Face Recognition using Local Ternary Pattern for Low Resolution Image

Face Recognition system is one of an effective system which is used for the identification of a person from a video frame or a digital image. The concept of FRS is primarily used in the field of security systems and hence can be get compared to other systems which are biometric in nature such as IRIS Recognition or Fingerprint systems. Apart from these techniques, Local Binary Patterns (LBP) also plays a vital role for the image recognition as well as texture recognition purposes based on Texture Spectrum Model. But, it can only develop the feature matrix from 0 and 1 on the basis of centre pixel value only. Hence, to compute upper and lower binary codes of regions, the concept of Local Ternary patterns (LTP) is being implemented in this paper. LTP is an extension to LBP which used a threshold value for selection of upper and lower binary code. In this paper, the simulations have been carried out to achieve a better improvement in the blurred images by using Discrete Wave Transformation (DWT) and LTP techniques. DWT and LTP approach for face recognition provides better FRR, FAR and Accuracy than other approaches used for low resolution images.