Browsing by Author "Tijani, I. B."

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    Detection of vascular intersection in retina fundus image using modified cross point number and neural network technique
    (IEEE, 2008-05-13) Iqbal, Muhammad I.; Aibinu, A. M.; Nilsson, Mikael; Tijani, I. B.; Salami, Momoh-Jimoh E.
    Vascular intersection can be used as one of the symptoms for monitoring and diagnosis of diabetic retinopathy from fundus images. In this work we apply the knowledge of digital image processing, fuzzy logic and neural network technique to detect bifurcation and vein-artery cross-over points in fundus images. The acquired images undergo preprocessing stage for illumination equalization and noise removal. Segmentation stage clusters the image into two distinct classes by the use of fuzzy c-means technique, neural network technique and modified cross-point number (MCN) methods were employed for the detection of bifurcation and cross-over points. MCN uses a 5x5 window with 16 neighboring pixels for efficient detection of bifurcation and cross over points in fundus images. Result obtained from applying this hybrid method on both real and simulated vascular points shows that this method perform better than the existing simple cross-point number (SCN) method, thus an improvement to the vascular point detection and a good tool in the monitoring and diagnosis of diabetic retinopathy.
  • Thumbnail Image
    Item
    Hybrid algorithm for NARX network parameters' determination using differential evolution and genetic algorithm
    (IOP Publishing, 2013) Salami, Momoh-JImoh E.; Tijani, I. B.; Abdullateef, A. I.; Aibinu, M. A.
    A hybrid optimization algorithm using Differential Evolution (DE) and Genetic Algorithm (GA) is proposed in this study to address the problem of network parameters determination associated with the Nonlinear Autoregressive with eXogenous inputs Network (NARX-network). The proposed algorithm involves a two level optimization scheme to search for both optimal network architecture and weights. The DE at the upper level is formulated as combinatorial optimization to search for the network architecture while the associated network weights that minimize the prediction error is provided by the GA at the lower level. The performance of the algorithm is evaluated on identification of a laboratory rotary motion system. The system identification results show the effectiveness of the proposed algorithm for nonparametric model development.