A no-reference objective image sharpness metric based on a filter bank of Gaussian derivative wavelets

The quality of an image depends on various attributes such as sharpness (or blurriness), naturalness, colorfulness, and contrast, etc. To develop a so-called no-reference objective image quality metric by incorporating all attributes of images without referring to the original ones is a difficult task. Hence, the purpose of this paper focuses on the development of a no-reference objective image sharpness metric. A high-frequency component preserving the full information of a given input image is extracted first and then applied to a multi-channel filter bank. The proposed method measures image sharpness based on the output of this filter bank. A snug frame composed of the Gaussian derivative wavelets is applied to construct the filter bank. The output of the filter bank not only contains the complete information of the input but also manifests prominent image features. Experimental results show that the metric predicts well in isotropic and uniform blur case. Validated by the high correlation between the metric values and the subjective test scores, the performance of the proposed metric is comparable to that of human subjects.