Search

Disease severity prediction is essential in clinical diagnosis nowadays, as correct understandings of the onset and progression of disease are priceless in treatment planning. In this study, dementia disease, one of the most severe non-communicable diseases worldwide, is focused. A novel dementia disease severity prediction scheme is proposed using new big ranking and learning techniques. To be specific, arterial spin labeling, an emerging functional-magnetic resonance imaging technique, is adopted to provide image-based clinical data. There are two steps composed of the whole scheme. First, a single-pixel based method is presented to correct the partial volume effect in arterial spin labeling images. The advantage of this method is that, problems of blurring and brain detail loss can be well tackled. Second, novel big pair-wise ranking and learning techniques is proposed to realize the dementia disease severity prediction task using arterial spin labeling images after partial volume effects correction. Extensive experiments using a big database composed of images acquired from 350 real demented patients are carried out with several conventional methods being compared. Comprehensive statistical analysis is performed and it suggests that the new scheme is promising in dementia disease severity prediction. [ABSTRACT FROM AUTHOR]

In this paper, we proposes high payload reversible data hiding scheme using weighted matrix. Generate cover image by enlarge the size of original image using interpolation. Partition the original image into (3 × 3) pixel blocks and cover image into (5 × 5) blocks. Perform modular sum of entry-wise multiplication of an (3 × 3) original pixel block with same size shared predefined weighted matrix W . At most twelve multiplication operations are performed to embed 48 bits secret data within an (5 × 5) overlapped pixel block. In each operation, the data embedding position are identified and stored the positional value by modifying three least significant bits of the interleaved pixel of cover image to hide large amount of secret data. For i -th block i = 1, 2, …, we update the weighted matrix W i +1 as W i +1 = ( W i × κ − 1) mod 9, where gcd ( κ , 9) = 1. The original pixels are not effected during data embedding in our scheme which assure reversibility. The proposed scheme provides average embedding payload 2.97 bits per pixel (bpp) with good visual quality measured by peak signal to noise ratio (PSNR) guaranteed to be higher than 37.97 dB. Finally, we compare our scheme with other state-of-the-art methods and obtain reasonably better performance in terms of data embedding capacity. [ABSTRACT FROM AUTHOR]

Many saliency models consider the feature extraction as the algorithmic core and the performance of their methods relies on the selection of the features to a great extent. However, there can hardly be a set of features effective to pop out the salient regions under various visual environments. Moreover, because saliency is not tuned to certain visual features, a location winning the spatial competition in any feature space can be defined as salient. Instead of seeking for or learning the features to highlight the difference between the salient areas and the background, we focus more on the sparsity and uniqueness carried by the original image itself, the source of all the features, to propose a nonlocal reconstruction-based saliency model. In the proposed approach, the saliency is measured by the sparse reconstruction residual of representing the central patch with a linear combination of its surrounding patches sampled in a nonlocal manner. In addition, this is generalized to model the global aspect saliency, which provides a complement to the nonlocal saliency and improves the performance further. As a generalization of Itti et al.׳s classical center–surround comparison scheme, the proposed approach performs well on images where Itti et al.׳s method fails, as well as on general natural images. Numerical experiments show the proposed approach produces better results compared with the state-of-the-art algorithms on three public databases. [ABSTRACT FROM AUTHOR]

Computer generated holograms are often exploited to implement optical encryption schemes. This paper proposes the integration of dynamic visual cryptography (an optical technique based on the interplay of visual cryptography and time-averaging geometric moiré) with Gerchberg–Saxton algorithm. A stochastic moiré grating is used to embed the secret into a single cover image. The secret can be visually decoded by a naked eye if only the amplitude of harmonic oscillations corresponds to an accurately preselected value. The proposed visual image encryption scheme is based on computer generated holography, optical time-averaging moiré and principles of dynamic visual cryptography. Dynamic visual cryptography is used both for the initial encryption of the secret image and for the final decryption. Phase data of the encrypted image are computed by using Gerchberg–Saxton algorithm. The optical image is decrypted using the computationally reconstructed field of amplitudes. [ABSTRACT FROM AUTHOR]

Abstract: Anisotropic blur and mis-registration frequently happen in multi-focus images due to object or camera motion. These factors severely degrade the fusion quality of multi-focus images. In this paper, we present a novel multi-scale weighted gradient-based fusion method to solve this problem. This method is based on a multi-scale structure-based focus measure that reflects the sharpness of edge and corner structures at multiple scales. This focus measure is derived based on an image structure saliency and introduced to determine the gradient weights in the proposed gradient-based fusion method for multi-focus images with a novel multi-scale approach. In particular, we focus on a two-scale scheme, i.e., a large scale and a small scale, to effectively solve the fusion problems raised by anisotropic blur and mis-registration. The large-scale structure-based focus measure is used first to attenuate the impacts of anisotropic blur and mis-registration on the focused region detection, and then the gradient weights near the boundaries of the focused regions are carefully determined by applying the small-scale focus measure. Experimental results clearly demonstrate that the proposed method outperforms the conventional fusion methods in the presence of anisotropic blur and mis-registration. [Copyright &y& Elsevier]

Highlights: [•] We propose genetic algorithm-based image steganography approach in spatial domain. [•] The stego image quality of the proposed method is tunable. [•] Our scheme is simple, and feasible for adaptive steganographic applications. [•] Obtained results are satisfactory from the aspect of PSNR and capacity. [Copyright &y& Elsevier]

Abstract: This paper aims at introducing a novel efficient approach for high-quality and fast image restoration by combining both a greedy strategy and a global optimization strategy, based on a pyramidal representation of the image. A coarse version of the input image is first restored by exemplar-based method using a greedy strategy. From the low-resolution inpainted image, higher resolutions are interpolated and refined by a global optimization strategy. Experimental results on natural images demonstrate the effectiveness of the proposed method. Moreover, a comparison with some methods of the state-of-the-art confirms its superiority in terms of image quality and computational time. [Copyright &y& Elsevier]

Abstract: In this paper, we propose a steganographic scheme based on the varieties of coefficients of discrete cosine transformation of an image. The major problem of hiding data in the high-frequency coefficients of discrete cosine transformation is that rounding errors will be added into the spatial-domain image, and thus cannot be transformed back to the correct modified coefficients of the discrete cosine transformation. To solve this problem, we use integer mapping to implement our discrete cosine transformation. Thus, the image recovered from the modified coefficients can be transformed again to the correct data hidden coefficients. [Copyright &y& Elsevier]

Accurate boundary detection is a critical step of the image segmentation process. While most edge detectors rely on the presence of strong intensity gradients, this criteria can limit robustness in many real world cases. In this work we propose a scheme which makes use of a combination of both global and local segmentation methods to capture the boundary of target objects in low contrast images. This approach has several advantages: the globally convex segmentation scheme is immune from initial conditions and is easily adapted to the data. The addition of a segmentation scheme based on local curve evolution produces a solution which is shown to help preserve topology between the initial and target shape, a property lacking in globally convex segmentation schemes. Experimental results show that the proposed method achieves enhanced performance compared to classical data-driven segmentation schemes proposed in the literature. [ABSTRACT FROM AUTHOR]

Abstract: We describe how oriented Basic Image Feature Columns (oBIF Columns) can be used for writer identification and how this texture-based scheme can be enhanced by encoding a writer's style as the deviation from the mean encoding for a population of writers. We hypothesise that this deviation, the Delta encoding, provides a more informative encoding than the texture-based encoding alone. The methods have been evaluated using the IAM dataset and by making entries to two top international competitions for assessing the state-of-the-art in writer identification. We demonstrate that the oBIF Column scheme on its own is sufficient to gain a performance level of 99% when tested using 300 writers from the IAM dataset. However, on the more challenging competition datasets, significantly improved performance was obtained using the Delta encoding scheme, which achieved first place in both competitions. In our characterisation of the Delta encoding, we demonstrate that the method is making use of information contained in the correlation between the written style of different textual elements, which may not be used by other methods. [Copyright &y& Elsevier]

Abstract: Using the original and ‘symmetrical face’ training samples to perform representation based face recognition was first proposed in [1]. It simultaneously used the original and ‘symmetrical face’ training samples to perform a two-step classification and achieved an outstanding classification result. However, in [1] the “symmetrical face” is devised only for one method. In this paper, we do some improvements on the basis of [1] and combine this “symmetrical faces” transformation with several representation based methods. We exploit all original training samples, left “symmetrical face” training samples and right “symmetrical face” training samples for classification and use the score fusion for ultimate face recognition. The symmetry of the face is first used to generate new samples, which is different from original face image but can really reflect some possible appearance of the face. It effectively overcomes the problem of non-sufficient training samples. The experimental results show that the proposed scheme can be used to improve a number of traditional representation based methods including those that are not presented in the paper. [Copyright &y& Elsevier]

Abstract: In recent years, a number of chaos-based image cryptosystems have been proposed to meet the increasing demand for real-time secure image transmission. In this paper, an improved diffusion scheme named continuous diffusion strategy is proposed to promote the efficiency of the conventional permutation–diffusion type image cipher. The new scheme contains a supplementary diffusion procedure after the conventional diffusion process and the control parameters are altered by the cipher image after the first diffusion procedure. As a result, the difference can be introduced at the beginning and spread out to the whole image, and hence the same level of security can be achieved with fewer overall rounds. Moreover, to further enhance the confusion effect of the diffusion operation, an intensive diffusion approach is proposed, using stretched key stream elements to perform a cyclic shift to the cipher pixels. Extensive cryptanalysis has been performed using differential analysis, key space analysis, key sensitivity analysis and various statistical analyses. Experiment results demonstrate that the new scheme has a high level of security and fast encryption speed for practical image encryption. [Copyright &y& Elsevier]

Highlights: [•] Introduced the combination operation on permutations and proved its properties. [•] Proposed an effective method to generate large permutation based on combination. [•] Proposed an image encryption scheme based on large permutation and diffusion. [•] The proposed encryption scheme has high security and significant speed advantage. [Copyright &y& Elsevier]

Abstract: Current energy-saving color design approaches can be classified into two categories, namely, context-aware dimming and color remapping. The former darkens individual regions with respect to the user interactions, and the latter replaces the color set with a new color set that yields lower energy consumption. Both schemes have drawbacks: color dimming tends to cause loss of perceptual quality, and color remapping is an offline color design process. This paper introduces a novel saliency-guided color dimming scheme for OLED displays in both the context of 3D visualization and 2D visualization. The key idea is to eliminate undesired details while enhancing the visually salient features of each frame on-the-fly by leveraging the color and spatial information. A parallelizable image-space salient region detection algorithm is introduced to make the entire process GPU-friendly and real-time. We apply our approach on several representative visualization scenarios and conduct a preliminary user study. Experimental results demonstrate the effectiveness, efficiency, and quality of our approach. [Copyright &y& Elsevier]

Abstract: In this paper, an efficient mapping model based on singular value decomposition (SVD) is proposed for face hallucination. We can observe and prove that the main singular values of an image at one resolution have approximately linear relationships with their counterparts at other resolutions. This makes the estimation of the singular values of the corresponding high-resolution (HR) face images from a low-resolution (LR) face image more reliable. From the signal-processing point of view, this can effectively preserve and reconstruct the dominant information in the HR face images. Interpolating the other two matrices obtained from the SVD of the LR image does not change either the primary facial structure or the pattern of the face image. The corresponding two matrices for the HR face images can be constructed in a “coarse-to-fine” manner using global reconstruction. Our proposed method retains the holistic structure of face images, while the learned mapping matrices, which are represented as embedding coefficients of the individual mapping matrices learned from LR-HR training pairs, can be seen as holistic constraints in the reconstruction of HR images. Compared to state-of-the-art algorithms, experiments show that our proposed face-hallucination scheme is effective in terms of producing plausible HR images with both a holistic structure and high-frequency details. [Copyright &y& Elsevier]

This study shows that a recently proposed image watermarking scheme in the paper “Optimized gray-scale image watermarking using DWT-SVD and Firefly Algorithm” ( Agarwal, Mishra, Sharma, & Bedi, 2014 ) has a fundamental flaw in its design. The extracted watermark is not the embedded watermark, actually it is determined by the reference watermark, which leads to false positive detection problem. Hence, it invalidates the objective of the scheme and cannot be used for the protection of rightful ownership. The weakness of the scheme designing is illustrated with an example. [ABSTRACT FROM AUTHOR]