A Neural Network based Classifier on the Cell Broadband Engine

Open Access
Rajamohan, Srijith
Graduate Program:
Electrical Engineering
Master of Science
Document Type:
Master Thesis
Date of Defense:
August 10, 2009
Committee Members:
  • Suman Datta, Thesis Advisor
  • Vijaykrishnan Narayanan, Thesis Advisor
  • classifier
  • Neural Network
  • Cell
The Cell processor is a novel heterogeneous multicore architecture that holds the distinction of being part of the most powerful distributed computing system capable of sustaining over 4.5 Petaflops as well as in the RoadRunner supercomputer. This thesis hopes to meet the demands of the challenging task of face detection. In this thesis a MLP neural network classifier is implemented on the multicore Cell architecture. Scalability of the code is addressed and every effort has been made to ensure that the code is not dependant on the structure of the network.