Image Resolution Enhancement Using NSCT Based Learning With LBP as Feature Model

In this paper, we propose a new technique for feature preserving spatial resolution enhancement of an image captured at low spatial resolution. We use a training database containing low resolution (LR) images and their high resolution (HR) versions. In an image, different features like edges, corners, curves and junctions are important to convey its local geometry. We use Local Binary Pattern (LBP) operator to represent texture of an image. The missing high resolution details of the low resolution observation are learnt in form of Non-subsampled Contourlet Transform (NSCT) coefficients of the high resolution images in the training database. We demonstrate the effectiveness of the proposed technique by conducting experiments on real world gray scale images. The results are compared with existing learning-based approaches. The proposed technique can be used in applications such as medical imaging, remote surveillance, wildlife sensor networks where the transmission bandwidth, the camera cost and the memory are main constraints.