Zhaohui Che

No-reference image quality assessment for photographic images of consumer device

In this paper we study common, camera-specific kinds of distortions and propose a no-reference image quality assessment algorithm for photographic images produced by consumer devices. Those real consumer-type images, being different from simulated-distortion images, are with realistic artifacts and quality ranges. We find that the state-of-the-art no-reference image quality assessment approaches do not perform well on those photographic images, and propose an approach that achieves high prediction performance on a dataset of consumer-centric images. The proposed method, with no need for the original image, is able to reveal camera-specific problems and differentiate consumer cameras.

Stereoscopic image quality assessment with the dual-weight model

Three-dimensional (3D) imaging technology has been growingly prevalent in todays world. But objective quality assessment of 3D images is a challenging task. In this paper, we try to investigate and develop the 2D image quality metrics into stereoscopic image quality assessment using a dual-weight model within the concept of free energy principle and human visual system. On the basis of a psychological measure, the free energy is a principle telling where supervises more and attracts human attention. We believe that the “surprise” can account for the binocular rivalry and introduce the inter-weight. Also, the intra-weight is computed according to the physiology of human visual system. We first do the intra-weighted 2D image quality assessment. With inter-weights, we then calculate the combining score of stereopairs. Experiments are performed on the symmetric LIVE3D-I and asymmetric LIVE3D-II databases. Results confirm that the proposed 3D IQA technique, appropriately computing and combining the inter- and intra-weights, is able to faithfully predict the visual quality of stereopairs.

Perceptually meaningful quadtree decomposition using dual-homogeneous criteria

Saliency detection is a powerful tool for many applications because it can provide valuable information reflecting human visual attention. Quadtree decomposition is a classical and efficient image operator which can divide image into square blocks in different scales. The blocks in smaller size contain valuable information such as edges and texture. In image lossy compression or similar applications, we want to reserve valuable regions (salient objects) as much as we can, meanwhile, we also want to suppress non-salient regions human are not interested in. For these reasons, we propose an ingenious saliency guided quadtree decomposition model. The proposed model outperforms traditional methods in two aspects: It reserves the salient regions well, and the edges can be protected simultaneously.

Influence of Spatial Resolution on State-of-the-Art Saliency Models

Visual attention has been widely investigated and applied in recent decades. Various computation models have been proposed to modeling visual attention, but most researches are conducted under the assumption that given images have few limited spatial resolutions. Spatial resolution is an important feature of image. Image resolution may have some influence on visual attention, and it may also affect the effectiveness of visual attention models. The influence of spatial resolution on saliency models has not been systematically investigated before. In this paper, we discuss two problems related to image resolution and saliency: 1 Most saliency models contain down-sampling function which changes the resolution of original images to lower the computational complexity and keep the formalization of the algorithm. In the first part, we discuss the influence of the down-sampling ratio on the effectiveness of classic saliency models. 2 In the second part, we investigate the effectiveness of saliency models in images of various resolutions. A dataset which provides images and corresponding eye movement data in various spatial resolutions is used in this part. We apply the default rescaling parameters and keep them unchanged. Then we analyze the performance of classic models on 8 resolution levels. In summary, we systematically investigate and analyze problems concerning spatial resolution in the research of saliency modeling. The results of this work can provide a guide to the use of classic models in images of different resolutions and they are helpful to computational complexity optimization.

Subjective and objective quality assessment for color changed images

Color change is an important factor in image quality assessment (IQA). Changing color manually is a hard work and improper process may destroy the perceptual quality. This problem has been largely ignored in traditional IQA which mainly focused on evaluating blur, noise and compression loss, etc. One bottleneck in this field is the lack of databases. Existing databases have very few distortion types related with color changes. In this paper, we construct a new database specially designed for color changed IQA problem. We then test a majority of state-of-the-art IQA methods on this database and prove that existing algorithms are indeed not suitable for color changed distortion. Therefore, we build a novel matrix considering features in color theory and statistics. The new model can effectively evaluate the color changed image quality.

Closing the gap: Visual quality assessment considering viewing conditions

Most of existing visual quality assessment algorithms are tested on standard databases that are created in controlled viewing conditions (e.g. display device, viewing distance and lighting). This implies that all the recoded subjective scores are only valid for the specific settings used in the database. However, with the prevalence of mobile devices, the practical viewing environments can significantly vary from moment to moment. It is our daily experience that the same image can look drastically different on dissimilar devices under changed viewing distance and/or lighting conditions. In other words, a gap exists between the eyes and the visual contents behind the screen in current research of quality assessment. Therefore, in this work, we perform subjective quality evaluation with varied actual viewing conditions. To make the research reproducible, we build a prototype system to record what the eyes really see from the screen and construct the viewing environment-changed image database. The database will be made available to the public. Meanwhile we design a dedicated effective environment-assessing algorithm. We believe that this work will benefit the research of visual quality assessment towards more practical applications.