Super-resolution: a comprehensive survey

Abstract

Super-resolution, the process of obtaining one or more high-resolution images from one or more low-resolution observations, has been a very attractive research topic over the last two decades. It has found practical applications in many real world problems in different fields, from satellite and aerial imaging to medical image processing, to facial image analysis, text image analysis, sign and number plates reading, and biometrics recognition, to name a few. This has resulted in many research papers, each developing a new superresolution algorithm for a specific purpose. The current comprehensive survey provides an overview of most of these published works by grouping them in a broad taxonomy. For each of the groups in the taxonomy, the basic concepts of the algorithms are first explained and then the paths through which each of these groups have evolved are given in detail, by mentioning the contributions of different authors to the basic concepts of each group. Furthermore, common issues in super-resolution algorithms, such as imaging models and registration algorithms, optimization of the cost functions employed, dealing with color information, improvement factors, assessment of super-resolution algorithms, and the most commonly employed databases are discussed.

Super-resolution pipeline for fast adjudication in watchlist screening

“…They address the aliasing artifacts that are present in lowresolution images (due to under-sampling process) by simulating the image formation model. In contrast, most singleimage SR algorithms rely on the prior learning-optimization of a set of training images to estimate the mapping function between low-and high-resolution images [17]. In the spatial domain, multi-image SR algorithms map from low-to highresolution using two main steps: registration (image alignment and correction) and reconstruction (image fusion) based on the estimation of the degradation function.…”

“…For single-image case (if any) image is upscaled using some interpolation algorithm. Projection onto Convex Sets (POCS) is employed as an iterative reference reconstruction (fusion) algorithm and is suitable for real time applications [17]. It is based on defined constraints which are convex sets to restrict the space of the super-resolved image.…”

Real-Time Video Scaling Based on Convolution Neural Network Architecture

“…Video Scaling techniques can be partitioned into two parts such as multi-frame and single-frame based approaches [33], [14]. Single image based approach mostly utilizes interpolation or example techniques due to their least computational cost.…”

Multiple-image super-resolution on mobile devices: an image warping approach

“…However, most SR works are performed on the spatial domain and it is here where intuitive techniques are proposed [16]. The concept of Shift-Add fusion is simple.…”

“…The theoretical basis for SR is best explained in the frequency domain [4,16]. However, most SR works are performed on the spatial domain and it is here where intuitive techniques are proposed [16].…”