Search

Quantification of the eye’s anterior segment morphology from optical coherence tomography (OCT) images is crucial for research and clinical decision-making, including the diagnosis and monitoring of many ocular diseases. Structural parameters, such as tissue thickness and area are the most common metrics used to quantify these medical images, and tissue segmentation is required before these metrics can be extracted. Currently, swept source OCT allows the capture of cross-sectional images that encompass the entire anterior segment with a high level of detail. However, the manual annotation of tissue boundaries is time-consuming. In this work, an algorithm based on graph-search theory combined with boundary-specific image transformation is applied for the segmentation of anterior segment OCT images. We demonstrate that the method can reliably segment 5 different tissue boundaries in healthy eyes with low boundary error (mean error below 1 pixel across all boundaries). The technique can be used to extract clinically relevant parameters such as central corneal and crystalline lens thickness as well as anterior chamber depth and area, with a high level of agreement with manual segmentation (normalized errors below 1.6%). The proposed method provides a tool that can support clinical and research OCT data analysis.

Computer vision as a scientific discipline refers to the theories and technologies for creating artificial systems that receive information from an image. Despite the fact that this discipline is quite young, its results have penetrated almost all areas of life. Computer vision is closely related to other practical fields like image processing, the input of which is two-dimensional images obtained from a camera or artificially created. This form of image transformation is aimed at noise suppression, filtering, color correction and image analysis, which allows you to directly obtain specific information from the processed image. This information may include searching for objects, keypoints, segments, and annexes

Цель исследования - анализ трансформации городского образа на примере Кёнигсберга-Калининграда. Новизна исследования - в применение градостроительной схемы К.Линча (1960) к анализу разновременной пространственной структуры крупного старого немецкого города Кёнигсберга и послевоенного (современного) Калининграда. The aim of the research is to analyze the transformation of the urban image on the example of Königsberg-Kaliningrad. The novelty of the research lies in the application of the urban planning scheme of K. Linch (1960) to the analysis of the multi-temporal spatial structure of the large old German city of Königsberg and post-war (modern) Kaliningrad.

In today’s era, the large amount of data’s are stored in cloud. But nowadays securely storing data to the cloud is an important task. But the unauthorized parties are tries to decrypt the data from the cloud. So, the data stored in cloud should be secure from any malicious activities. Based on these requests, we propose a framework named reversible data hiding (RDH). It mainly works on encrypted images based on reversible image transformation (RIT). The server insert data into the cloud. On that time, he will add some additional information for the protection of data. In the given framework, the original data can be embedded in another image named carrier image with equal size. So, the hacker tries to pullout the data, he will only get the carrier image. So we can accommodate more data’s into a single storage space. The method picks the data to be hide and a carrier image with equal size. The cover image is embedded into the original image based on the LSB insertion algorithm. i.e., the original data’s are dissolved into the carrier image using a secret hiding key. The cover image is encrypted using blow fish encryption algorithm. It includes normal RDH method and RDH with RIT. Using RIT, it gives high visual quality and security. The technique is mainly used in medical imagery, military imagery for reliable data storage.

This work focuses on a method to hide images in videos in a manner that the secret image can be losslessly recovered from the target image with minimal distortion. This lossless recovery can be done by using Reversible Data Hiding (RDH). To ensure the privacy of the video owner, the target media is encrypted before applying RDH. Reversible Image Transformation (RIT) is a framework used to ensure that. Audio Steganography techniques are used for further encryption. This can be used in cloud technology so that the cloud may add information into the target video without compromising its integrity. RDH also has other possible applications in the field of military, defence and medicine.

Linear array cameras are widely used in the fields of aerial and space remote sensing. The ground‐to‐image transformation plays a significant role in the geometric processing procedures of linear pushbroom images, which is time‐consuming due to multiple iterative calculations. Moreover, the detectors of the linear array may exhibit a curved shape due to distortions, resulting in a more complex ground‐to‐image transformation. This paper presents a novel ground‐to‐image transformation algorithm for linear pushbroom images based on a rigorous sensor model. We regard the distance between the back‐projected image point and the curved shape linear array as the “generalized distance.” The idea of generalized distance prediction was developed to efficiently determine the best scan line corresponding to the ground point. The plane analytic geometry is employed to solve the practical problems of the ground‐to‐image transformation, including the iterative direction, convergence condition, and subpixel interpolation. The applications of the proposed method in orthorectification and image matching were investigated. The proposed method was comprehensively tested using airborne images, Earth observation satellite images, and planetary images. The experimental results demonstrate that the proposed method (1) delivers high accuracy (better than 1 × 10−4 pixels) and efficiency, (2) is better than the cutting‐edge method based on the geometric constraints of central perspective plane for the linear pushbroom images with distortions, and (3) provides promising applications for generating orthophotos and image matching. The proposed method is easy to implement and efficient, which can greatly enhance the geometric processing abilities for various types of linear pushbroom images.

© 2019 Elsevier Ltd Particle Image Velocimetry is commonly used to compute velocity fields in several areas including fluid mechanics, hydraulics and geophysics. However, acquired images often contain deformations caused either by camera lenses or placement. In this work the most popular digital transformation methods used to remove/reduce these deformations are benchmarked and suggestions tailoring specific transformations to different types of deformations are made. This article also shows the reduction of the error associated to the first and second order statistics, in the case of two-dimensional Particle Image Velocimetry, when the transformation techniques are applied to the computed velocity fields, and not the raw images, a common option in available commercial software.

In this research, we have carried out various experiments to perform mutual transformation between a domain of Ikebana (Japanese traditional flower arrangement) photos and other domains of images (landscapes, animals, portraits) to create new artworks via a variation of CycleGAN - a GANs technique based on cycle-consistency loss. A pre-trained process on object detection was added to improve the efficiency by avoiding over-transformation.

In this paper, we propose three methods for removing floating particles from underwater images. The first two methods are based on Generative Adversarial Networks (GANs). The first method uses CycleGAN which can be trained with an unpaired dataset, and the second method uses pix2pixHD that is trained with a paired dataset created by adding artificial particles to underwater images. The third method consists of two-step process – particle detection and image inpainting. For particle detection, an image segmentation neural network U-Net is trained by using underwater images added with artificial particles. Using the output of U-Net, the particle regions are repaired by an image inpainting network Partial Convolutions. The experimental results showed that the methods using GANs were able to remove floating particles, but the resolution became lower than that of the original images. On the other hand, the results of the method using U-Net and Partial Convolutions showed that it is capable of accurate detection and removal of floating particles without loss of resolution.

Image projection is the premise of image transformation, and also the premise of image preprocessing, splicing, operation and fusion. When the camera collects the image, it usually carries out complex motion, and different motion modes will produce different effects on the scene imaging. The motion mode of camera is usually divided into several relatively simple motion modes: translation motion, lens zooming, horizontal sweeping, vertical sweeping and rotation motion. Through image distortion restoration technology, the interference factors that affect image quality are solved, and these factors are removed by technical means, so as to ensure the space invariance of image projection.

In the past decade, convolutional neural networks (CNNs) have achieved great practical success in image transformation tasks, including style transfer, semantic segmentation, etc. CNN-based style transfer, which denotes transforming an image into a desired output image according to a user-specified style image, is one of the most popular techniques in image transformation. It has led to to many successful industrial applications with significant commercial impacts, such as Prisma and DeepArt. Figure 1 shows the general workflow of the CNN-based style transfer. Given a content image and a user-specified style image, the content features and style features can be extracted using a pre-trained CNN, and then be merged to generate the stylized image. The CNN model is trained for generating a stylized image that has similar content features as the content image's and similar style features as the style image's. In this example, we can see the content image is captured at a lake in the daytime, while the style image is another similar scene captured at dusk. After performing style transfer, the content image is successfully transformed to the dusky scene while keeping the content unchanged as the content image.

In the past decade, convolutional neural networks (CNNs) have achieved great practical success in image transformation tasks, including style transfer, semantic segmentation, etc. CNN-based style transfer, which denotes transforming an image into a desired output image according to a user-specified style image, is one of the most popular techniques in image transformation. It has led to to many successful industrial applications with significant commercial impacts, such as Prisma and DeepArt. Figure 1 shows the general workflow of the CNN-based style transfer. Given a content image and a user-specified style image, the content features and style features can be extracted using a pre-trained CNN, and then be merged to generate the stylized image. The CNN model is trained for generating a stylized image that has similar content features as the content image's and similar style features as the style image's. In this example, we can see the content image is captured at a lake in the daytime, while the style image is another similar scene captured at dusk. After performing style transfer, the content image is successfully transformed to the dusky scene while keeping the content unchanged as the content image.

International audience; We present an algorithm for augmenting the shape of the vocal tract using 3D static and 2D dynamic speech MRI data. While static 3D images have better resolution and provide spatial information, 2D dynamic images capture the transitions. The aim of this work is to combine strong points of these two types of data to obtain better image quality of 2D dynamic images and extend the 2D dynamic images to the 3D domain. To produce a 3D dynamic consonant-vowel (CV) sequence, our algorithm takes as input the 2D CV transition and the static 3D targets for C and V. To obtain the enhanced sequence of images , the first step is to find a transformation between the 2D images and the mid-sagittal slice of the acoustically corresponding 3D image stack, and then find a transformation between neighbouring sagittal slices in the 3D static image stack. Combination of these transformations allows producing the final set of images. In the present study we first examined the transformation from the 3D mid-sagittal frame to the 2D video in order to improve image quality and then we examined the extension of the 2D video to the 3rd dimension with the aim to enrich spatial information.

Face sketch recognition is one of the most challenging heterogeneous face recognition problems. The large domain difference of hand-drawn sketches and color photos along with the subjectivity/variations due to eye-witness descriptions and skill of sketch artists makes the problem demanding. Therefore, despite several research attempts, sketch to photo matching is still considered an arduous problem. In this research, we propose to transform a hand-drawn sketch to a color photo using an end to end two-stage generative adversarial model followed by learning a discriminative classifier for matching the transformed images with color photos. The proposed image to image transformation model reduces the modality gap of the sketch images and color photos resulting in higher identification accuracies and images with better visual quality than the ground truth sketch images.

Sketch-based image retrieval (SBIR) technique has progressed by deep learning to learn cross-modal distance metrics that relate sketches and photos from a large number of sketch-photo pairs. However, datasets of sketch-photo pairs are small, as acquisition of a large number of such pairs is expensive. To alleviate the issue, data augmentation via image transformation such as scaling, flipping, rotation, and deformation has been widely adopted. Still, insufficiency in training set seems to have impeded deep learning from achieving its full potential for SBIR. In this paper, we propose a novel data augmentation approach dedicated for SBIR. A deep neural network called Photo2Sketch (P2S) converts photos into line drawings that are visually similar to those sketched by human. An artificially augmented training dataset of sketch-photo pairs is generated at low cost by feeding photos from a large image corpus into the P2S. Experiments evaluate quality of sketch-like images generated by the P2S as well as efficacy of the proposed data augmentation algorithm under SBIR scenario. In particular, retrieval accuracy is significantly improved when the proposed algorithm is combined with the data augmentation by image transformation

This paper proposes image compression using discrete cosine transform (DCT) for the format of joint photographic expert groups (JPEG) or JPG, portable network graphic (PNG) and bitmap (BMP). These three extensions are the most popular types used in current image processing storage. The purpose of image compression is to produce lower memory usage or to reduce memory file. This process removes redundant information of each pixel. The challenge for image compression process is to maintain the quality of images after the compression process. Hence, this article utilizes the DCT technique to sustain the image quality and at the same time reduces the image storage size. The effectiveness of the DCT technique has been reasonable over some real images and the implementation of the technique has been compared with different types of image extensions. Matlab software is an important platform for this project in order to write a program and perform the progress of project phase by phase to achieve the expected results. Based on the tested images, the DCT technique in image compression is capable to reduce the image storage memory in average about 50% of each image tested.

This paper presents a novel Gabor phase based illumination invariant extraction method aiming at eliminating the effect of varying illumination on face recognition. Firstly, It normalizes varying illumination on face images, which can reduce the effect of varying illumination to some extent. Secondly, a set of 2D real Gabor wavelet with different directions is used for image transformation, and multiple Gabor coefficients are combined into one whole in considering spectrum and phase. Lastly, the illumination invariant is obtained by extracting the phase feature from the combined coefficients. Experimental results on the Yale B and the CMU PIE face database show that our method obtained a significant improvement over other related methods for face recognition under large illumination variation condition.

Robots are used frequently nowadays to reduce the human effort due to its efficient capability and performance. However, interaction with human-friendly environment is needed to integrate the robot accurately. Recently, robotic vision plays an important role in the real-world phenomenon to achieve this goal. This is because robotic vision has the capability to identify and determine the accurate positions of all related objects within the working area of the robot. In robotic vision, images are used as input, which analyzes the content of images. These have been used for the output of the system based on the criteria of image analysis, image transformation, and image understanding. Hence, the system may be able to capture the related information using the motion of the objects. And it updated this information for verification, tracking, acquisition, and extractions of images to adopt in the database. The main objective of this work is to enable the algorithm to understand the visual world. In order to achieve the goal, we have proposed an algorithm which is based on computer vision theories. Experiments have been performed on real-world image which shows our algorithm has better performance.

Information hiding technology has always been a hot research field in the field of information security. How to hide information in images is also a major research field of image transformation. Today’s information hiding technology has the problems of complicated encryption technology and large amount of information. Based on the genetic algorithm, this paper studies the image information hiding technology based on genetic algorithm, performs image simulation on different types of mothers, and displays the results. For the images that need to be hidden, no matter how the size changes, the mother image can achieve good information hiding. The change of the parent image is not large. Compared with the Least Significant Bit (LSB) technique, this method has a larger peak signal-to-noise.

Presently, a new reversible image transformation technique is proposed, this technique not only enhances the quality of the encrypted image but also it can restore the secret image in lossless manner. Furthermore, is delegate data to the cloud. Therefore, the cloud can easily add additional data into the encrypted image by any RDH methods. It is very crucial to secure data and allow the cloud server to manage the data at the meantime. Under such demands, proposes a method of Reversible Data Hiding in Encrypted Image based on Reversible Image Transformation. Other from all existing encryption methods, RIT based method allow user to transmute the data of original image into another target image with the same size. Which secure the original image, the transmuted image which appears like the target image which is used as the encrypted image, and the transmutation can be done between the micro blocks with small size, which enhance the quality of the encrypted image.

Subject of Study. The key feature of spherical panoramas is the maximum possible angle of view (360 × 180 degrees). Omnidirectional optical-electronic systems are able to produce images that show most of this space, but these images are different from the canonical spherical panoramas. This paper proposes a method of converting the circular images obtained by omnidirectional optical systems, to canonical spherical panoramic images with the use ofcalibration procedure of omnidirectional optical-electronic system. Method. The process of spherical panoramas creation consists of three steps.The first step includesthe forming of virtual surface in the object space corresponding to a field of view of the spherical panorama. The surface is defined by the three-dimensional array of pixels.At the second step the coordinates of images of these points in the plane of the detectorare specified. At the third step pixel-by-pixel forming of the output imageisperformed from the original omnidirectional image with the use of coordinates obtained at the second step. Main Results. We have considered the geometric projectionmodel of spherical panoramas. The algorithm has been proposed calculating the three-dimensional array of pixels, characterizing the field of view of a spherical panorama and convenient for practical usage. The developed method is designed to work with omnidirectional optical-electronic systems both with catadioptric optical systems and with fisheye lens. Experimental results confirming the validity of this method are presented. The reprojection mean-square error was equal to 0.794 px. Practical Relevance. The proposed method can be applied in technologies of creating of virtual tours or panoramas of streets where spherical panoramic images are the standard for storing visual information. The method may also find its application in various fields of robotics, orientation and navigation of space vehicles and UAVs.

The multi-lens multispectral cameras (MSCs), such as Micasense Rededge and Parrot Sequoia, can record multispectral information by each separated lenses. With their lightweight and small size, which making they are more suitable for mounting on an Unmanned Aerial System (UAS) to collect high spatial images for vegetation investigation. However, due to the multi-sensor geometry of multi-lens structure induces significant band misregistration effects in original image, performing band co-registration is necessary in order to obtain accurate spectral information. A robust and adaptive band-to-band image transform (RABBIT) is proposed to perform band co-registration of multi-lens MSCs. First is to obtain the camera rig information from camera system calibration, and utilizes the calibrated results for performing image transformation and lens distortion correction. Since the calibration uncertainty leads to different amount of systematic errors, the last step is to optimize the results in order to acquire a better co-registration accuracy. Due to the potential issues of parallax that will cause significant band misregistration effects when images are closer to the targets, four datasets thus acquired from Rededge and Sequoia were applied to evaluate the performance of RABBIT, including aerial and close-range imagery. From the results of aerial images, it shows that RABBIT can achieve sub-pixel accuracy level that is suitable for the band co-registration purpose of any multi-lens MSC. In addition, the results of close-range images also has same performance, if we focus on the band co-registration on specific target for 3D modelling, or when the target has equal distance to the camera.

There is an image filtration using an Ateb-Gabor method realized. Filtration in the frequency area requires implementation of such changes sequence: discrete two-dimensional image transformation from spatial area to the frequency area. Filtration makes it possible to realize more combinations of filtering and conduct more qualitative transformation. Proposed method allows to expend a Gabor filter performance because it contains much more combinations. There was an Ateb-Gabor filtration provided to the one case of images and to the one periodic Ateb-function.

Medical imaging is the technique that is used to produce images of the human body or parts for clinical purpose. The CT image provides thorough information of structure of lungs, which could be used for better surgical preparation for treating Lung Cancer features. This work proposes a method for segmentation of lungs from the given CT images. Here a curve evolutional framework for segmentation is used. The flow fields driving the curves are based on the distributions of features in the inner and outer regions bounded by the curves. The segmentation method is automatic and shows good result. Computer Tomography (CT) is one of the most efficient medical diagnostic methods and has currently a widespread usage. Multi-slice CT scanning technology has revolutionized screening of the lungs and inspires necessitate for pulmonary image analysis. Segmentation of the lungs and lobes is a prerequisite for such image analysis in chest CT scans. CT scan is more appropriate for showing the detailed information of the parts of human body and it is used for various applications such as detection, classification etc. The analysis of lungs in CT image is used to detect the airway and the vessel present in the lungs [6]. The human lungs are subdivided into distinct pulmonary lobes separated by thin barriers, called fissures. Each lobe contains separate supply branches for both vessels and airways. The segmentation of anatomical parts is important in the detection of pathological abnormalities present in the lungs. Segmentation of the pulmonary lobes is relevant in many clinical applications. Accurate lung segmentation allows for the detection and quantification of abnormalities within the lungs. The lobes are separately supplied by the first subdivisions of the bronchial tree after the main bronchi. The lobes function independently within the lungs. Segmentation of the pulmonary lobes is relevant in clinical practice and particularly challenging for cases with severe diseases or incomplete fissures. Pulmonary fissure is a boundary between the lobes in the lungs. Its segmentation is of clinical interest as it facilitates the assessment of lung disease on a lobar level. A new approach has been made for segmenting the major fissures in both lungs on thin-section computed tomography (CT) [5]. An image transformation called “ridge map” is proposed for enhancing the appearance of fissures on CT. A curve-growing process, modeled by a Bayesian network, is described that is influenced by both the features of the ridge map and prior knowledge of the shape of the fissure. The process is implemented in an adaptive regularization framework that balances these influences and reflects the causal dependencies in the Bayesian network using an entropy measure. This method effectively alleviates the problem of inappropriate weights of

Many people who have visual disturbance, called as “low vision”, such as old sight and cataract feel inconvenient in daily life. For example, climbing stairs, walking the street, reading information on a newspaper or a book, etc.. Especially, they have a difficulty to read characters without using optical aids. In this paper, we propose a novel algorithm to transform a character appearance for compensating the differences of visual perception such as blurring. In the proposed method, characters in an image are transformed into the characters which can be recognized with ease according to pre-learning rules. In the experiments, we apply the proposed method to various characters and fonts. The experimental results show the effectiveness of proposed method.

This paper proposes finding optimal ways of recognizing geometric shapes (square, circle, rectangle, etc.) of files in different formats. Shapes recognition is a field of artificial intelligence, which includes all representation and decision techniques to automate the process of identifying similarities between objects or phenomena. An application of shapes recognition requires the definition of descriptors and choosing a distance. An application of shapes recognition is done in two phases: learning and recognition. By learning to calculate the set of descriptors of known (reference). In recognition stage, we calculate the same set of descriptors for unknown form and compared with known shape descriptors. The comparison is made through a distance. Recognition is a decision problem if you say unknown form is the same as the reference shape and if you say unknown form is different form the reference. Classification of forms is done in five steps: Reading and image display, image transformation in binary image and get its negative, Discover crossing thresholds between levels of color, Determine the properties of objects and Classification of objects by shape. Are three methods: Chain codes, the conventional Hough transform, method Harris Corner Detection

This paper presents a human belly surface image matching technology based on stereo vision system. Our aim is to still recover the human belly surface shape accurately and efficiently in spite of the influences derived from illumination variation, blur affection and image transformation on the 2D images concerning the frontal and rear sides of belly acquired by only four cameras. For this purpose, we devote to develop an innovative stereo matching algorithm which includes two major phases. In the first phase, a novel feature descriptor is constructed by substituting blur invariance into new illumination-robust combined invariants based on Hu-moment, which can avoid the various influences from the actual photographing. The external factors that may cause the images non-ideal such as the illumination variation, the camera photosensitive discrepancy from the different perspective images, the motion of the organism, noise and so on. Meanwhile, the pixel gradient-level intensity is brought into similarity meas...

In this paper, we explore the PET-MRI 3D surface image registration method for the longitudinal study of cancer. Image registration gives new information to the radiologists by matching images with different modalities. The objective of this research is to develop 3D surface registration for PET/CT and MR images acquired by different medical imaging systems at different times. Our proposed method consists of three stages: 1) global surface matching, 2) fine adjustment stage and 3) image transformation. In this study, we perform the global registration with surface points within CT and MRI using ICP algorithm. Then we perform the fine adjusts stage using ECICP algorithm. Finally, PET image was transformed with various parameters and overlaid onto MR image.

This paper presents a semantic segmentation based method for automatically synthesizing two-tone cartoon portraits in black-and-white style. Synthesizing two-tone portraits from photographs can be considered as a heterogeneous image transformation problem, of which the result should be vivid portraits with distinct freehand-like features, such as clean backgrounds and continuous lines. To achieve this goal, our system connects two separate subsystems together, namely semantic segmentation and portrait synthesis. In the semantic segmentation phase, photographs are segmented into background, hair and skin regions using multiple segmentations method. In the portrait synthesis phase, we treat different regions with different strategies. Experimental results demonstrate that our system can precisely segment the input photo and produce visually desired two-tone portraits.

Current image transformation and recoloring algorithms try to introduce artistic effects in the photographed images, based on user input of target image(s) or selection of pre-designed filters. These manipulations, although intended to enhance the impact of an image on the viewer, do not include the option of image transformation by specifying the affect information. In this paper we present an automatic image-transformation method that transforms the source image such that it can induce an emotional affect on the viewer, as desired by the user. Our proposed novel image emotion transfer algorithm does not require a user-specified target image. The proposed algorithm uses features extracted from top layers of deep convolutional neural network and the user-specified emotion distribution to select multiple target images from an image database for color transformation, such that the resultant image has desired emotional impact. Our method can handle more diverse set of photographs than the previous methods. We conducted a detailed user study showing the effectiveness of our proposed method. A discussion and reasoning of failure cases has also been provided, indicating inherent limitation of color-transfer based methods in the use of emotion assignment. Project Page: http://im.itu.edu.pk/affective-image-transfer/, Comment: Published at British Machine Vision Conference (BMVC) 2017

By using Convolutional Neural Networks (CNNs), the semantics and styles of images can be separated and recombined to create fascinating images. In this paper, an image transformation network for style transfer is proposed, which consists of convolution layers, deconvolution layers and Fusion modules composed of two 1 × 1 convolution layers and a residual block. The output of each layer in the network is normalized using batch normalization to speed up the training process. Compared with other networks, our network has fewer parameters and better real-time performance while generating similar quality images.

The efficient use of satellite images for practical purposes depends upon the ability to combine them for the purposes of analysis. However, this process involves many computationally intensive computer vision algorithms. Assembling large images usually requires costly hardware with a large amount of memory. This paper addresses this problem by proposing a system that implements algorithms required in the process of image stitching, viz. algorithms for feature detection, feature description, feature matching, image transformation and image compositing, in parallel on a distributed memory architecture. The proposed system aims to improve the performance of the image stitching process while providing a more easily scalable system.

This chapter presents the image registration methods that can be used when validating atherosclerotic plaque characterization methods with histological images or micro-CT images. Both histological images and micro-CT images can be used as gold standard for detecting the plaque; however, the lack of blood flow when acquiring those images leads to deformed and shrunken images. Rigid and nonrigid image transformation models have been developed for the registration of different image modalities which represent the same objects. These models are explained and presented in detail. The application of a modified nonrigid image transformation model for inflating the histological and micro-CT images is also included.

This paper presents a new hardware-oriented approach for the extraction of disparity maps from stereo images. The proposed method is based on the herein named Adaptive Census Transform that exploits adaptive support weights during the image transformation; the adaptively weighted sum of SADs is then used as the dissimilarity metric. Quality tests show that the proposed method reaches significantly better accuracy than alternative hardware-oriented approaches. To demonstrate the practical hardware feasibility, a specific architecture has been designed and its implementation has been carried out using a single FPGA chip. Such a VLSI implementation allows a frame rate up to 68fps to be reached for 640x480 stereo images, using just 80,000 slices and 32 RAM blocks of a Virtex6 chip.

A lot of feature-based correspondence matching methods have been proposed in the field of computer vision, image processing and pattern recognition. These methods are also effective for biometric recognition. In general, in the case of feature-based matching methods, the matching score is calculated as a ratio between the number of feature points and corresponding points. These methods need to normalize image deformation by fitting an image transformation model to images according to the correspondence between images. Then, the matching score is calculated from the normalized images so as to take into consideration image deformation. On the other hand, this paper proposes a score calculation method which calculates a matching score from positional information of corresponding point pairs. The proposed method does not need any deformation model defined for each biometric trait to handle image deformation. The combination of the matching scores defined by the number of corresponding points and the positional information improves the performance of biometric recognition algorithms, since these scores play a complementary role in decision. Through a set of experiments using a palmprint image database, we demonstrate that the proposed method exhibits efficient performance for biometric recognition.

Image segmentation is a technique to locate certain objects or boundaries within an image. Image segmentation plays a crucial role in many medical imaging applications. There are many algorithms and techniques have been developed to solve image segmentation problems. Spectral pattern is not sufficient in high resolution image for image segmentation due to variability of spectral and structural information. Thus the spatial pattern or texture techniques are used. Thus the concept of Holder Exponent for segmentation of high resolution medical image is an efficient image segmentation technique. The proposed method is implemented in Matlab and verified using various kinds of high resolution medical images. The experimental results shows that the proposed image segmentation system is efficient than the existing segmentation systems. &nbsp