Venue : Coral Deira - Dubai, Deira, Dubai, UAE..  &  Date : April 24 ~ 25, 2015

Accepted Papers

  • Lossless Color Image Compression With Mojette Transform Using Different Projection A Preliminary Study
    Djamel Eddine Boukhari1 and Mourad Kezai2,1University of Biskra, Algeria and 2University of Setif, Algeria.
    The Mojette transform is a discrete, exact and redundant Radon transform. The application of Mojette transform in image compression is based on the image projection similarity using different directions. The techniques using in this paper of lossless image compression based on Mojette transform called "intra-projection coding", "inter-projections coding" and "differential coding" are proposed in literature. In the latter case, for this paper, we propose the mean coding of projections to improve the Mojette transform in image color with different projections and we calculate the difference between the projections and the mean. The results obtained are better compared to the inter-projection coding in most cases, and comparable with CALIC.
  • Using Global Motion Informatin for Modification of the Adaptive Rood Pattern Search Algorithm
    Mehrnaz Fani, Fatemeh Ghofrani and Mehran Yazdi, Shiraz University, Iran.
    In this paper we present an efficient block based motion estimation algorithm called Modified Adaptive Rood Pattern Search with use of the Global Motion Information (GMI-MARPS). This method can be used in video codecs to reduce the temporal redundancies and to enhance their compression ratio. In the proposed method, first a fast Global Motion Estimation (GME) algorithm based on cross power spectrum is performed. The results show this simple ME procedure is efficient in most of the scenarios. The achieved motion vector (MV) is then assigned to the blocks with the global motion (GMBs). This step can significantly increase the compression ratio and reduce the computational cost of the following stages. Next, the motion of the blocks that cannot be described by the global motion (NGMBs) are found by introducing the MARPS algorithm. This algorithm involves three main steps; 1) Prediction, that utilizes MVs of four spatial neighbors and the GM vector. 2) Initial search, which is based on rood pattern search. 3) Refined search, which deploys the unite-size rood pattern search. The proposed method is then compared with the well-known fast block matching Algorithms (BMAs). Note that, the combination of different BM searching algorithms with the initial GME provides significant improvement in their performances, regarding their computational cost and the MSE of the predicted frames. In empirical simulation we have compared the performance of our method with other BM algorithms after combining them with the initial GMI. Experimental results show the capability of our method and its superiority over the other algorithms that exist in the literature.
  • A Modified Method for Predictivity of Heart Rate Variability
    Mazhar B. Tayel and Eslam I AlSaba,Alexandria University,Egypt.
    Heart Rate Variability (HRV) plays an important role for reporting several cardiological and non-cardiological diseases. Also, the HRV also has a prognostic value and is therefore very important in modeling the cardiac risk. The nature of the HRV is chaotic, stochastic, it remains highly controversial. Because the HRV has utmost importance, it needs a sensitive tool to analyze the variability. In previous work, Rosenstein and Wolf had used the lyapunov exponent as a quantitative measure for HRV detection sensitivity. However, the two methods diverge in determining the HRV sensitivity. This paper introduces a modification to both the Rosenstein and Wolf methods to overcome their drawbacks. The introduced Mazhar-Eslam algorithm increases the sensitivity to HRV detection with better accuracy.