International Journal of Computer Science and Technology

International Journal of Computer Science and Technology | Vol. 7, No. 1, January 2005 | pp. 1–8

DOI: 10.46882/2005/IJCST/070104

Research Article

Title: Face Recognition Under Varying Illumination Using Principal Component Analysis and Local Binary Patterns

Names of Authors: Sofia Alvarez¹, Jean-Pierre Dubois²

Authors’ Affiliations: ¹Departamento de Ingeniería de Sistemas, Universidad Nacional de Colombia, Bogotá, Colombia; ²Laboratoire d'Informatique, Université de Paris-Sud, Orsay, France

Abstract:
Variations in environmental lighting conditions present a persistent barrier to the high-accuracy deployment of automated facial recognition systems. This research describes a hybrid feature extraction model that merges the global dimensionality reduction capabilities of Principal Component Analysis (PCA) with the local structural robustness of Local Binary Patterns (LBP). The pre-processing pipeline employs a modified histogram equalization step to balance intensity gradients before extracting the multi-scale LBP histograms. Testing was performed using the Extended Yale Face Database B, which contains extreme illumination variances. The combined PCA-LBP approach yielded an overall verification accuracy rate of 98.64%, outpacing standalone PCA implementations by 11.3% and standard Linear Discriminant Analysis (LDA) frameworks by 6.7%. The classification phase, driven by a nearest-neighbor classifier using Chi-square distance metrics, completed profiling tasks in approximately t = 14.5 ms per image. This performance baseline makes the algorithm highly suitable for integration into live biometric surveillance workflows.

Keywords: Face Recognition, Principal Component Analysis, Local Binary Patterns, Feature Extraction, Illumination Invariance

Manuscript Timeline: Received October 15, 2004; Revised November 30, 2004; Accepted December 18, 2004; Published January 20, 2005

Citation: Alvarez, S., & Dubois, J. P. (2005). Face Recognition Under Varying Illumination Using Principal Component Analysis and Local Binary Patterns. International Journal of Computer Science and Technology, 7(1), 1–8.

International Journal of Computer Science and Technology | Vol. 6, No. 12, December 2004 | pp. 89–96

DOI: 10.46882/2004/IJCST/061203

Research Article

Title: A Novel Fault-Tolerant Scheduling Architecture for Real-Time Heterogeneous Distributed Systems

Names of Authors: Amara Chinedu¹, Takahiro Tanaka²

Authors’ Affiliations: ¹Department of Computer Science, University of Nigeria, Nsukka, Nigeria; ²Department of Information Science, University of Tokyo, Tokyo, Japan

Abstract:
Ensuring reliability in heterogeneous distributed computing systems without sacrificing deadline constraints is critical for safety-critical real-time applications. This paper introduces a primary-backup fault-tolerant scheduling architecture based on an adaptive passive replication paradigm. The model dynamically balances task execution parameters across heterogeneous processing nodes by monitoring node state vector metrics in real-time. Mathematical modeling shows that the architecture guarantees 100% recovery for single-node failures while maintaining a predictable response envelope. Computational tests under heavy workload profiles (utilization factors U > 0.85) demonstrated a significant reduction in deadline miss ratios by a factor of 3.4 compared to conventional reliable scheduling heuristics. The system processing overhead added by the backup synchronization mechanism was kept minimal at Δ = 3.8 ms ± 0.4 ms. This proves that the strategy is viable for mission-critical deployments like aerospace navigation and automated industrial control systems.

Keywords: Fault Tolerance, Heterogeneous Systems, Distributed Computing, Real-Time Scheduling, Passive Replication

Manuscript Timeline: Received September 01, 2004; Revised October 20, 2004; Accepted November 12, 2004; Published December 28, 2004

Citation: Chinedu, A., & Tanaka, T. (2004). A Novel Fault-Tolerant Scheduling Architecture for Real-Time Heterogeneous Distributed Systems. International Journal of Computer Science and Technology, 6(12), 89–96.

International Journal of Computer Science and Technology | Vol. 6, No. 11, November 2004 | pp. 81–88

DOI: 10.46882/2004/IJCST/061102

Research Article

Title: Dynamic Channel Allocation in Multi-Cell Cellular Systems Using Enhanced Particle Swarm Optimization

Names of Authors: Marcus Vance¹, Elena Rostova²

Authors’ Affiliations: ¹Department of Electrical and Electronic Engineering, University of Manchester, Manchester, UK; ²Faculty of Computer Science, Saint Petersburg State University, Saint Petersburg, Russia

Abstract:
Electromagnetic interference and spectrum scarcity pose substantial challenges to modern cellular network infrastructures. This study introduces an Enhanced Particle Swarm Optimization (EPSO) algorithm designed for dynamic channel allocation (DCA) in multi-cell environments. Unlike traditional heuristic approaches that frequently converge into local minima, the EPSO incorporates a chaotic mutation operator that scales adaptively with the system workload. Simulations were evaluated across a standard 49-cell hexagonal grid topology under varying traffic intensities ranging from 10 to 120 Erlangs. The empirical findings show that the proposed EPSO framework drops the call blocking probability (P_b) to less than 0.015 under peak traffic loads, yielding a 14.2% reduction compared to standard genetic algorithms. The average spectrum utilization efficiency improved by μ = 18.7%, while the computational execution time required to reach convergence stabilized at t = 42 ms. This allows for near real-time adaptations to rapid fluctuations in user density.

Keywords: Channel Allocation, Particle Swarm Optimization, Cellular Networks, Call Blocking Probability, Spectrum Efficiency

Manuscript Timeline: Received August 04, 2004; Revised September 15, 2004; Accepted October 11, 2004; Published November 22, 2004

Citation: Vance, M., & Rostova, E. (2004). Dynamic Channel Allocation in Multi-Cell Cellular Systems Using Enhanced Particle Swarm Optimization. International Journal of Computer Science and Technology, 6(11), 81–88.

International Journal of Computer Science and Technology | Vol. 6, No. 10, October 2004 | pp. 73–80

DOI: 10.46882/2004/IJCST/000063

Article Type: Research Article

Title: An Optimized Metaheuristic Swarm Framework for the Multi-Objective Layout Design of Factory Logistics

Names of Authors: Mateo Silva¹, Lucia Rossi²

Authors’ Affiliations: ¹Department of Industrial Engineering, University of Chile, Santiago, Chile; ²Department of Computer Science, University of Bologna, Italy

Abstract: Designing physical spatial layouts for modern production plants is a complex optimization task that directly affects internal material transport overhead, assembly throughput, and overall operational safety. This paper introduces an optimized multi-objective Ant Colony Optimization (ACO) metaheuristic tailored for structural facility layout problems. The AHM algorithm relies on an adaptive pheromone update mechanism that switches exploration patterns depending on layout congestion indexes to prevent local stagnation. We benchmarked our framework using standard industrial layout instances alongside real automated guided vehicle tracking logs. The proposed ACO model secured an overall 14.8% reduction in material handling costs and minimized idle transit delay parameters by 9.2% when compared to standard Tabu Search strategies. Computationally, the routine compiles optimal spatial configurations in under 3.5 s, establishing an agile system for factory floor reconfiguration.

Keywords: Facility Layout Problem, Ant Colony Optimization, Factory Logistics, Metaheuristics, Space Allocation

Manuscript Timeline: Received: July 20, 2004; Revised: August 30, 2004; Accepted: September 24, 2004; Published: October 18, 2004

Citation: Silva, M., & Rossi, L. (2004). An Optimized Metaheuristic Swarm Framework for the Multi-Objective Layout Design of Factory Logistics. International Journal of Computer Science and Technology, 6(10), 73–80.

International Journal of Computer Science and Technology | Vol. 6, No. 9, September 2004 | pp. 65–72

DOI: 10.46882/2004/IJCST/000062

Article Type: Original Research Paper

Title: An Empirical Investigation into Web Service Composition Using Automated Semantic AI Planners

Names of Authors: Amina Bekele¹, Jean-Claude Mpasa²

Authors’ Affiliations: ¹Department of Information Science, Addis Ababa University, Ethiopia; ²Department of Computer Science, University of Kinshasa, Democratic Republic of the Congo

Abstract: Combining individual business web services into an integrated enterprise workflow remains an expensive, manual engineering process due to a lack of explicit semantic descriptors. This research conducts a comparative empirical evaluation of three automated Artificial Intelligence planning systems—SHOP2, FastForward, and an optimized heuristic search planner—tailored for automated Web Service composition. Utilizing a testbed described in OWL-S with 500 unique transactional service endpoints, we measured planning duration, graph complexity, and success rates under conflicting operational criteria. The heuristic search planner outperformed competing architectures by mapping dependency strings inside a unified semantic abstraction layer, achieving a service composition generation rate of under 180 ms per workflow request. Our analysis demonstrates that utilizing ontological relationships minimizes graph compilation errors by 32.5%, providing a practical blueprint for automated e-commerce and logistics software integration.

Keywords: Web Services, Semantic Web, AI Planning, Ontology, OWL-S, Service-Oriented Architecture

Manuscript Timeline: Received: June 18, 2004; Revised: July 29, 2024; Accepted: August 21, 2004; Published: September 12, 2004

Citation: Bekele, A., & Mpasa, J.-C. (2004). An Empirical Investigation into Web Service Composition Using Automated Semantic AI Planners. International Journal of Computer Science and Technology, 6(9), 65–72.

International Journal of Computer Science and Technology | Vol. 6, No. 8, August 2004 | pp. 57–64

DOI: 10.46882/2004/IJCST/000061

Article Type: Research Article

Title: Design and Testing of a Fault-Tolerant Distributed Shared Memory Protocol for Linux Clusters

Names of Authors: Sophie Dubois¹, Hans-Werner Schmidt²

Authors’ Affiliations: ¹Department of Applied Cryptography, ENS Paris, France; ²Fraunhofer Institute for Secure Information Technology, Darmstadt, Germany

Abstract: Distributed Shared Memory (DSM) models simplify parallel programming pipelines but are prone to complete cluster failure if any individual computing node goes offline. This study outlines a robust, page-based Software DSM protocol that incorporates an asynchronous checkpointing infrastructure to guarantee fault tolerance. The software architecture tracks page modification actions using a relaxed, lazy release consistency model, saving differential page updates to non-volatile local storage blocks. Fault recovery evaluations conducted on a 32-node Linux workstation grid showed that our protocol handles unexpected node terminations cleanly, restoring cluster processing state within 1.45 s of failure detection. The performance penalty introduced by the checkpointing code loop caused a negligible throughput drop of less than 2.8% under typical matrix factorization workloads, offering a secure environment for long-running industrial compute loops.

Keywords: Distributed Shared Memory, Fault Tolerance, Workstation Clusters, Checkpointing, Consistency Models

Manuscript Timeline: Received: May 14, 2004; Revised: June 22, 2004; Accepted: July 19, 2004; Published: August 15, 2004

Citation: Dubois, S., & Schmidt, H.-W. (2004). Design and Testing of a Fault-Tolerant Distributed Shared Memory Protocol for Linux Clusters. International Journal of Computer Science and Technology, 6(8), 57–64.