HVS Based Face Recognition Using Slant Transform

There are many person identification technologies like password, PIN, key, and token are used in many applications. Present, popular identification technology is face. In many applications the database is large. Hence, recognition with high speed is major challenge. This paper presents a recognition using HVS features in transform domain. Human visual system identifies perceptual important information in the images. Slant transform basis vector is sawtooth. It efficiently represents linear brightness variations along an image line. Hence, in this work HVS features in slant transform domain is explored for face recognition. Feature vector is based on HVS parameters. In this method image is decomposed into subblocks using Slant Transform. Important elements are identified using HVS weightage. Experiments are performed on bench mark face databases. Proposed method has better recognition performance than existing methods. Retrieval time is also less.

CBIR Using Slant Transform Using DC & AC Coefficients

In the CBIR- (Content Based Image Retrieval) technique requires low-level or primitive features- color, texture, or other data that can be taken from its image Extracting feature vectors of database images as well as query image can be calculated with the help of slant transform by considering DC & 3 AC coefficients obtained in each block of an image. Slant transform represents the gradual brightness changes in an image line effectively. By calculating the difference between feature vector data base and feature vector for a query by using the distance measuring techniques. The vector of the smaller distance is the closest to query image. The experiment is performed in the Corel 500 Image Database. Finally, CBIR results are evaluated by the recall, precision, and F-Score.