Face recognition of blurred images using image enhancement and texture features
Deeksha Kapil, Abhilasha · 2015
Face recognition is an existing and one of most prominent technique of biometrics that includes processing of an image and to be matched with different database. In this procedure, the user matches one person identity with several database images. Various approaches like Discrete Wavelet Transform (DWT), Discrete Cosine Transform (DCT), Local Ternary Pattern (LTP) have been used for the purpose of face recognition. These approaches use different features for face recognition purpose. The features used for face recognition are shape, distance between two traits of face and texture features. Texture features are particularly susceptible to the resolution of images, when the resolution changes the calculated textures are not accurate. Texture features computed for low resolution images does not provide better feature information. So there is a big issue in face recognition for low resolution images. In the proposed work, EULBP (Equalized Uniform Local binary Pattern) has been implemented for the purpose of low resolution images, but it does not provide better results up to an extent. To improve the recognition accuracy, the aim of this research is to study various approaches and development of new approach used for recognition purpose on low resolution images using texture features of an image which can provide better results.