. Current imaging architectures and digital image processing solutions will not be able to meet these ever increasing demands. There is a need to develop novel imaging architectures and image processing solutions to address these requirements. In this work we propose analog signal processing as a solution to this problem. The analog processor is not suggested as a replacement to a digital processor but it will be used as an augmentation device which works in parallel with the digital processor, making the system faster and more efficient. In order to show the merits of analog processing the highly computational Normalized Cross Correlation algorithm is implemented. We propose two novel modifications to the algorithm and a new imaging architecture which, significantly reduces the computation time.

The precise orientation-estimation of humans relative to the pose of a monocular camera system is a challenging task due to the general aspects of camera calibration and the deformable nature of a human body in motion. Thus, novel approaches of Deep Learning for precise object pose-estimation in robotics are hard to adapt to human body analysis. In this work, a hybrid approach for the accurate estimation of a human body rotation relative to a camera system is presented, thereby significantly improving results derived from poseNet by applying analysis of optical flow in a frame to frame comparison. The human body in-place rotating in T-pose is thereby aligned in the center, applying object tracking methods to compensate for translations of the body movement. After 2D skeleton extraction, the optical flow is calculated for a region of interest (ROI) area aligned relative to the vertical skeleton joint representing the spine and compared frame by frame. To evaluate the eligibility of the clothing as a fundament for good feature, the local pixel homogeneity is taken into consideration to restrict the optical flow to heterogeneous regions with distinctive features like imprint patterns, buttons or buckles besides local illumination changes. Based on the mean optical flow with a coarse approximation of the axial body shape as ellipsis, an accuracy between 0.1° and 2.0° by a target rotation of 10° for orientation-estimation is achieved on a frame-toframe comparison evaluated and validated on both, Computer Generated Imagery (CGI) renderings and real-world videos of people wearing clothing of varying feature appropriateness.

Not many decades ago, Pattern Recognition and Image Analysis (PR&IA) addressed with simple tasks applying shallow models. But things are changing, and quickly. Then, this highly dynamic discipline has been expanding greatly, also helped by the emergence of newer application such as in robotics, biometrics or multimedia systems. Just now, PR&IA tasks run the complete gamut: from preprogramed works to the stimulating challenge of getting computers to learn as they go. At their most formidable, PR&IA tasks require computers to look, interpret and report back. We are at a transition point where PR&AI are suddenly at the forefront. Progress has come about thanks in part to steady advance in the technologies needed to help machines understand visual data, including machine learning and data mining techniques. The papers included in this special issue provide a snapshot of image analysis and pattern recognition research today. They are the very best of the 6th Iberian Conference on Pattern Recognition and Image Analysis (IbPRIA 2013), held on 5-7 June, 2013 in Madeira, Portugal. IbPRIA 2013 attracted 181 papers from 34 different countries. After the reviewing process, 105 papers were accepted for presentation in the conference. A selection of the best scored and presented at the conference was invited to submit to this special issue a substantially extended and revised version of the conference paper and the resulting manuscripts were sent out for full review. The process, including required revisions, was in accordance with the standing editorial policy of Neurocomputing, resulting in the final versions of the ten papers accepted and appearing in this special issue.

As we know the image can be processed with the help of different types of coding for example mat-lab. Here in this paper we are primarily focusing on some common filtering methodologies [5] related to image contour in clustered conditions. For filtering purpose in this paper we have used three different filtering technologies such as prewitt [3], sobel [3], canny [3] filtering. But on the other hand we have used both colored [1] and non-colored [3] images for clustering operations. Our main aim in this paper to show variations of signal to noise ratios for the colored and non-colored contour images with and without filtering. As per my request study the discussion of results very carefully to realize the deeper meaning of the journal [4].

This work considers sampled data of continuous-domain Gaussian processes. We derive a maximum-likelihood estimator for identifying autoregressive moving average parameters while incorporating the sampling process into the problem formulation. The proposed identification approach introduces exponential models for both the continuous and the sampled processes. We construct a likelihood function from a digitally-filtered version of the available data which is asymptotically exact. This function has several local minima that originate from aliasing, plus a global minimum that corresponds to the maximum-likelihood estimator. We further compare the performance of the proposed algorithm with other currently available methods.

A new approach to align an image of a medical object with a given prototype is proposed. Visual appearance of the images, after equalizing their signals, is modelled with a new Markov-Gibbs random field with pairwise interaction model. Similarity to the prototype is measured by a Gibbs energy of signal co-occurrences in a characteristic subset of pixel pairs derived automatically from the prototype. An object is aligned by an affine transformation maximizing the similarity by using an automatic initialization followed by gradient search. Experiments confirm that our approach aligns complex objects better than popular conventional algorithms.

One of the final steps in a display production line is the image alignment that includes the visual adjustment of the geometric parameters and the color of the image. Mea-surement of geometric characteristics using machine vision is a necessary function in the automatic alignment of dis-plays ’ image in the factory. A critical part in the measure-ment of the geometric attributes is to precisely locate a test pattern position on the display screen. In this paper we in-troduce novel patterns as fuzzy test patterns and present a novel algorithm to precisely locate the fuzzy test pattern in captured images of the display screen. We experimentally show that the application of the proposed fuzzy test pattern and its associated locating algorithm increases the precision and robustness of the geometric measurements of a display like a TV display. The use of this new measurement method in an auto-alignment system increases the adjustment accu-racy, improves the reliability of the alignment s...