A Methodology and a System for Adaptive, Integrated Speech and Image Learning and Recognition.
Akbar Ghobakhlou, Nikola Kirilov Kasabov · The International Journal of Computers, Systems and Signal · 2004
ABSTRACT The paper presents a novel approach towards building adaptive speech processing systems based on the evolving connectionist systems paradigm (ECoS). The methodology proposed in this paper is applied on speech and image recognition and person identification based on speech and image. Adaptive connectionist classifier (ACC) is an implementation of the ECoS paradigm. The process of learning and recognition in ACC is achieved through a local adaptation of each module in interaction with the other modules. They can accommodate new input data and new classes through local element tuning. New connections and neurons are created during the adaptive learning process of the system. Experiments are conducted to illustrate this concept. It is demonstrated that a system can adapt to new data and add new outputs at any time of its operation without having to build the network from “scratch”. The system is robust to forgetting when new output classes are added. The methodology is tested on several case studies and results are reported.