Cars Identification from partial images using norm angle of feature points.

Abstract

Image processing is being widely used in various practical fields such as face recognition and medical diagnosis. Image recognition could be based on full shape or partial shape imaging depending on the area where image processing is applied. The application of interest in this paper is to use  partial shape recognition to determine the brand and model of a car. Since the task of tracking partial car images is extremely important for security purposes .partial shape recognition only works on part of a car picture and uses sub matrix matching algorithms. The proposed research aims to improving on existing sub-matrix matching by norm angle calculations of features adding. We will demonstrate that the proposed new approach results in higher detection efficiency and lower error. Previous studies could do this but with the presence of at least 30% of the size of the car to give correct detection with limited efficiency. Our aim is to able to get correct detection results with minimum of 20% of car size coverage.