Abstract
This paper presents an image recognition scheme using Hopfield net and ellipse fitting procedure. The method is based on geometrical and visual analysis of the image features. The system is commenced with capturing image using a digital camera. On image enhancement, processing and analysis, images of elliptical shape are being approximated as 'ellipse' using principal component analysis. Hebbian learning has been employed for ellipse fitting and recognition has been established by means of Hopfield Neural Network. Experimental results demonstrate that our method is capable of recognizing elliptical image patterns efficiently.