International Journal of Computer and Applications

In the real world, a realistic setting for computer vision or multimedia recognition problems is that we have some classes containing lots of training data and many classes contain a small amount of training data. Therefore, how to use frequent classes to help learning rare classes for which it is harder to collect the training data is an open question. Learning with Shared Information is an emerging topic in machine learning, computer vision and multimedia analysis. There are different level of components that can be shared during concept modeling and machine learning stages, such as sharing generic object parts, sharing attributes, sharing transformations, sharing regularization parameters and sharing training examples, etc. Regarding the specific methods, multi-task learning, transfer learning and deep learning can be seen as using different strategies to share information. These learning with shared information methods are very effective in solving real-world large-scale problems. This special issue aims at gathering the recent advances in learning with shared information methods and their applications in computer vision and multimedia analysis. Both state-of-the-art works, as well as literature reviews, are welcome for submission. Papers addressing interesting real-world computer vision and multimedia applications are especially encouraged. Topics of interest include, but are not limited to:  • Multi-task learning or transfer learning for large-scale computer vision and multimedia analysis • Deep learning for large-scale computer vision and multimedia analysis • Multi-modal approach for large-scale computer vision and multimedia analysis • Different sharing strategies, e.g., sharing generic object parts, sharing attributes, sharing transformations, sharing regularization parameters and sharing training examples, • Real-world computer vision and multimedia applications based on learning with shared information, e.g., event detection, object recognition, object detection, action recognition, human head pose estimation, object tracking, location-based services, semantic indexing. • New datasets and metrics to evaluate the benefit of the proposed sharing ability for the specific computer vision or multimedia problem. • Survey papers regarding the topic of learning with shared information.  Authors who are unsure whether their planned submission is in scope may contact the guest editors prior to the submission deadline with an abstract, in order to receive feedback.

IEEE transactions on pattern analysis and machine intelligence

Fast Fourier Transform

Computed tomography (CT) imaging is an essential tool in various clinical diagnoses and radiotherapy treatment planning. Since CT image intensities are directly related to positron emission tomography (PET) attenuation coefficients, they are indispensable for attenuation correction (AC) of the PET images. However, due to the relatively high dose of radiation exposure in CT scan, it is advised to limit the acquisition of CT images. In addition, in the new PET and magnetic resonance (MR) imaging scanner, only MR images are available, which are unfortunately not directly applicable to AC. These issues greatly motivate the development of methods for reliable estimate of CT image from its corresponding MR image of the same subject. In this paper, we propose a learning-based method to tackle this challenging problem. Specifically, we first partition a given MR image into a set of patches. Then, for each patch, we use the structured random forest to directly predict a CT patch as a structured output, where a new ensemble model is also used to ensure the robust prediction. Image features are innovatively crafted to achieve multi-level sensitivity, with spatial information integrated through only rigid-body alignment to help avoiding the error-prone inter-subject deformable registration. Moreover, we use an auto-context model to iteratively refine the prediction. Finally, we combine all of the predicted CT patches to obtain the final prediction for the given MR image. We demonstrate the efficacy of our method on two datasets: human brain and prostate images. Experimental results show that our method can accurately predict CT images in various scenarios, even for the images undergoing large shape variation, and also outperforms two state-of-the-art methods.

Estimating CT Image From MRI Data Using Structured Random Forest and Auto-Context Model

Nearly all nontrivial real-world systems are nonlinear dynamical systems. Chaos describes certain nonlinear dynamical systems that have a very sensitive dependence on initial conditions. Chaotic systems are always deterministic and may be very simple, yet they produce completely unpredictable and divergent behavior. Systems of nonlinear equations are difficult to solve analytically, and scientists have relied heavily on visual and qualitative approaches to discover and analyze the dynamics of nonlinearity. Indeed, few fields have drawn as heavily from visualization methods for their seminal innovations: from strange attractors, to bifurcation diagrams, to cobweb plots, to phase diagrams and embedding. Although the social sciences are increasingly studying these types of systems, seminal concepts remain murky or loosely adopted. This article has three aims. First, it argues for several visualization methods to critically analyze and understand the behavior of nonlinear dynamical systems. Second, it uses these visualizations to introduce the foundations of nonlinear dynamics, chaos, fractals, self-similarity and the limits of prediction. Finally, it presents Pynamical, an open-source Python package to easily visualize and explore nonlinear dynamical systems' behavior.

Visual Analysis of Nonlinear Dynamical Systems: Chaos, Fractals, Self-Similarity and the Limits of Prediction

Numerous algorithms have been proposed in the fields of visual cryptography and steganography with the goals of improving security, reliability, and efficiency. Visual cryptography involves dividing the image into random shares without the use of keys. Since it can be employed by anyone without any cryptographic knowledge and does not require any computations while decrypting, many researches' have been focused on it. Steganography is the art, science, or practice in which messages, images, or files are hidden inside other messages, images, or files. The person sending the hidden data and the person meant to receive the data are the only ones who know about it; but to everyone else, the object containing the hidden data just seems like an everyday normal object. Ultimately they both are ways of hiding data from prying eyes. This paper proposes a method to combine these two strong methodologies of keeping the image and the data secret from the prying eyes. We can develop an algorithm which uses the LSB method of steganography to hide text data inside an image, and then use that image as a secret image for visual cryptography method. The secret image with hidden data will be split up into shares using RKO technique for Visual Cryptography. Then when these shares are re-assembled or decoded to reconstruct the original image the revealed image will still contains the hidden data. The receiver then extracts the hidden data from the revealed image using reverse LSB.

Steganography Combined with RKO Technique for Visual Cryptography

Colorization is the term which means to color any gray scale image. In digital image processing terms colorization of grayscale image is nothing but assigning the red, green and blue color component values to each gray value of grayscale image [1, 2]. Colorization is very difficult because it involves assigning three-dimensional (RGB) pixel values to an image which varies along only one dimension (luminance or intensity) [3]. Since different colors may have the same luminance value but vary in hue or saturation, the problem of colorizing grayscale images has no inherently ‛unique’ solution [4]. So user intervention becomes trivial in colorization process to select the ‛better’ match (red, green and blue values) for every grayscale pixel. In [1, 2] authors have proposed some of the simple approaches for colorization of grayscale images, where user intervention is needed only to select the source color image to be used to generate color palette. In [1, 5] different color spaces and pixel window sizes are worked out for coloring grayscale images. All these techniques gives the results subjective to the source color image considered for coloring are very heavy with respect to time complexity. To speed up the process different search algorithms are proposed in [1, 4]. But everywhere the size of source color image is assumed to be equal to or more than to be colored target grayscale image. The paper presents novel colorization technique where this size dependency of source color image and target grayscale image are taken out. The proposed technique generates color palette using vector quantization codebook generation approach. Here Kekre’s fast codebook generation (KFCG) [16, 20, 21] algorithm is used.

Kekre’s fast codebook generation in VQ with various color spaces for colorization of grayscale images

The paper presents the reduction of feature vector size of the image by using Wavelet Pyramids based image retrieval techniques for Walsh Transform. The colour averaging methods like row & column mean (RCM), forward diagonal mean (FDM) and row column & forward diagonal mean (RCFDM) are applied on image wavelets generated at four levels of decomposition. The proposed content based image retrieval (CBIR) techniques are tested on a generic image database having 1000 images spread across 11 categories. For each proposed CBIR technique 55 queries (randomly selected 5 per category) are fired on the image database. To compare the performance of image retrieval techniques average precision and recall values are computed for all queries. When these results are compared with the colour averaging based image retrieval techniques applied on the original image itself, it has been observed that the precision recall crossover value for wavelet pyramid level 1 is almost same (up to 3 decimal places) for FDM and RCFDM. However the size of the feature vector in the proposed CBIR methods is significantly less than the original image. Thus the proposed CBIR methods prove to be better in terms of reduced computational complexity. In the discussed image retrieval methods, Walsh wavelet pyramid level 1 for RCFDM gives the highest performance as indicated by the precision recall crossover point.

/pdf/reduction-in-feature-vector-size-of-colour-averaging-based-1i02klg4d6.pdf

Reduction in Feature Vector Size of Colour Averaging based Image Retrieval Techniques using Walsh Wavelet Pyramid Levels

This paper presents a new blind secure video watermarking algorithm using DCT (discrete cosine transform). In this proposed algorithm cover video is divided into frames and watermark is inserted into selected frames. For selected video frames two-dimensional 8×8 discrete cosine transform is carried out on luminance component. Finally binary watermark is embedded into mid frequency DC coefficients by adjusting coefficients DCT(4,3) and DCT(5,2). Experimental results shows that the proposed algorithm is imperceptible as well as robust against wide variety of signal and video processing attacks like Gaussian noise, Salt -pepper noise, Gaussian filter, Median filter, Histogram Equalization etc. The technique is fairly acceptable and watermarked video is of good quality, achieves high PSNR. The technique shows efficient extraction of watermark with NC value of the retrieved watermark as 1.

/pdf/a-new-dct-based-color-video-watermarking-using-luminance-3bq2y01dlv.pdf

A New DCT based Color Video Watermarking using Luminance Component

This paper proposes simple and novel image compression method using transforms. Discrete cosine Transform (DCT) and Discrete Kekre Transform (DKT) are applied on image individually. Wavelet transform of DCT and DKT is generated using Kekre's algorithm of wavelet generation. Wavelet transform gives better energy compaction than individual transform. It is reflected in reconstructed image quality of wavelet transforms. Further hybridization of two transform is used and hybrid wavelet transform using DKT and DCT is generated and applied on images. It combines properties of both, Kekre transform and DCT. Hybrid wavelet transform gives lesser error than wavelet transform and individual orthogonal transform.

https://ieeexplore.ieee.org/iel7/6937353/6949036/06949245.pdf

Tanuja Sarode

Papers

Steganography Combined with RKO Technique for Visual Cryptography

Kekre’s fast codebook generation in VQ with various color spaces for colorization of grayscale images

Reduction in Feature Vector Size of Colour Averaging based Image Retrieval Techniques using Walsh Wavelet Pyramid Levels

A New DCT based Color Video Watermarking using Luminance Component

Performance superiority of hybrid DKT-DCT wavelet compared to DKT, DCT individual transforms and their wavelets in image compression