We present a novel fusion method to improve the spatial resolution of multispectral (MS) images, where the fused spectral images integrate the spectral information and spatial details from the original MS images and panchromatic (PAN) image, respectively. Band by band, a spectral image with high resolution is reconstructed from the original spectral image to take advantage of super-resolution technology. The pan-sharpening method via Amélioration de la Resolution Spatiale par Injection de Structures concept is further applied to obtain the fused images from the reconstructed spectral images and PAN image. Performance of the proposed method has been evaluated on the public optical satellite QuickBird images. Experimental results show that the fused spectral images both preserve the spatial details of high-resolution from the PAN image and have higher spectral resolution than the original spectral images.

Although many histogram equalization (HE) based methods have been developed to overcome the problem of mean-shift, they often suffer from over-enhancement caused by an abrupt jump of gray levels. Consequently, mean brightness cannot be preserved with higher accuracy. In this paper, one simple and efficient method, brightness-preserving weighted subimages (BPWSI), for brightness preservation and contrast enhancement of gray level images is proposed. First, based on brightness preserving bi-histogram equalization, two new subimages are defined. Then one brightness-preserving way to calculate two weight coefficients is given. Finally, the output image is defined as the weighted sum of two defined subimages. Extensive simulations show that BPWSI can preserve the mean brightness with higher accuracy than many other HE-based methods while enhancing the contrast efficiently. Furthermore, due to its simplicity, BPWSI can be implemented on real-time systems.

Range scanner inspection of manufactured artifacts like propeller blades produces data used to determine conformity to design. Inspection data contain both large- and small-scale feature information. Because small-scale feature information is irrelevant to analysis of large-scale features, determination of large-scale shape conformance is easier if small-scale information is removed. We discuss the combination of registration and smoothing methods that have been implemented in software for analysis of large-scale shape conformance of manufactured marine propeller blades. We show that although inspection systems produce errors and noise, we are able to create surfaces that accurately represent actual artifact large-scale features. To our knowledge the implementation of these methods has not been accomplished previously in an industrial inspection environment.

We present a study of JPEG baseline coding. It aims to determine the minimum storage needed to buffer the JPEG Huffman code bits of 8-bit image blocks. Since DC is coded separately, and the encoder represents each AC coefficient by a pair of run-length/AC coefficient level, the net problem is to perform an efficient search for the optimal run-level pair sequence. We formulate it as a two-dimensional, nonlinear, integer programming problem and solve it using a branch-and-bound based search method. We derive two types of constraints to prune the search space. The first one is given as an upper-bound for the sum of squares of AC coefficients of a block, and it is used to discard sequences that cannot represent valid DCT blocks. The second type constraints are based on some interesting properties of the Huffman code table, and these are used to prune sequences that cannot be part of optimal solutions. Our main result is that if the default JPEG compression setting is used, space of minimum of 346 bits and maximum of 433 bits is sufficient to buffer the AC code bits of 8-bit image blocks. Our implementation also pruned the search space extremely well...

JPEG2000, the new image coding standard, is more efficient when compared with the old JPEG standard. However, most of the computations and memory usage in post-compression rate distortion (PCRD) optimization scheme in JPEG2000 is redundant. We propose an efficient rate control approach for the JPEG2000 encoder. In this approach, optimal rate distortion (RD) slope threshold is selected using the minimum RD slope of lower level horizontally and vertically low pass (LL₀) subband of the wavelet decomposed image. Also, the passes of the code blocks of other subbands, whose RD slopes are less than the optimal RD slope, are skipped. The proposed approach reduces memory requirements and encoding time when compared with the existing rate control approaches for JPEG2000 standard, and also provides the same image quality as that of the PCRD approach used in the JPEG2000 standard.

A robust image watermarking scheme based on the discrete wavelet transform (DWT), discrete cosine transform (DCT), and singular value decomposition (SVD) is proposed. The cover image is transformed from the RGB color space into the YCbCr color space to obtain a gray-level image. The DWT is then applied to the gray-level image to obtain a low-frequency (LL) subband image. Next, the DCT is applied to the LL subband image to obtain the frequency components. Finally, SVD is used on the obtained frequency components to embed the watermark. Experimental results show that the proposed method is robust against cropping, rotation, Gaussian noise additions, histogram equalization, JPEG compression, and JPEG2000 compression attacks. Therefore, the proposed method is very suitable for applications of transaction tracking, data authentication, and video surveillance.

We propose a page layout analysis algorithm to classify a scanned document into different regions such as text, photo, or strong lines. The proposed scheme consists of five modules. The first module performs several image preprocessing techniques such as image scaling, filtering, color space conversion, and gamma correction to enhance the scanned image quality and reduce the computation time in later stages. Text detection is applied in the second module wherein wavelet transform and run-length encoding are employed to generate and validate text regions, respectively. The third module uses a Markov random field based block-wise segmentation that employs a basis vector projection technique with maximum a posteriori probability optimization to detect photo regions. In the fourth module, methods for edge detection, edge linking, line-segment fitting, and Hough transform are utilized to detect strong edges and lines. In the last module, the resultant text, photo, and edge maps are combined to generate a page layout map using K-Means clustering. The proposed algorithm has been tested on several hundred documents that contain simple and complex page layout structures and contents such as articles, magazines, business cards, dictionaries, and newsletters, and compared against state-of-the-art page-segmentation techniques with benchmark performance. The results indicate that our methodology achieves an average of ∼ 89% classification accuracy in text, photo, and background regions.

The low-contrast images taken by digital cameras or camera phones are not always satisfactory due to the limitation of the capturing devices or improper illumination/exposure conditions. Conventional image contrast enhancement methods may either fail to produce satisfactory and undistorted images, or they cannot improve every region of interest appropriately, especially faces. In this paper, a histogram equalization (HE) approach exploiting the just-noticeable-difference (JND) model of the human visual system (HVS), denoted by JND-HE, is proposed for generic image contrast enhancement. Further, the proposed JND-HE approach is combined with the exposure correction (EC) method (denoted by JND-HE-EC) for face image enhancement. The proposed JND-HE-EC approach can improve the contrast in face regions and provide proper illumination in the background. Experimental results on both generic images and faces have shown that our proposed approach can produce more pleasing and appealing enhanced images than other methods.

To make progress in understanding knot theory, we need to interact with the projected representations of mathematical knots, which are continuous in three dimensions (3-D) but significantly interrupted in the projective images. One way to achieve such a goal is to design an interactive system that allows us to sketch two-dimensional (2-D) knot diagrams by taking advantage of a collision-sensing controller and explore their underlying smooth structures through a continuous motion. Recent advances of interaction techniques have been made that allow progress in this direction. Pseudohaptics that simulate haptic effects using pure visual feedback can be used to develop such an interactive system. We outline one such pseudohaptic knot diagram interface. Our interface derives from the familiar pencil-and-paper process of drawing 2-D knot diagrams and provides haptic-like sensations to facilitate the creation and exploration of knot diagrams. A centerpiece of the interaction model simulates a physically reactive mouse cursor, which is exploited to resolve the apparent conflict between the continuous structure of the actual smooth knot and the visual discontinuities in the knot diagram representation. Another value in exploiting pseudohaptics is that an acceleration (or deceleration) of the mouse cursor (or surface locator) can be used to indicate the slope of the curve (or surface) of which the projective image is being explored. By exploiting these additional visual cues, we proceed to a full-featured extension to a pseudohaptic four-dimensional (4-D) visualization system that simulates the continuous navigation on 4-D objects and allows us to sense the bumps and holes in the fourth dimension. Preliminary tests of the software show that main features of the interface overcome some expected perceptual limitations in our interaction with 2-D knot diagrams of 3-D knots and 3-D projective images of 4-D mathematical objects.

Disparity estimation is a highly complex and time consuming process in multiview video encoders. Since multiple views taken from a two-dimensional camera array need to be coded at every time instance, the complexity of the encoder plays an important role besides its rate-distortion performance. In previous papers we have introduced a new frame type called the D (derived) frame that exploits the strong geometrical correspondence between views, thereby reducing the complexity of the encoder. By employing the D frames instead of some of the P frames in the prediction structure, significant complexity gain can be achieved if the threshold value, which is a keystone element to adjust the complexity at the cost of quality and/or bit-rate, is selected wisely. A new adaptive method to calculate the threshold value automatically from existing information during the encoding process is presented. In this method, the threshold values are generated for each block of each D frame to increase the accuracy. The algorithm is applied to several image sets and 20.6% complexity gain is achieved using the automatically generated threshold values without compromising quality or bit-rate.

Architectural distortion is an important sign of early breast cancer. Due to its subtlety, it is often missed during screening. We propose a method to detect architectural distortion in prior mammograms of interval-cancer cases based on statistical measures of oriented patterns. Oriented patterns were analyzed in the present work because regions with architectural distortion contain a large number of tissue structures spread over a wide angular range. Two new types of cooccurrence matrices were derived to estimate the joint occurrence of the angles of oriented structures. Statistical features were computed from each of the angle cooccurrence matrices to discriminate sites of architectural distortion from falsely detected regions in normal parts of mammograms. A total of 4,224 regions of interest (ROIs) were automatically obtained from 106 prior mammograms of 56 interval-cancer cases and 52 mammograms of 13 normal cases with the application of Gabor filters and phase portrait analysis. For each ROI, Haralick's 14 features were computed using the angle cooccurrence matrices. The best result obtained in terms of the area under the receiver operating characteristic (ROC) curve with the leave-one-patient-out method was 0.76; the free-response ROC curve indicated a sensitivity of 80% at 4.2 false positives per patient.

Although two-dimensional principal component analysis (2DPCA) extracts image features directly from 2D image matrices rather than one dimensional vectors, 2DPCA is only based on the whole images to preserve total variances by maximizing the trace of feature covariance matrix. Thus, 2DPCA cannot extract localized components, which are usually important for face recognition. Inspired by nonnegative matrix factorization (NMF), which is based on localized features, we propose a novel algorithm for face recognition called two-dimensional nonnegative principal component analysis (2DNPCA) to extract localized components and maintain the maximal variance property of 2DPCA. 2DNPCA is a matrix-based algorithm to preserve the local structure of facial images and has the nonnegative constraint to learn localized components. Therefore, 2DNPCA has both advantages of 2DPCA and NMF. Furthermore, 2DNPCA solves the time-consuming problem by removing the restriction of minimizing the cost function and extracting only the base matrix. The nearest neighbor (NN) classifier and linear regression (LR) classifier are used for classification and extensive experimental results show that 2DNPCA plus NN and 2DNPCA plus LR are both very efficient approaches for face recognition.

We present a nondestructive optical technique for quantitatively imaging a focused ultrasonic field in continuous-wave mode within clear media. We used a Nomarski interferometer configuration coupled with a couple charge device array and multichannel lock-in detection. A two-dimensional full-field image of pressure amplitude can be obtained in only one acquisition. This method was experimentally demonstrated with a high-frequency ultrasonic transducer and experimental measurements of acoustic pressure obtained at 27 MHz. This technique is very interesting because it does not introduce any disturbance in the ultrasonic field and does not need any scanning mechanisms to acquire images. It can be useful for probing high-frequency acoustic fields and for calibrating HF-piezo-electric transducers.

Reversible data hiding deals with the insertion of auxiliary information into a host data without causing any permanent degradation to the original signal. In this contribution a high capacity reversible data hiding scheme, based on the classical difference expansion insertion algorithm, is presented. The method exploits a prediction stage, followed by prediction errors modification, both in the spatial domain and in the S-transform domain. Such two step embedding allows us to achieve high embedding capacity while preserving a high image quality, as demonstrated in the experimental results.

We present a new formulation to solve a defect detection problem on images using multiple reference images. The reference images are defect-free images obtained from the same position of other products. The defect detection problem is reformulated as a binary labeling problem, where each pixel is labeled with "one" if it contains a defect and with "zero" otherwise. The formulation of the energy function used for the labeling problem is defined. Then, the graph-cuts algorithm is used to obtain the optimal label set minimizing the energy function that becomes the defect detection result. The presented approaches are robust to noises taken from several sources, including image-taking, transmission process, environmental lighting, and pattern variation. It does not suffer from the alignment problem for the conventional comparison methods using references. These approaches are illustrated with real data sets, semiconductor wafer images collected by scanning electron microscope equipment, and compared to other defect detection approach.

From the texture analysis of image content, we propose a steganalytic method to detect spatial domain steganography in grayscale images. First of all, based on the local linear vectors, which are selected carefully and sensitive to image texture, images are decomposed into several textural detail subbands by the local linear transform (LLT). Then the statistical distribution of the LLT coefficient is modeled by using the generalized Gaussian distribution. Finally, novel textural features of the LLT coefficient histogram and cooccurrence matrix are extracted for steganalyzers implemented by the support vector machine. Extensive experiments are performed on four diverse uncompressed image databases and seven typical spatial domain steganographic algorithms, such as the highly undetectable stego. The results reveal that the proposed scheme is universal for detecting spatial domain steganography. By comparison with other well-known feature sets, our presented feature set offers the best performance under most circumstances.

We present a methodology to calculate the color appearance of advertising billboards set in indoor and outdoor environments, printed on different types of paper support and viewed under different illuminations. The aim is to simulate the visual appearance of an image printed on a specific support, observed in a certain context and illuminated with a specific source of light. Knowing in advance the visual rendering of an image in different conditions can avoid problems related to its visualization. The proposed method applies a sequence of transformations to convert a four channels image (CMYK) into a spectral one, considering the paper support, then it simulates the chosen illumination, and finally computes an estimation of the appearance.

We present a new framework for an interactive image delineation technique, which we term as interactive texture-snapping system (IT-SNAPS). One of the unique features of IT-SNAPS stems from the fact that it can effectively aid the user in accurately segmenting images with complex texture, without placing undue burden on the user. This is made possible through the formulation of IT-SNAPS, which enables it to be user-aware, i.e., it unobtrusively elicits information from the user during the segmentation process, and hence, adapts itself on-the-fly to the boundary being segmented. In addition to generating an accurate segmentation, it is shown that the framework of IT-SNAPS allows for extraction of useful information post-segmentation, which can potentially assist in the development of customized automatic segmentation algorithms. The afore mentioned features of IT-SNAPS are demonstrated on a set of texture images, as well as on a real-world biomedical application. Using appropriate segmentation protocols in conjunction with expert-provided ground truth, experiments are designed to quantitatively evaluate and compare the segmentation accuracy and user-friendliness of IT-SNAPS with another popular interactive segmentation technique. Promising results indicate the efficacy of IT-SNAPS and its potential to positively impact a broad spectrum of computer vision applications.

We propose an image encryption technique for color images by using conventional logarithms operator, chaotic map and 4-weighted fractional Fourier transform (4-WFRFT). The color image to be encrypted is initially separated into its primary color channels, red, green, and blue; each channel is concealed inside the same random mask by using the base-changing axiom of logarithms. Subsequently, each logarithmic-concealed channel is encrypted independently by using double random 4-WFRFT plane encoding technique. This technique utilizes two random phase mask that are developed by using a chaotic map. The parameters of RPMs and transform orders of 4-WFRFT utilized in each logarithm channel are considered as secret keys for encryption/decryption process. The robustness of the proposed technique is analyzed and discussed by presenting the experimental results.

Bidimensional empirical mode decomposition (BEMD) technique decomposes an image into several bidimensional intrinsic mode functions (BIMFs) and a bidimensional residue (BR). Classical BEMD methods require some form of surface interpolation to estimate envelope surfaces, which causes various problems. On the other hand, existing surface interpolation-based BEMD methods are proposed and/or are suitable for gray-scale images only. This paper presents a novel BEMD approach for color images known as color BEMD (CBEMD), which employs order-statistics filter (OSF)-based envelope estimation to avoid some of the difficulties, otherwise encountered in classical BEMD approaches. The CBEMD can decompose a color image into several color BIMFs and a color BR based on hierarchical local spatial variation of image intensity and color. Since the color BR represents the trend of a color image in terms of global intensity and color distribution, it is utilized for trend adjustment of color images. Both formulation of the CBEMD and finding and adjusting the color image trend are known to be the first approaches in corresponding issues. Experimental results with several real images demonstrate the potential of the proposed CBEMD method for color image processing, which include trend adjustment of color images.

We present a new calibration model of both static distortion and projection for a scanning-electron microscope (SEM). The proposed calibration model depends continuously on the magnification factor. State-of-the-art methods have proposed models to solve the static distortion and projection model but for a discrete set of low and high magnifications: at low magnifications, existing models assume static distortion and perspective projection. At high magnifications, existing models assume an orthogonal projection without presence of static distortion. However, a magnification-continuous model which defines continuous switch from low to high magnifications has not yet been proposed. We propose a magnification-continuous static calibration model of the SEM. The static distortion and intrinsics of the projection matrix are modeled by partial differential equations (PDEs) with respect to magnification. The approach is applied with success to the JEOL-JSM 820 in a secondary electron imaging mode for magnification ranging from 100× to 10k×. The final RMS reprojection error is about 0.9 pixels. This result together with two application-based experiments: the consistent measurements of the bending of a cantilever and a 3-D reconstruction of a nano-ball emphasize the relevance of the proposed approach.

A method for preferred color reproduction with a psychophysical study based upon a real-life scene demonstrated application is presented. A color correction matrix optimization algorithm is first introduced which applies additional hue constraints on top of the CIE delta E magnitude error to emphasize the importance of hue in the subjective quality of color reproduction. An application is implemented by applying matrices optimized for the three dominant memory colors: skin, sky and green based on each pixel's categorical classification which is obtained through an explicitly defined boundary model. A psychophysical study was carried out to systematically investigate observers' hue preference in terms of the reproduction of individual memory colors as well as when multiple memory colors exist in the scene simultaneously. Conclusions suggest that observers' preference for individual memory colors is consistent with regard to scene composition. When a scene with multiple memory colors is evaluated, skin is given the highest priority in the overall preference scale, followed by sky and then green. In practice, the optimized preferred color reproduction may be achieved by first locating the matrices for individual memory colors through a rank order study, and then applying them based on each pixel's classification.

The problem of illumination estimation for color constancy and automatic white balancing of digital color imagery can be viewed as the separation of the image into illumination and reflectance components. We propose using nonnegative matrix factorization with sparseness constraints to separate these components. Since illumination and reflectance are combined multiplicatively, the first step is to move to the logarithm domain so that the components are additive. The image data is then organized as a matrix to be factored into nonnegative components. Sparseness constraints imposed on the resulting factors help distinguish illumination from reflectance. The proposed approach provides a pixel-wise estimate of the illumination chromaticity throughout the entire image. This approach and its variations can also be used to provide an estimate of the overall scene illumination chromaticity.

This paper presents a method to measure dynamic deformable three-dimensional (3-D) surfaces from a calibrated stereo image sequence. This proposed method adopts a patch-based expansion technique that performs 3-D measuring and tracking by computing the motion and disparity independently. Compared with traditional expansion techniques, the proposed expansion technique always spreads from the most reliable region, which makes the measurement more accurate and dense. The performance is evaluated on simulated experiments, and the effectiveness is demonstrated by different real surfaces.

The projection axes produced by conventional locality preserving projection (LPP) are not orthogonal though many dimension reduction methods favor the use of orthogonal projection axes. Orthogonal LPP (OLPP) has been found to perform well in document indexing but suffers from a much higher computational complexity than conventional LPP. This is somewhat because current OLPP algorithm must solve the same number of eigen equations as the number of required projection axes. In contrast, conventional LPP obtains all of the projection axes by solving just one eigenequation. A further drawback of current OLPP algorithm is that, since it requires a number of matrix operations, it also produce more rounding errors than conventional LPP. Four main theoretical contributions are presented. First, a new, more computationally efficient algorithm for implementing OLPP is proposed. Second, for the first time the solution property of conventional LPP is shown. Third, another form of current OLPP algorithm is described. Finally, it is shown that if the projection axes of conventional LPP and the OLPP are paired in sequence, the projection axis of the OLPP has a greater ability to preserve locality than the paired projection axis of conventional LPP.

Simultaneous denoizing and segmentation of ink-jet printing images are challenging tasks for ink-jet printing pattern analysis. Thus, a unified denoizing-segmentation algorithm is proposed for ink-jet printing images. The energy functional for the proposed algorithm consists of three terms: the Beltrami flow, the phase field, and the shape prior. The Beltrami flow is used to enhance image features while preserving natural fine structures and the phase field model is introduced to extract specific pattern structures within ink-jet printing images. The shape prior term for the deformable framework through a nonlinear energy representation is designed to attract different shapes towards the Beltrami flow and the phase field at given directions. A fuzzy optimization method, called multi-start fuzzy optimization method (MSFOM), is also proposed to numerically solve the unified denoizing-segmentation model. MSFOM is a hybrid algorithm, combining fuzzy logic and genetic optimization, which is able to find the suboptimal global minimum with a low computational cost. Experimental results demonstrate that the proposed algorithm is better than the existing schemes, and it offers effective noise removal in noisy ink-jet printing images while maintaining fine structures of patterns.

Image noise removal is the first step in image processing, pattern recognition, and computer vision. A novel algorithm is proposed to remove noise on images based on anisotropic diffusion and subpixel approaches. Firstly, the subpixel difference of an image is defined and the Euler-Lagrange equation is solved. Then, the diffusion equation is solved numerically using an iterative approach. Finally, the noise is removed after the diffusion procedure is finished. The experiments show that the proposed algorithm yields better signal-to-noise ratio and has no blocky effect and less generated speckle noise in the results than the other methods do. In addition, it is easy to implement, takes less iterations, and has low computational complexity.

Although single image resolution enhancement, otherwise known as super-resolution, is widely regarded as an ill-posed inverse problem, we re-examine the fundamental relationship between a high-resolution (HR) image acquisition module and its low-resolution (LR) counterpart. Analysis shows that partial HR information is attenuated but still exists, in its LR version, through the fundamental averaging-and-subsampling process. As a result, we propose a modified Laplacian filter (MLF) and an intensity correction process (ICP) as the pre and post process, respectively, with an interpolation algorithm to partially restore the attenuated information in a super-resolution (SR) enhanced image image. Experiments show that the proposed MLF and ICP provide significant and consistent quality improvements on all 10 test images with three well known interpolation methods including bilinear, bi-cubic, and the SR graphical user interface program provided by Ecole Polytechnique Federale de Lausanne. The proposed MLF and ICP are simple in implementation and generally applicable to all average-subsampled LR images. MLF and ICP, separately or together, can be integrated into most interpolation methods that attempt to restore the original HR contents. Finally, the idea of MLF and ICP can also be applied for average, subsampled one-dimensional signal.

We report a system for performing critical-dimension (CD) measurements of glass panels that uses a substepping system to generate a sequence of lower-resolution images and a fast, edge-directed image reconstruction algorithm to combine these images into a higher-resolution image. A large working distance and large aperture of microscope objective is required in glass panel manufacturing, to measure very small distances at high-level repeatability in a short time, which in turn allows only low magnification objectives. Low-resolution images are obtained when the camera of the CD measurement system is moved at step intervals smaller than the normal pixel size of the camera sensor. We propose a fast, edge-directed image registration (IR) algorithm to find the subpixel accuracy information for full-size images to be registered. The number of processed pixels is only about 5% to 10% of the number of pixels in the image, and the algorithm runs noniteratively. Thus, the subpixel IR algorithm is faster than other methods. In addition, a weighting calculation method for fast and robust edge-directed image interpolation algorithm is proposed to form a high-resolution image. Our experimental results prove that the proposed method offers faster processing time than the standard process and acceptable repeatability of CD measurements.

Abstract unavailable.

Abstract unavailable.

A book review of a recently released title.