We propose a simple kernel based nearest neighbor approach for handwritten digit classification. The "distance" here is actually a kernel defining the similarity between two images. We carefully study the effects of different number of neighbors and weight schemes and report the results. With only a few nearest neighbors (or most similar images) to vote, the test set error rate on MNIST databas...

The image quality assessment has been widely used in image processing. Several researches have been developed and carried considering the Human Visual System (HVS). Under the hypothesis that human visual perception is extremely adapted to retrieve structural information from a scene, the SSIM index is the most widely used in this area, which leads to a better correlation with HVS. Despite its r...

The structural similarity index (SSIM) has been shown to be a superior objective image quality metric. A web-based pilot experiment was conducted with the goal of quantifying, through the use of a sample of human participants, a trend in SSIM values showing when the human visual system can begin to perceive distortions applied to reference images. The just noticeable difference paradigm was use...

Numerical experiments indicate that images, in general, possess a considerable degree of affine self-similarity, that is, blocks are well approximated in root mean square error (RMSE) by a number of other blocks when affine greyscale transformations are employed. This has led to a simple L-based model of affine image self-similarity which includes the method of fractal image coding (cross-scale...

Although image quality assessment (IQA) has observed much research and advancement in the past, not much progress has been made towards the development of quality assessment metrics for range images. The current metrics being used to assess the quality of range images are reminiscent to the old IQA metrics such as mean squared error that have since been replaced with metrics which better correl...

In this supplementary file, we first present more quantitative results for experiments that employed the 91-image dataset as the training set. SSIM and IFC [1] are reported. Then we follow the literature to compare with seven stateof-the-art super-resolution algorithms trained with different datasets. All results are evaluated in terms of PSNR, SSIM and IFC on the Set5 [2], Set14 [3] and BSD200...

Starting instead from a symmetric-antisymmetric reformulation we first show SSIM to be a contrast or visibility function in the classic sense. Furthermore, the previously enigmatic structural covariance is revealed to be the difference of variances. The second step, eliminating the intrinsic quadratic nature of SSIM, allows a near linear correlation with human observer scores, and without invok...

The structural similarity (SSIM) index has been shown to be an useful tool in a wide variety of applications that involve the assessment of image quality and similarity. However, in-depth studies are still lacking on how to incorporate it for signal representation and approximation problems, where minimal mean squared error is still the dominant optimization criterion. Here we examine the probl...

A novel image threshold selection approach based on structural similarity (SSIM) is proposed. The thresholded image is obtained first, then comparison regions are extracted based on the local variance of the neighborhood of the thresholded image. Due to the characteristic of comparison regions, the conventional SSIM expression is simplified as a nonparametric form, and the partial SSIM (PSSIM) ...