Abstract

CHANG TORTOLERO, Oscar Guillermo. Detection of complex objects by using neural networks and ocular microtremor. Rev. Fac. Ing. UCV [online]. 2013, vol.28, n.4, pp.49-55. ISSN 0798-4065.

We present a tracking-recognition system in which two cascaded, independently trained neural networks, cooperate in order to process images and produce the real time reliable recognition of an object that moves in the real world and is visualized through a web cam. The first net specializes in tracking one specific object and once trained it participates in a closed loop control system in which received images directly control eye movements. This arrangement artificially reproduces a phenomenon similar to the ocular micro tremor (OMT) characteristic in mammals’ eyes. The micro tremor signals are stored in short term memory elements and become the input to a second net which converges into a single “concept cell”, whose activity determines the presence of the selected object. The method has been tested using real time real world images under rough visual conditions which include complex background, complex objects and variations in scaling, tilt and perspective. Thanks to its acute artificial vision, our proposed system could be used in relevant social activities such as: Medicine (automatic recognition of tumors and fractures, help to vision impaired people), Agriculture (crop health), Ecology (fire detection and unusual environment behavior), construction (building progress measurement), security (intruder detection), military (moving target tracking and detection), among others.

Keywords : Robotic vision; Neural nets; Complex pattern recognition; Ocular micro tremor; Concept cell.

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )