Image Alignment

Wepresenttheapplicationof differentnonlinearimagedeformationmodelsto thetaskof imagerecognition.Thedeformation models are especially suited for local changes as they often occur in the presence of image object variability. We show that, among the discussed models, there is one approach that combines simplicity of implementation, low-computational complexity, and highly competitive performance across various real-world image recognition tasks. We show experimentally that the model performs very well for

Find article (software manual) at:  http://www.iaees.org/publications/journals/np/articles/2020-5(1-2)/1-Zhang-Abstract.asp                                                                                                                                               

Download software at:                       
http://www.iaees.org/publications/software/                                       

In present study, I developed a powerful Matlab-based software package, imageProcAnal (Version 1.0), for image processing and analysis. Several modules were available for uses. Functions to resize, crop, rotate, dilate, pixelate and watermark images are included in Basic module. Adjustment of image edge, contrast and gamma factor is available in Adjustment module. In Denoising module, four elementary denoising methods and their joint uses are provided for image filtering and denoising. In Edge and Contour Detection module, five methods for detection of image edge and contour are provided. Also three methods for image sharpening are available in Sharpening module. Segmentation module includes three methods for image segmentation. Connected components recognition can be made in Object Recognition module. In Analysis module, Pearson correlation measure was provided to align images. Pearson correlations and p values among images can be obtained. Mutiple methods across different modules can be sequentially and jointly used to achieve ideal image. Finally, some examples for image processing can be found in Demo module. In addition to conventional methods, I proposed two new methods for image segmentation, i.e., TDOS (Two-Dimensional Ordered Segmentation) and MWAS (Moving Windows Averaging Segmentation). In TDOS, which is based on two-dimensional ordered cluster analysis, the pixels belonging to the same segment (class) are adjoint and are not separated by any of remaining segments. MWAS is based on moving windows averaging method for transect boundary detection. Algorithms of TDOS and MWAS were given, and full Matlab codes of MWAS were given.

Image alignment is without doubt the most crucial step in SAR Interferometry. Interferogram formation requires images to be coregistered with an accuracy of better than 1/8 pixel to avoid significant loss of phase coherence. Conventional interferometric precise coregistration methods for full-resolution SAR data (Single-Look Complex imagery, or SLC) are based on the cross-correlation of the SLC data, either in the original complex form or as squared amplitudes. Offset vectors in slant range and azimuth directions are computed on a large number of windows, according to the estimated correlation peaks. Then, a two-dimensional polynomial of a certain degree is usually chosen as warp function and the polynomial parameters are estimated through LMS fit from the shifts measured on the image windows. In case of rough topography and long baselines, the polynomial approximation for the warp function becomes inaccurate, leading to local misregistrations. Moreover, these effects increase with the spatial resolution and then with the sampling frequency of the sensor, as first results on TerraSAR-X interferometry confirm. An improved, DEM-assisted image coregistration procedure can be adopted for providing higher-order prediction of the offset vectors. Instead of estimating the shifts on a limited number of patches and using a polynomial approximation for the transformation, this approach computes pixel by pixel the correspondence between master and slave by using the orbital data and a reference DEM. This study assesses the performance of this approach with respect to the standard procedure. In particular, both analytical relationships and simulations will evaluate the impact of the finite vertical accuracy of the DEM on the final coregistration precision for different radar postings and relative positions of satellites. The two approaches are compared by processing real data at different carrier frequencies and using the interferometric coherence as quality figure.

This thesis concerns the use of spatial and tonal adaptivity in improving the resolution of aliased image sequences under scene or camera motion. Each of the five content chapters focuses on a different subtopic of super-resolution: image registration (chapter 2), image fusion (chapter 3 and 4), super-resolution restoration (chapter 5), and super-resolution synthesis (chapter 6).

Image alignment is one of the most widely used techniques in computer vision. Applications range from optical flow, tracking and layered motion, to mosaic construction, medical image registration, and face model fitting. The original image alignment algorithm was the Lucas-Kanade algorithm. Since then, numerous extensions have been made to it. In particular, Baker and Matthews recently proposed the inverse compositional algorithm, an efficient algorithm applicable to most 2D image alignment problems. In this report, we investigate whether the 2D inverse compositional algorithm can be generalized to 2.5D and 3D. By 3D we mean volumetric data consisting of a dense 3D array of voxels. By 2.5D we mean a surface in 3D represented by a collection of 3D surface points. We show that the inverse compositional algorithm is easily generalized to 3D. On the other hand, while algebraically it appears as though the 2.5D case may be treated similarly, doing so violates one of the assumptions in th...

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 ...

Aerial mapping is attracting more attention due to
the development in unmanned aerial vehicles and
their availability and also vast applications that
require a wide aerial photograph of a region in a
specific time. This paper concentrates on image
registration of aerial images. The proposed method
is able to sample automatically and align the sensed
images to form the final map. The results are
compared with satellite images that it shows a
reasonable performance with geometrically correct
registration.

In this paper, we seek to expand the use of direct methods in real-time applications by proposing a vision-based strategy for pose estimation of aerial vehicles. The vast majority of approaches make use of features to estimate motion. Conversely, the strategy we propose is based on a MR (Multi-Resolution) implementation of an image registration technique (Inverse Compositional Image Alignment ICIA) using direct methods. An on-board camera in a downwards-looking configuration, and the assumption of planar scenes, are the bases of the algorithm. The motion between frames (rotation and translation) is recovered by decomposing the frame-to-frame homography obtained by the ICIA algorithm applied to a patch that covers around the 80% of the image. When the visual estimation is required (e.g. GPS drop-out), this motion is integrated with the previous known estimation of the vehicles' state, obtained from the on-board sensors (GPS/IMU), and the subsequent estimations are based only on the vision-based motion estimations. The proposed strategy is tested with real flight data in representative stages of a flight: cruise, landing, and take-off, being two of those stages considered critical: take-off and landing. The performance of the pose estimation strategy is analyzed by comparing it with the GPS/IMU estimations. Results show correlation between the visual estimation obtained with the MR-ICIA and the GPS/IMU data, that demonstrate that the visual estimation can be used to provide a good approximation of the vehicle's state when it is required (e.g. GPS drop-outs). In terms of performance, the proposed strategy is able to maintain an estimation of the vehicle's state for more than one minute, at real-time frame rates based, only on visual information.

The availability of accurately aligned, whole-body anatomical (CT) and functional (PET) images could have a significant impact on diagnosing and staging malignant disease and on identifying and localizing metastases. Computer algorithms to align CT and PET images acquired on different scanners are generally successful for the brain, whereas image alignment in other regions of the body is more problematic. A combined PET/CT tomograph with the unique capability of acquiring accurately aligned functional and anatomical images for any part of the human body has been designed and built. The PET/CT scanner was developed as a combination of a Siemens Somatom AR.SP spiral CT and a partial-ring, rotating ECAT ART PET scanner. All components are mounted on a common rotational support within a single gantry. The PET and CT components can be operated either separately, or in combined mode. In combined mode, the CT images are used to correct the PET data for scatter and attenuation. Fully quanti...

Image alignment refers to finding the best transformation from a fixed reference image to a new image of a scene. This process is often guided by similarity measures between images, computed based on the image data. However, in time-critical applications state-of-the-art methods for computing similarity are too slow. Instead of using all the image data to compute similarity, one can use a subset of pixels to improve the speed, but often this comes at the cost of reduced accuracy. This makes the problem of image alignment a natural application domain for deliberation control using anytime algorithms. However, almost no research has been done in this direction. In this paper, we present anytime versions for the computation of two common image similarity measures: mean squared difference and mutual information. Off-line, we learn a performance profile specific to each measure, which is then used on-line to select the appropriate amount of pixels to process at each optimization step. Wh...

Improved X-ray sources and optics now enable X-ray imaging resolution down to ∼50 nm for laboratory-based X-ray microscopy systems. This offers the potential for submicrometre resolution in tomography; however, achieving this resolution presents challenges due to system stability. We describe the use of software methods to enable submicrometre resolution of approximately 560 nm. This is a very high resolution for a modest laboratory-based point-projection X-ray tomography system. The hardware is based on a scanning electron microscope, and benefits from inline X-ray phase contrast to improve visibility of fine features. Improving the resolution achievable with the system enables it to be used to address a greater range of samples.

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.

A new approach for the measurement of the red blood cell (RBC) velocity from capillary video by using optical flow estimation has been developed. An image registration function based on mutual information was used for stabilizing images in order to cope with slight finger movement during video acquisition. After image alignment, a skeleton extraction algorithm implemented by thinning was followed which enabled tracking blood flow entirely in arteriolar and venular limbs, and the curved segment as well. Optical flow and cross-correlation approaches were applied individually for velocity estimation of twelve microcirculation videos acquired independently from three healthy volunteers. The RBC velocity of 12 vessels at three given measurement sites (arteriolar, curve and venular sites) in a 45-second period of occlusion-release condition of vessel were examined. There were four stages of flow conditions: resting (T(1)), pre-occlusion (T(2)), post-occlusion (T(3)) and release (T(4)). The results from both approaches revealed that the velocity difference among the three sites were not significant. The pattern of distribution of RBC velocity was also reported. The correlation coefficient (r) of the velocity calculated using optical flow and cross-correlation in four stages of blood flow conditions and the overall correlation were: 1-window: r(T1)=0.68, r(T2)=0.67, r(T3)=0.92, r(T4)=0.88 and r(All)=0.79; 2-window: r(T1)=0.84, r(T2)=0.88, r(T3)=0.87, r(T4)=0.93 and r(All)=0.88. The averaged velocity results showed no significant differences between optical flow and 2-window cross-correlation in all flow conditions. Optical flow estimation is not only independent to the direction of flow, but also able to calculate the intensity displacement of all pixels. The proposed velocity measurement system has been shown to provide complete velocity information for the whole vessel limb which demonstrates the advantage of measuring blood flow at the level of microcirculation more accurately.

Shape matching and point correspondence recovering play a fundamental role in applications like pattern and object recognition, shape classication, image alignment and regis- tration, visual information data mining, and many other computer vision related areas, which recently became active research disciplines due to the increasing needs in many practical elds. This article introduces the Contour-Point Signature (CPS), a novel technique