Image quality assessment using a vector quantization histogram

Image quality assessment (IQA) evaluates the quality of an image by computing the difference between the reference and distorted images. There are three categories of IQA methods: full-reference, reduced-reference (RR), and no-reference. This paper proposes a vector quantization (VQ) histogram method, which is an RR IQA method. A histogram is generated by counting the number of vectors in each quantized region, which is obtained by VQ processing of an image. This histogram is used as an effective RR feature for IQA. To show the effectiveness of the proposed IQA metric, we compare the results with differential mean opinion score data for laboratory for image and video engineering (LIVE) data images. Experiments with LIVE data images for various types of test images show that the proposed metric gives better performance than the conventional methods such as the structural similarity, mean squared error, and singular value decomposition.