Force Field Feature Extraction Using Fast Algorithm for Face Recognition Performance

Face recognition is method of recognizing individuals by facial expressions. It has become essential for security and surveillance applications, including banks, organisations, workplaces, and social areas, and is needed everywhere. In face recognition, there are a variety of difficulties faced, including face shape, age, sex, lighting, and other variable factors. Another problem is that the scale of the servers for these apps is relatively limited. Education and acknowledgment, thus, are increasingly complicated. In recent years, many unchanged features have been proposed in the literature, in this paper approach the use of the fast algorithm as local descriptors, and as we shall see, it is not only fixed-size features, but also offers the advantage of being highly efficient. The proposed approach allows distinguishing the destination after converting the image to the HSV system, after which the force field features will be extracted using the fast algorithm and then classification by using the distance for three methods (Manhattan, Euclidean, and Cosine) through which a comparison is made to choose the best resolution, as it was found that the resulting accuracy of the two dataset (ORL and UFI) is 99.9%.