Explore the vast range of titles available.

The object contour is a significant cue for identifying and categorising objects. The current work is motivated by indicative researches that attribute object contours to edge information. The spatial continuity exhibited by the edge pixels belonging to the object contour make these different from the noisy edge pixels belonging to the background clutter. In this study, the authors seek to quantify the object contour from a relative count of the adjacent edge pixels that are oriented in the four possible directions, and measure using exponential functions the continuity of each edge over the next adjacent pixel in that direction. The resulting computationally simple, low-dimensional feature set, called as ‘edge continuity features’, can successfully distinguish between object contours and at the same time discriminate intra-class contour variations, as proved by the high accuracies of object recognition achieved on a challenging subset of the Caltech-256 dataset. Grey-to-RGB template matching with City-block distance is implemented that makes the object recognition pipeline independent of the actual colour of the object, but at the same time incorporates colour edge information for discrimination. Comparison with the state-of-the-art validates the efficiency of the proposed approach.

A visual cryptography scheme (VCS) splits the secret image into several shares. Stacking a certain number of shares will reveal the secret image. The principle algorithm (deterministic algorithm) retains the size of shares, and the revealed image is at least double the size of the secret image with good visual quality. In contrast, the probabilistic algorithm has the size of the revealed image equal to that of the secret image with the noisily revealed image. Multi-pixel algorithm cast solutions for the trade off problems between the good visual quality and the non-expansion size of the revealed image. In this paper, we introduce new algorithms for multi-pixel to improve the quality of preceding algorithms, generalize the multi-pixel algorithm, provide evidence of the best form of multi-pixel, and contribute to the thin line problem.

For the robust recognition of noisy face images, in this study, the authors improved the fast neighbourhood component analysis (FNCA) model by introducing a novel spatially smooth regulariser (SSR), resulting in the FNCA-SSR model. The SSR can enforce local spatial smoothness by penalising large differences between adjacent pixels, and makes FNCA-SSR model robust against noise in face image. Moreover, the gradient of SSR can be efficiently computed in image space, and thus the optimisation problem of FNCA-SSR can be conveniently solved by using the gradient descent algorithm. Experimental results on several face data sets show that, for the recognition of noisy face images, FNCA-SSR is robust against Gaussian noise and salt and pepper noise, and can achieve much higher recognition accuracy than FNCA and other competing methods.

The vast majority of computer vision algorithms use some form of optimization, as they intend to find some solution which is “best” according to some criterion. Consequently, the field of optimization is worth studying for everyone being seriously interested in computer vision. In this chapter, some expressions being of widespread use in literature dealing with optimization are clarified first. Furthermore, a classification framework is presented, which intends to categorize optimization methods into the four categories continuous, discrete, combinatorial, and variational, according to the nature of the set from which they select their solution. This categorization helps to obtain an overview of the topic and serves as a basis for the structure of the remaining chapters at the same time. Additionally, some concepts being quite common in optimization and therefore being used in diverse applications are presented. Especially to mention are so-called energy functionals measuring the quality of a particular solution by calculating a quantity called “energy”, graphs, and last but not least Markov Random Fields.

Energy functions consisting of a pixel-wise sum of data-driven energies as well as a sum of terms affected by two adjacent pixels (aiming at ensuring consistency for neighboring pixels) are quite common in computer vision. This kind of energy can be represented well by Markov Random Fields (MRFs). If we have to take a binary decision, e.g., in binary segmentation, where each pixel has to be labeled as “object” or “background,” the MRF can be supplemented by two additional nodes, each representing one of the two labels. The globally optimal solution of the resulting graph can be found by finding its minimum cut (where the sum of the weights of all severed edges is minimized) in polynomial time by maximum flow algorithms. Graph cuts can be extended to the multi-label case, where it is either possible to find the exact solution when the labels are linearly ordered or the solution is approximated by iteratively solving binary decisions. An instance of the max-flow algorithm, binary segmentation, as well as stereo matching and optical flow calculation, which can both be interpreted as multi-labeling tasks, is presented in this chapter. Normalized cuts seeking a spectral, i.e., eigenvalue solution, complete the chapter.

To achieve multiple color images encryption, a secure double-color-image encryption algorithm is designed based on the quaternion multiple parameter discrete fractional angular transform (QMPDFrAT), a nonlinear operation and a plaintext-related joint permutation-diffusion mechanism. QMPDFrAT is first defined and then applied to encrypt multiple color images. In the designed algorithm, the low-frequency and high-frequency sub-bands of the three color components of each plaintext image are obtained by two-dimensional discrete wavelet transform. Then, the high-frequency sub-bands are further made sparse and the main features of these sub-bands are extracted by a Zigzag scan. Subsequently, all the low-frequency sub-bands and high-frequency fusion images are represented as three quaternion signals, which are modulated by the proposed QMPDFrAT with three quaternion random phase masks, respectively. The spherical transform, as a nonlinear operation, is followed to nonlinearly make the three transform results interact. For better security, a joint permutation-diffusion mechanism based on plaintext-related random pixel insertion is performed on the three intermediate outputs to yield the final encryption image. Compared with many similar color image compression-encryption schemes, the proposed algorithm can encrypt double-color-image with higher quality of image reconstruction. Numerical simulation results demonstrate that the proposed double-color-image encryption algorithm is feasibility and achieves high security.

Homogeneous band- or pixel-based feature selection, which exploits the difference between spectral or spatial regions to select informative and low-redundant bands, has been extensively studied in classifying hyperspectral images (HSIs). Although many models have proven effective, they rarely simultaneously exploit homogeneous spatial and spectral information, which are beneficial to extract potential low-dimensional characteristics even under noise. Moreover, the employed vectorial transformation and unordered assumption destroy the implicit knowledge of HSIs. To solve these issues, a dual homogeneous pixel patches-based methodology termed PHSIMR was created for selecting the most representative, low-redundant, and informative bands, integrating hybrid superpixelwise adjacent band grouping and regional informative mutuality ranking algorithms. Specifically, the adjoining band grouping technique is designed to group adjacent bands into connected clusters with a small homogeneous pixel patch containing several homolabeled adjacent spatial points. Hence, the processing is efficient, and the superpixelwise adjoining band grouping can perceptually and quickly acquire connected band groups. Furthermore, the constructed graph and affiliated group avoid vectorial transformation and unordered assumption, protecting spectral and spatial contextual information. Then, the regional informative mutuality ranking algorithm is employed on another larger pixel patch within each homogeneous band group, acquiring the final representative, low-redundant, and informative band subset. Since the employed dual patches consist of homolabeled spatial pixels, PHSIMR is a supervised methodology. Comparative experiments on three benchmark HSIs were performed to demonstrate the efficiency and effectiveness of the proposed PHSIMR.

Offline handwriting recognition is a well-known challenging task in the optical character recognition field due to the difficulty caused by various unconstrained handwriting styles and limited training data. In order to learn invariant feature representations for handwriting, we propose a novel method to incorporate pixel-level rectification into a CNN- and RNN-based recognizer. We also propose an adjacent output mixup method for RNN layer’s training to improve the generalization ability of the recognizer, i.e., the previous output of an RNN layer is added to the current output with random weights. We additionally adopt a series of techniques including pre-training, data augmentation and language model to significantly expand the training data scale, and further analyze their contributions to the improvement in the model performance. The proposed method performs well on four public offline handwriting benchmarks, including the IAM, Rimes, IFN/ENIT and CASIA-HWDB datasets.

This paper presents an efficient medical image indexing and retrieval method using two new proposed feature descriptors named as threshold local binary AND pattern (TLBAP) and local adjacent neighborhood average difference pattern (LANADP). In basic local binary pattern (LBP), every center pixel is considered as a threshold to generate the binary pattern, whereas in the proposed method a threshold value is calculated using the highest pixel intensity of the neighboring pixels to construct the threshold local binary pattern (TLBP). Thereafter, logical AND operation is performed between LBP and TLBP pattern to produce TLBAP pattern. The objective of the other feature descriptor named here as LANADP is to explore the relationship of neighboring pixels with its adjacent neighbors in vertical, horizontal and diagonal directions. In the proposed work, both TLBAP and LANADP features are concatenated in the form of the histograms to generate the final features vector and the performance of the system is evaluated. To test the effectiveness of the proposed method, three publicly available medical image databases, namely OASIS-MRI brain images, NEMA-CT images and VIA/ELCAP-CT images, are used. Two measures, viz. average retrieval precision and average retrieval rate, have been used to evaluate the performance of the method proposed which is further compared with some existing local pattern-based methods. The experimental results show that the proposed methods give better results as compared to the other existing methods considered in this study.