No-Reference Sharpness Assessment of Camera-Shaken Images by Analysis of Spectral Structure

Abstract

The tremendous explosion of image-, video-, and audio-enabled mobile devices, such as tablets and smart-phones in recent years, has led to an associated dramatic increase in the volume of captured and distributed multimedia content. In particular, the number of digital photographs being captured annually is approaching 100 billion in just the U.S. These pictures are increasingly being acquired by inexperienced, casual users under highly diverse conditions leading to a plethora of distortions, including blur induced by camera shake. In order to be able to automatically detect, correct, or cull images impaired by shake-induced blur, it is necessary to develop distortion models specific to and suitable for assessing the sharpness of camera-shaken images. Toward this goal, we have developed a no-reference framework for automatically predicting the perceptual quality of camera-shaken images based on their spectral statistics. Two kinds of features are defined that capture blur induced by camera shake. One is a directional feature, which measures the variation of the image spectrum across orientations. The second feature captures the shape, area, and orientation of the spectral contours of camera shaken images. We demonstrate the performance of an algorithm derived from these features on new and existing databases of images distorted by camera shake.

Blind Sharpness Prediction Based on Image-Based Motion Blur Analysis

“…Let B( f z φ , fˆz φ ) be the DFT coefficients at frequencies ( f z φ , fˆz φ ), and let B(0, 0) be the DC component. The contrast Con z φ and variance Var z φ of the DFT coefficients are defined on the z φ -axis, respectively, as [19] …”

“…From (19), it can be seen that the size of the motion blur window W m relies on L w . If L w is too small, we can observe only the DFT distribution on a straight line without including the spatial distribution of DFT coefficients over the spatial domain.…”

“…Fig. 9(a)-(d) show the estimated motion directions by using the simple gradient-based methods [32] and [33], the radon transform-based method [23] and the motion blur window-based method of (19). Fig.…”

“…Various techniques have been studied to measure the sharpness of videos, and are categorized into two types: (1) edge-based methods [14]- [16], which measure the spread of edges under the assumption that there is at least one edge in the frame, and (2) spectral-based methods [17], [18], which measure the spectral distribution under the assumption that the edge and texture regions include high frequency components. However, most works measured the focal blur or out-of-focus blur, and the high frequency components are removed due to the video compression when the camera is stationary, such as in Stationary camera scene [19]. Therefore, it is difficult to find solid work for the sharpness assessment due to the motion blur, particularly in non-stationary scenes or for moving objects.…”

No-reference image quality assessment using interval type 2 fuzzy sets

“…However, determination of the effective parameter set is an important issue. No-reference IQA algorithms [60][61][62][63][64][65][66][67][68][69][70][71][72][73][74][75][76][77] attempt to perform quality evaluation for specific type of distorted images [60][61][62] and designed specifically for JPEG or JPEG2000 compression artifacts [15]. Two prominent works by Wang et al [10] and Sheikh [9] have been reported relating to no-reference image quality assessment.…”