Taha H. Rassem

Block-based discrete wavelet transform-singular value decomposition image watermarking scheme using human visual system characteristics

Digital watermarking has been suggested as a way to achieve digital protection. The aim of digital watermarking is to insert the secret data into the image without significantly affecting the visual quality. This study presents a robust block-based image watermarking scheme based on the singular value decomposition (SVD) and human visual system in the discrete wavelet transform (DWT) domain. The proposed method is considered to be a block-based scheme that utilises the entropy and edge entropy as HVS characteristics for the selection of significant blocks to embed the watermark, which is a binary watermark logo. The blocks of the lowest entropy values and edge entropy values are selected as the best regions to insert the watermark. After the first level of DWT decomposition, the SVD is performed on the low-low sub-band to modify several elements in its U matrix according to predefined conditions. The experimental results of the proposed scheme showed high imperceptibility and high robustness against all image processing attacks and several geometrical attacks using examples of standard and real images. Furthermore, the proposed scheme outperformed several previous schemes in terms of imperceptibility and robustness. The security issue is improved by encrypting a portion of the important information using Advanced Standard Encryption a key size of 192-bits (AES-192).

Completed Local Ternary Pattern for Rotation Invariant Texture Classification

Despite the fact that the two texture descriptors, the completed modeling of Local Binary Pattern (CLBP) and the Completed Local Binary Count (CLBC), have achieved a remarkable accuracy for invariant rotation texture classification, they inherit some Local Binary Pattern (LBP) drawbacks. The LBP is sensitive to noise, and different patterns of LBP may be classified into the same class that reduces its discriminating property. Although, the Local Ternary Pattern (LTP) is proposed to be more robust to noise than LBP, however, the latters weakness may appear with the LTP as well as with LBP. In this paper, a novel completed modeling of the Local Ternary Pattern (LTP) operator is proposed to overcome both LBP drawbacks, and an associated completed Local Ternary Pattern (CLTP) scheme is developed for rotation invariant texture classification. The experimental results using four different texture databases show that the proposed CLTP achieved an impressive classification accuracy as compared to the CLBP and CLBC descriptors.

Object class recognition using combination of color SIFT descriptors

Classifying the unknown image into the correct related class is the aim of the object class recognition systems. Two main points should be kept in mind to implement a class recognition system. Which descriptors that have a higher discriminative power that needs to be extracted from the images? Which classifier can classify these descriptors successfully? The most famous image descriptor is the Scale Invariant Feature Transform (SIFT). Although, SIFT has a high performance, it is partially an illumination invariant. Adding local color information to SIFT descriptors are then suggested to increase the illumination invariant, these descriptors can be called color SIFT descriptors. In this paper, different color SIFT descriptors were implemented to evaluate their performance in the object class recognition systems. This is due to the fact that some descriptors may have a good performance in one class and bad performance in another class at the same time. All possible combinations of these descriptors were used. Some combinations of color SIFT descriptors achieved remarkable classification accuracy. Non linear χ2-kernel support vector machine is used as a learning classifier and bag-of-features representation is used to represent the image features in this paper.

A new reliable optimized image watermarking scheme based on the integer wavelet transform and singular value decomposition for copyright protection

Although image watermarking schemes based on singular value decomposition (SVD) demonstrate high robustness and imperceptibility, they are exposed to the false positive problem (FPP). This drawback mostly occurs when embedding steps depend on singular values while singular vectors are used as secret keys. In this study, a new reliable SVD-based image watermarking scheme that uses integer wavelet transform (IWT) is proposed to overcome FPP and fulfil all watermarking requirements. Unlike in other schemes, the S and V matrices of the watermark are used as secret keys, whereas the S singular vector of the watermark is embedded into the singular values of the host image. The additional secret key is obtained from the watermarked image during the embedding process to increase security and avoid FPP completely. To improve the robustness, as well as achieve balance between robustness and imperceptibility, multi-objective ant colony optimization (MOACO) is utilized to find the optimal scaling factors, namely, multiple zooming factors. Results of the robustness, imperceptibility, and reliability tests demonstrate that the proposed IWT-SVD-MOACO scheme outperforms several previous schemes and avoids FPP completely.

Performance evaluation of RDWT-SVD and DWT-SVD watermarking schemes

Digital image watermarking protects content by embedding a signal (i.e., owner information) into the host image without noticeable degradation in visual quality. To develop any image watermarking scheme, there some important requirements should be achieved such as imperceptibly, robustness, capacity, security, and, etc. Generally, the watermarking scheme based on wavelet transform domain shows an advantage in human perception and good imperceptibility and robustness. Due to this fact, this paper presents two blind image watermarking schemes based on DWT-SVD and RDWT-SVD. To evaluate their performance, these schemes are exposed to different geometric and non-geometric attacks. Although, DWT-SVD and RDWT-SVD showed robust against all attacks, RDWT-SVD is better than DWT-SVD, especially for geometrical attacks.

Performance evaluation of Completed Local Ternary Patterns (CLTP) for medical, scene and event image categorisation

The Completed Local Ternary Pattern descriptor (CLTP) was proposed to overcome the drawbacks of the Local Binary Pattern (LBP). It used for rotation invariant texture classification and demonstrated superior classification accuracy with different types of texture datasets. In this paper, the performance of CLTP for image categorisation is studied and investigated. Different image datasets are used in the experiments such as the Oliva and Torralba datasets (OT8), Event sport datasets, and 2D HeLa medical images. The experimental results proved the superiority of the CLTP descriptor over the original LBP, and different new texture descriptors such as Completed Local Binary Pattern (CLBP) in the image categorisation task. In 2D HeLa medical images, the proposed CLTP achieved the highest state of the art classification rate reaching 95.62%.

New color image histogram-based detectors

Detecting an interest point in the images to extract the features from it is an important step in many computer vision applications. For good performance, these points have to be robust against any transformation that can be done on the images such as viewpoint change, scaling change, rotation, and illumination and, etc. Many of the suggested interest point detectors are measuring the pixel-wise differences in the image intensity or image color. Lee and Chen [1] used image histogram representation instead of pixel representation to detect the interest points. They used the gradient histogram and the RGB color histogram representation. In this work, different color models histogram representation such as Ohta-color histogram, HSV-color histogram, Opponent color histogram and Transformed-color histogram are implemented and used in the proposed interest point detector. These detectors are evaluated by measuring their repeatability and matching score between the detected points in the image matching task and the classification accuracy in the image classification task. It is found that as compared with intensity pixels detectors and Lees histogram detectors, the proposed histogram detectors performed better under some image conditions such as illumination change, blur and some other conditions.

Security analyses of false positive problem for the SVD-based hybrid digital image watermarking techniques in the wavelet transform domain

Singular Value Decomposition (SVD) comprises many important mathematical properties that are useful in numerous applications. Newly developed SVD-based watermarking schemes can effectively maintain minor changes despite the large altered singular values S caused by the attacks. Due to the stability and the properties of S, most of the researchers prefer to embed into S. However, despite satisfying the stability and robustness criteria, SVD-based image watermarking can still encounter false positive problems (FPP). Avoiding FPPs is one of the popular research topics in the field of SVD-based image watermarking. Satisfying robustness and imperceptibility requirements, as well as preventing FPPs, in SVD-based image watermarking is crucial in applications such as copyright protection and authentication. In this paper, false positive problem is studied, analysed and presented in detail. Different schemes are studied and classified based on the probability of exposure to false positive problem. All types of SVD-based embedding algorithms that leads to false positive problem and the related potential attacks has been evaluated using the reliability test as well as all solutions to false positive problem are reviewed. To understand how the attacks can threaten the rightful ownership and how to avoid these attacks, the three potential attacks of false positive problem has been demonstrated using recent proposed watermarking schemes. The main perspective of this paper is to gather all the issues belong to the false positive problem with SVD-based schemes.

Combined Support Vector Machine and Pattern Matching for Arabic Islamic Hadith Question Classification System

The dimensional phase of Arabic Language, question answering (QA) involves an intrinsic form of question classification (QC) that functions to perform an important task in question answering system (QAS). The purpose of QC is to precisely assign labels to questions that are majorly dependent on the form of answer type. Moreover, classification of user’s question is a herculean task based on the tractability that natural language (NL) affords with different forms. The information enshrined in a group of words is not sufficient to effectively classify the question in the quote. Until now, few reports have focused on QC for Arabic Language question answering (QA). The earlier report has employed the technique of handcrafted rules and keyword matching for QC. Nonetheless, these procedures are considered obsolete in terms of applying it to new territories. In this paper, we present a question-answering system combining aims on a combination of model fixed on Support Vector Machine (SVM) and pattern-based Matching techniques for Arabic Language question classification (ALQC). The Islamic Hadith purview on QA in the study was focusing on the effect of a feature set on the performance of SVM for QC. About five patterns were employed in the analysis together with the classification of three types of questions, namely “Who”, “Where” and “What”. The dataset employed in this study consisted of 200 questions on Arabic Islamic Hadith derived from Sahih Al-Bukhari. The performance generated for the F-measure values for “Who”, “Where” and “What” were 88.39%, 87.66% and 87.93% respectively. In this research work, we evaluate the metric performance to combine the SVM with Pattern Matching to get the accuracy answer. The outcome of this answer reflected that the proposed prototype of SVM and pattern-based approach is indispensible from the field of QC in the Arabic language.

One Parameter at a time Combinatorial Testing Strategy Based on Harmony Search Algorithm OPAT-HS

Software testing is required to verify and validate systems. Combinatorial testing in one of the significant testing techniques. Design and select test cases for combinatorial testing considered as combinatorial problem. Even though there are some existing optimization algorithm based combinatorial testing strategies, that minimize the number of test cases, but most of these strategies based on one-test-at-a-time (OTAT) approach and none of them has adopted the one-parameter-at-a-time (OPAT) approach. Therefore, this paper will propose a new OPAT strategy based on Harmony Search Algorithm (HS) called OPAT-HS. OPAT-HS was originally designed only to support Covering Array (CA) and Mixed Covering Array (MCA) for uniform interaction strength. The result obtained in the experiments appears that OPAT-HS is always best at configurations with MCA notations.