Learning-based Super-resolution via Canonical Correlation Analysis
نویسندگان
چکیده
The task of image super-resolution is to up sample a low resolution (LR) image while recovering sharp edges and high frequency details. In this paper, a single image superresolution algorithm via canonical correlation analysis (CCA) is proposed. This method is based on the assumption that the corresponding LR and high resolution (HR) images have high correlation coefficients when transformed into a special space. The proposed approach includes two stages: training and testing. In the training stage, a couple of canonical bases for transformation are calculated with the prepared coupled training sets. In the testing stage, the HR image can be recovered by using the canonical bases obtained in the training stage. In addition, an iterative back projection algorithm is used to further improve the image quality. The experiments demonstrate that our algorithm can reconstruct richer details, with fewer artifacts. Moreover, this algorithm is of less complexity.
منابع مشابه
Super-resolution for Face Recognition Based on Correlated Features and Nonlinear Mappings
For the problem of low recognition rate on low resolution face images, a super-resolution method for face recognition based on correlated features and nonlinear mappings is proposed in this paper. Canonical correlation analysis (CCA) is applied to establish the correlated subspaces between the features of high and low resolution face images, and radial base functions (RBFs) are employed to cons...
متن کاملFace Recognition Using Cca on Nonlinear Features
The face recognition (FR) system plays a vital role in commercial & law enforcement applications. Image resolution is an important factor affecting face recognition performance. The performance of face recognition system degrades by low resolution of face images. To address this problem, a super resolution (SR) method was introduced by Hua Huang and Huiting He [7], which uses Canonical correlat...
متن کاملImproving Super-resolution Techniques via Employing Blurriness Information of the Image
Super-resolution (SR) is a technique that produces a high resolution (HR) image via employing a number of low resolution (LR) images from the same scene. One of the degradations that attenuates performance of the SR is the blurriness of the input LR images. In many previous works in the SR, the blurriness of the LR images is assumed to be due to the integral effect of the image sensor of the im...
متن کاملSurveillance Face Super-Resolution via Shape Clustering and Subspace Learning
In a learning-based super-resolution algorithm, suitable prior from the training database is a key issue. A novel face hallucination algorithm based on shape clustering and subspace learning for adaptive prior is proposed in this paper. We define face shape metrics with point distribution model by Hausdorff Distance, then a framework of adaptive prior and subspace learning is proposed to enhanc...
متن کاملImage Super-Resolution based on the Dual-Dictionary Learning and Sparse Representation Method
Learning based image super-resolution method is used to reconstruct high-frequency (HF) details from the prior model trained by a set of high-(HR) and low-resolution (LR) image patches. In this paper, newly proposed novel image superresolution method via dual-dictionary learning and sparse representation, which consists of the main dictionary learning and the residual dictionary learning, to re...
متن کامل