Rowayda A. Sadek

SVD Based Image Processing Applications: State of The Art, Contributions and Research Challenges

Singular Value Decomposition (SVD) has recently emerged as a new paradigm for processing different types of images. SVD is an attractive algebraic transform for image processing applications. The paper proposes an experimental survey for the SVD as an efficient transform in image processing applications. Despite the well-known fact that SVD offers attractive properties in imaging, the exploring of using its properties in various image applications is currently at its infancy. Since the SVD has many attractive properties have not been utilized, this paper contributes in using these generous properties in newly image applications and gives a highly recommendation for more research challenges. In this paper, the SVD properties for images are experimentally presented to be utilized in developing new SVD-based image processing applications. The paper offers survey on the developed SVD based image applications. The paper also proposes some new contributions that were originated from SVD properties analysis in different image processing. The aim of this paper is to provide a better understanding of the SVD in image processing and identify important various applications and open research directions in this increasingly important area; SVD based image processing in the future research.

Blind Synthesis Attack on SVD Based Watermarking Techniques

Singular value decomposition (SVD) is robust and reliable orthogonal matrix decomposition. Due to its conceptual and stability properties, it is becoming more and more popular in signal processing area especially in watermarking. Many watermarking and data hiding researches has struck the use of singular values in casting hidden information into the hosted image. This paper introduces a counterfeiting attack for the existed SVD based watermarking techniques that embed the watermark information into the singular values of the hosted image. This paper emphasizes the weakness of this class of techniques due to the vulnerability of singular values to a wide class of image processing operations as well as intentional attacks.

Multilayer Secured SIP Based VoIP Architecture

VoIP faces many emerging attacks and threats. securing SIP based VoIP is a major challenging task, hence confidentiality, integrity, availability, as well as authenticity must be provided. Focusing on three main critical attacks targeting SIP based VoIP infrastructure, which are Denial of service (DoS), man-in-the middle attack, and Authenticity based attacks. In this paper the main contribution is providing a secure efficient multilayer security architecture based on open source applications (snort, snortsam and iptables, as well as OPENVPN Tunnel), The architecture provides a secure reliable VoIP services for the enterprise network, that have been deployed based on asterisk PBX. The proposed security architecture aims to prevent the mentioned critical attacks, to provide CIAA security services, by proposing an adaptive rule based queuing polices. QoS is a major challenge, the paper also provides an enhancement for the proposed architecture to minimize the delay for more efficient secure communication, as well as preventing zero day attacks by exploiting method and updating Snort DB with attack signatures. QoS factors have been measured using OPNET simulators. The proposed architecture gives promising results when it comes to attacks prevention with 0.01% better performance results compared to previous work.

An efficient brain mass detection with adaptive clustered based fuzzy C-mean and thresholding

Image segmentation plays an important role in analyzing medical images. Brain tumor detection is one of the applications that require brain image segmentation. Due to the complex nature of brain magnetic resonance images (MRI), the accurate computer aided detection (CAD) system for brain tumor segmentation has a lot of advantages over manual segmentation as it requires a lot of time and its results may vary from expert to expert. In this paper, a novel automatic tumor detection technique is proposed using Fuzzy C-means(FCM) where the number of clusters is determined automatically and conformed threshold is used instead of global threshold. Also a preprocessing stage is added to enhance the detection accuracy. The performance of the proposed method is evaluated by calculating completeness, correctness and accuracy of brain magnetic resonance (MR) tumor images. The proposed technique gives promising results for completeness, correctness, accuracy and processing time. Also, a comparison between the proposed technique results and the global threshold technique results is given.

An Efficient Hybrid Detection System for Abnormal Masses in Digital and Analog Mammogram

Mammography is a type of radiography used on the breasts as screening method for women. The indicators for breast cancer aremasses and calcifications. Breast cancer screenings show that radiologists miss 8%–20% of the tumors. For this reason, development of systems for computer-aided detection (CAD) and computer-aided diagnosis (CADx) algorithms is the concern of a lot of researches currently being done. CAD and CADx algorithms assist radiologists in the decision between follow up and biopsy phases. An intelligent Image Processing Technique employed in systems that can help the radiology in detecting abnormal masses. This paper presents a general framework for mammography that will provide advantages for managing information and simplifying process in each layer for imaging technique. A method has been developed to make supporting tools used a framework as a reference model. This method will automatically segment and detect abnormal masses in analog and digital mammography images and compare between results.

Content Aware multi-MetricRouting for MPEG-4 video over IEEE 802.11e wireless mesh networks using cross-layer approach

Video streams over wireless channel require stringent QoS constraints, such as low packet loss and latency. It becomes the main challenge to design a network to achieve these QoS objectives. This paper proposes a cross-layer approach to enhance quality of MPEG-4 video stream transmission over an IEEE 802.11e wireless mesh networks. Most of the conventional routing schemes follow the philosophy of “shortest path”, where hop count is the metric to determine the best route but shortest-path routing schemes repeatedly use congested nodes and links, causing heavy loads, frequent link failures and interface queue overflow. In this paper Cross-layer approach among application layer, network layer and MAC layer has been proposed to discover the best QoS route to forward the highest priority video frames. The proposed modified protocol adds the QoS parameter ETX metric to hop count as a new routing metric to select the best QoS route during the route discovery phase. Proposed approach is compared with already developed schemes. Simulation results using NS2 Simulator toolgive high performance under light and heavy loads network traffic.

Comparative analysis of authentication techniques to Secure Low Level Reader Protocol (LLRP) connection

Radio Frequency Identification (RFID) is a technology that has risen to prominence over the past decade. EPCglobal Network provides a standardized way to collect, store, and distribute data among various organizations. Low-Level Reader Protocol (LLRP) is a standard interface between RFID Readers and Clients. LLRP implements Transport Layer Security (TLS) to setup a secure connection. Authentication techniques in TLS protocol are mainly based on public key cryptography. Recently TLS supports Kerberos cipher suites that provide Kerberos-based authentication. It is a symmetric key cryptography. This paper introduces RFID Access Control system (RFID-AC) as approach and assesses its security vulnerabilities. The paper experimentally implements TLS LLRP endpoints and examines different possible security solutions. Experiments focus on comparing the commonly used authentication technique; authenticated certificate-based versus Kerberos credentials-based. Analysis was done to determine the most suitable one for access control that can satisfy the security concerns and Quality of Services (QoS) issues.

Computer assisted surgery and therapy (CAST) for breast masses based on mammography

Breast cancer represents the most frequently diagnosed cancer in women. Mammography is the most commonly method for early detection of masses related to breast cancer. Correlation of information from multiple-view mammograms improves the performance of diagnosis by radiologists or by computer assisted systems. Detecting the location of masses accurately is highly important for radiologist to classify masses, for the surgeon to help in ease the procedure of an accurate surgery and in radio therapy process for less and efficient dose. In this paper, CAST is developed to accurately locate abnormal masses by quarter and clock segments. CAST will replace the manual localization method. The nipple is used as a reference point in localization process. A new simple nipple detection method is also proposed. The proposed CAST is examined on a new local database which was tested and classified by experts. The methodology achieved a sensitivity of 96% and specificity 73.33%.

A Robust Algorithm for Enhancement of Remotely Sensed Images Based on Wavelet Transform

In the field of remote sensing, removing noise from images is still a challenging research in image processing. Generally there is no common enhancement approach for noise reduction. Several approaches have been introduced and each has its own assumption, advantages and disadvantages. The speckle noise is usually found in the remote sensing images. This paper proposes an adaptive threshold method for image despeckling based on wavelet transform. The quality of the enhanced images in this paper is measured by the statistical quantity measures: Peak Signal-to-Noise Ratio (PSNR), and Mean Square Error (MSE). Experimental results showed that the proposed method demonstrates an improved denoising performance over related techniques according to increasing of PSNR values and decreasing of MSE values of enhanced images.