FSIM: A Feature Similarity Index for Image Quality Assessment
01 Aug 2011-IEEE Transactions on Image Processing (IEEE Trans Image Process)-Vol. 20, Iss: 8, pp 2378-2386
TL;DR: A novel feature similarity (FSIM) index for full reference IQA is proposed based on the fact that human visual system (HVS) understands an image mainly according to its low-level features.
Abstract: Image quality assessment (IQA) aims to use computational models to measure the image quality consistently with subjective evaluations. The well-known structural similarity index brings IQA from pixel- to structure-based stage. In this paper, a novel feature similarity (FSIM) index for full reference IQA is proposed based on the fact that human visual system (HVS) understands an image mainly according to its low-level features. Specifically, the phase congruency (PC), which is a dimensionless measure of the significance of a local structure, is used as the primary feature in FSIM. Considering that PC is contrast invariant while the contrast information does affect HVS' perception of image quality, the image gradient magnitude (GM) is employed as the secondary feature in FSIM. PC and GM play complementary roles in characterizing the image local quality. After obtaining the local quality map, we use PC again as a weighting function to derive a single quality score. Extensive experiments performed on six benchmark IQA databases demonstrate that FSIM can achieve much higher consistency with the subjective evaluations than state-of-the-art IQA metrics.
Citations
More filters
08 Oct 2016
TL;DR: In this paper, the authors combine the benefits of both approaches, and propose the use of perceptual loss functions for training feed-forward networks for image style transfer, where a feedforward network is trained to solve the optimization problem proposed by Gatys et al. in real-time.
Abstract: We consider image transformation problems, where an input image is transformed into an output image. Recent methods for such problems typically train feed-forward convolutional neural networks using a per-pixel loss between the output and ground-truth images. Parallel work has shown that high-quality images can be generated by defining and optimizing perceptual loss functions based on high-level features extracted from pretrained networks. We combine the benefits of both approaches, and propose the use of perceptual loss functions for training feed-forward networks for image transformation tasks. We show results on image style transfer, where a feed-forward network is trained to solve the optimization problem proposed by Gatys et al. in real-time. Compared to the optimization-based method, our network gives similar qualitative results but is three orders of magnitude faster. We also experiment with single-image super-resolution, where replacing a per-pixel loss with a perceptual loss gives visually pleasing results.
6,639 citations
Posted Content•
TL;DR: This work considers image transformation problems, and proposes the use of perceptual loss functions for training feed-forward networks for image transformation tasks, and shows results on image style transfer, where aFeed-forward network is trained to solve the optimization problem proposed by Gatys et al. in real-time.
Abstract: We consider image transformation problems, where an input image is transformed into an output image. Recent methods for such problems typically train feed-forward convolutional neural networks using a \emph{per-pixel} loss between the output and ground-truth images. Parallel work has shown that high-quality images can be generated by defining and optimizing \emph{perceptual} loss functions based on high-level features extracted from pretrained networks. We combine the benefits of both approaches, and propose the use of perceptual loss functions for training feed-forward networks for image transformation tasks. We show results on image style transfer, where a feed-forward network is trained to solve the optimization problem proposed by Gatys et al in real-time. Compared to the optimization-based method, our network gives similar qualitative results but is three orders of magnitude faster. We also experiment with single-image super-resolution, where replacing a per-pixel loss with a perceptual loss gives visually pleasing results.
5,668 citations
Cites background from "FSIM: A Feature Similarity Index fo..."
...More results (including FSIM [63] and VIF [64] metrics) are shown in the supplementary material....
[...]
...More results (including FSIM [63] and VIF [64]) are shown in the supplementary material....
[...]
Posted Content•
TL;DR: A new dataset of human perceptual similarity judgments is introduced and it is found that deep features outperform all previous metrics by large margins on this dataset, and suggests that perceptual similarity is an emergent property shared across deep visual representations.
Abstract: While it is nearly effortless for humans to quickly assess the perceptual similarity between two images, the underlying processes are thought to be quite complex. Despite this, the most widely used perceptual metrics today, such as PSNR and SSIM, are simple, shallow functions, and fail to account for many nuances of human perception. Recently, the deep learning community has found that features of the VGG network trained on ImageNet classification has been remarkably useful as a training loss for image synthesis. But how perceptual are these so-called "perceptual losses"? What elements are critical for their success? To answer these questions, we introduce a new dataset of human perceptual similarity judgments. We systematically evaluate deep features across different architectures and tasks and compare them with classic metrics. We find that deep features outperform all previous metrics by large margins on our dataset. More surprisingly, this result is not restricted to ImageNet-trained VGG features, but holds across different deep architectures and levels of supervision (supervised, self-supervised, or even unsupervised). Our results suggest that perceptual similarity is an emergent property shared across deep visual representations.
3,838 citations
Cites methods from "FSIM: A Feature Similarity Index fo..."
...In fact, they correspond far better than the commonly used metrics like SSIM and FSIM [57, 61], which were not designed to handle situations where spatial ambiguities are a factor [48]....
[...]
...This problem has been a longstanding goal, and there have been numerous perceptually motivated distance metrics proposed, such as SSIM [57], MSSIM [59], FSIM [61], and HDR-VDP [34]....
[...]
...They all perform better than traditional metrics 2, SSIM, and FSIM at 63.2%, 63.1%, 63.8%, respectively....
[...]
11 Jan 2018
TL;DR: In this paper, the authors introduce a new dataset of human perceptual similarity judgments, and systematically evaluate deep features across different architectures and tasks and compare them with classic metrics, finding that deep features outperform all previous metrics by large margins on their dataset.
Abstract: While it is nearly effortless for humans to quickly assess the perceptual similarity between two images, the underlying processes are thought to be quite complex. Despite this, the most widely used perceptual metrics today, such as PSNR and SSIM, are simple, shallow functions, and fail to account for many nuances of human perception. Recently, the deep learning community has found that features of the VGG network trained on ImageNet classification has been remarkably useful as a training loss for image synthesis. But how perceptual are these so-called "perceptual losses"? What elements are critical for their success? To answer these questions, we introduce a new dataset of human perceptual similarity judgments. We systematically evaluate deep features across different architectures and tasks and compare them with classic metrics. We find that deep features outperform all previous metrics by large margins on our dataset. More surprisingly, this result is not restricted to ImageNet-trained VGG features, but holds across different deep architectures and levels of supervision (supervised, self-supervised, or even unsupervised). Our results suggest that perceptual similarity is an emergent property shared across deep visual representations.
3,322 citations
21 Jul 2017
TL;DR: It is concluded that the NTIRE 2017 challenge pushes the state-of-the-art in single-image super-resolution, reaching the best results to date on the popular Set5, Set14, B100, Urban100 datasets and on the authors' newly proposed DIV2K.
Abstract: This paper introduces a novel large dataset for example-based single image super-resolution and studies the state-of-the-art as emerged from the NTIRE 2017 challenge. The challenge is the first challenge of its kind, with 6 competitions, hundreds of participants and tens of proposed solutions. Our newly collected DIVerse 2K resolution image dataset (DIV2K) was employed by the challenge. In our study we compare the solutions from the challenge to a set of representative methods from the literature and evaluate them using diverse measures on our proposed DIV2K dataset. Moreover, we conduct a number of experiments and draw conclusions on several topics of interest. We conclude that the NTIRE 2017 challenge pushes the state-of-the-art in single-image super-resolution, reaching the best results to date on the popular Set5, Set14, B100, Urban100 datasets and on our newly proposed DIV2K.
2,388 citations
Cites background from "FSIM: A Feature Similarity Index fo..."
...There is a large interest in the automatic assessment of the image quality and numerous measures have been proposed [41, 57, 34, 29]....
[...]
References
More filters
TL;DR: In this article, a structural similarity index is proposed for image quality assessment based on the degradation of structural information, which can be applied to both subjective ratings and objective methods on a database of images compressed with JPEG and JPEG2000.
Abstract: Objective methods for assessing perceptual image quality traditionally attempted to quantify the visibility of errors (differences) between a distorted image and a reference image using a variety of known properties of the human visual system. Under the assumption that human visual perception is highly adapted for extracting structural information from a scene, we introduce an alternative complementary framework for quality assessment based on the degradation of structural information. As a specific example of this concept, we develop a structural similarity index and demonstrate its promise through a set of intuitive examples, as well as comparison to both subjective ratings and state-of-the-art objective methods on a database of images compressed with JPEG and JPEG2000. A MATLAB implementation of the proposed algorithm is available online at http://www.cns.nyu.edu//spl sim/lcv/ssim/.
40,609 citations
TL;DR: The theory of edge detection explains several basic psychophysical findings, and the operation of forming oriented zero-crossing segments from the output of centre-surround ∇2G filters acting on the image forms the basis for a physiological model of simple cells.
Abstract: A theory of edge detection is presented. The analysis proceeds in two parts. (1) Intensity changes, which occur in a natural image over a wide range of scales, are detected separately at different scales. An appropriate filter for this purpose at a given scale is found to be the second derivative of a Gaussian, and it is shown that, provided some simple conditions are satisfied, these primary filters need not be orientation-dependent. Thus, intensity changes at a given scale are best detected by finding the zero values of delta 2G(x,y)*I(x,y) for image I, where G(x,y) is a two-dimensional Gaussian distribution and delta 2 is the Laplacian. The intensity changes thus discovered in each of the channels are then represented by oriented primitives called zero-crossing segments, and evidence is given that this representation is complete. (2) Intensity changes in images arise from surface discontinuities or from reflectance or illumination boundaries, and these all have the property that they are spatially. Because of this, the zero-crossing segments from the different channels are not independent, and rules are deduced for combining them into a description of the image. This description is called the raw primal sketch. The theory explains several basic psychophysical findings, and the operation of forming oriented zero-crossing segments from the output of centre-surround delta 2G filters acting on the image forms the basis for a physiological model of simple cells (see Marr & Ullman 1979).
6,893 citations
"FSIM: A Feature Similarity Index fo..." refers background in this paper
...The visual information in an image, however, is often very redundant, while the HVS understands an image mainly based on its low-level features, such as edges and zero-crossings [8-10]....
[...]
09 Nov 2003
TL;DR: This paper proposes a multiscale structural similarity method, which supplies more flexibility than previous single-scale methods in incorporating the variations of viewing conditions, and develops an image synthesis method to calibrate the parameters that define the relative importance of different scales.
Abstract: The structural similarity image quality paradigm is based on the assumption that the human visual system is highly adapted for extracting structural information from the scene, and therefore a measure of structural similarity can provide a good approximation to perceived image quality. This paper proposes a multiscale structural similarity method, which supplies more flexibility than previous single-scale methods in incorporating the variations of viewing conditions. We develop an image synthesis method to calibrate the parameters that define the relative importance of different scales. Experimental comparisons demonstrate the effectiveness of the proposed method.
4,333 citations
TL;DR: An image information measure is proposed that quantifies the information that is present in the reference image and how much of this reference information can be extracted from the distorted image and combined these two quantities form a visual information fidelity measure for image QA.
Abstract: Measurement of visual quality is of fundamental importance to numerous image and video processing applications. The goal of quality assessment (QA) research is to design algorithms that can automatically assess the quality of images or videos in a perceptually consistent manner. Image QA algorithms generally interpret image quality as fidelity or similarity with a "reference" or "perfect" image in some perceptual space. Such "full-reference" QA methods attempt to achieve consistency in quality prediction by modeling salient physiological and psychovisual features of the human visual system (HVS), or by signal fidelity measures. In this paper, we approach the image QA problem as an information fidelity problem. Specifically, we propose to quantify the loss of image information to the distortion process and explore the relationship between image information and visual quality. QA systems are invariably involved with judging the visual quality of "natural" images and videos that are meant for "human consumption." Researchers have developed sophisticated models to capture the statistics of such natural signals. Using these models, we previously presented an information fidelity criterion for image QA that related image quality with the amount of information shared between a reference and a distorted image. In this paper, we propose an image information measure that quantifies the information that is present in the reference image and how much of this reference information can be extracted from the distorted image. Combining these two quantities, we propose a visual information fidelity measure for image QA. We validate the performance of our algorithm with an extensive subjective study involving 779 images and show that our method outperforms recent state-of-the-art image QA algorithms by a sizeable margin in our simulations. The code and the data from the subjective study are available at the LIVE website.
3,146 citations