A.Nag, S. Biswas, D. Sarkar, P.P. Sarkar, B. Gupta
Interest in the area of pattern recognition has been increasing rapidly due to emerging applications, which are not only challenging but also demanding. Feature extraction is a special form of dimensionality reduction in pattern recognition. Our goal is to introduce a simple feature extraction technique for pattern recognition. Pattern recognition is the study of how machines can observe the environment, learn to distinguish patterns of interest from their background and take reasonable decision about the categories of the patterns. During last 50 years of research, attempts have been made by different researchers to develop a general-purpose machine pattern recognizer depending on different algorithms like template matching, statistical approach, artificial neural network etc. In this paper an approach of scale and translation independent feature extraction technique has been presented and analyzed with the help of Hopfield network. This technique is very useful for extraction of feature of shape of an object. It can be applied in the area of image processing, synthetic aperture radar, robotics etc., where detection of shapes of a digitized image or video stream are required.