Automatically learned quality assessment for images has recently become a hottopic due to its usefulness in a wide variety of applications such asevaluating image capture pipelines, storage techniques and sharing media.Despite the subjective nature of this problem, most existing methods onlypredict the mean opinion score provided by datasets such as AVA [1] and TID2013[2]. Our approach differs from others in that we predict the distribution ofhuman opinion scores using a convolutional neural network. Our architecturealso has the advantage of being significantly simpler than other methods withcomparable performance. Our proposed approach relies on the success (andretraining) of proven, state-of-the-art deep object recognition networks. Ourresulting network can be used to not only score images reliably and with highcorrelation to human perception, but also to assist with adaptation andoptimization of photo editing/enhancement algorithms in a photographicpipeline. All this is done without need for a "golden" reference image,consequently allowing for single-image, semantic- and perceptually-aware,no-reference quality assessment.