Virtual reality (VR) experiences often elicit a negative effect, cybersickness, which results in nausea, disorientation, and visual discomfort. To quantitatively analyze the degree of cybersickness depending on various attributes VR content (i.e., camera movement, field view, path length, frame reference, controllability), we generated reference (CYRE) with 52 scenes that represent different attributes. A protocol for evaluation was designed to collect subjective opinions from 154...

Being able to predict the degree of visual discomfort that is felt when viewing stereoscopic 3D (S3D) images an important goal toward ameliorating causative factors, such as excessive horizontal disparity, misalignments or mismatches between left and right views stereo pairs, conflicts different depth cues. Ideally, a model should account for factors capture geometries, distribution disparities, responses neurons. When modern displays, caused primarily by changes in binocular vergence while...

Previously, no-reference (NR) stereoscopic 3D (S3D) image quality assessment (IQA) algorithms have been limited to the extraction of reliable hand-crafted features based on an understanding insufficiently revealed human visual system or natural scene statistics. Furthermore, compared with full-reference (FR) S3D IQA metrics, it is difficult achieve competitive score predictions using extracted features, which are not optimized respect opinion. To cope this limitation conventional approach,...

What if we could interpret the cognitive state of a user while experiencing virtual reality (VR) and estimate from visual stimulus? In this paper, address above question by developing an electroencephalography (EEG) driven VR cybersickness prediction model. The EEG data has been widely utilized to learn representation brain activity. first stage, fully exploit advantages data, it is transformed into multi-channel spectrogram which enables account for correlation spectral temporal...

When coming up with phrases of movement, choreographers all have their habits as they are used to skilled dance genres. Therefore, tend return certain patterns the genres that familiar with. What if artificial intelligence could be help blend by suggesting various dances, and one matches choreographic style? Numerous task-specific variants autoregressive networks been developed for generation. Yet, a serious limitation remains existing algorithms can repeated given initial pose sequence,...

The human visual system perceives 3D depth following sensing via its binocular optical system, a series of massively parallel processing units, and feedback that controls the mechanical dynamics eye movements crystalline lens. process accommodation (focusing lens) vergence is controlled simultaneously symbiotically cross-coupled communication between two critical computation modalities. output responses these subsystems, which are induced by oculomotor control, used in clear stable cyclopean...

In a virtual reality (VR) environment, where visual stimuli predominate over other stimuli, the user experiences cybersickness because balance of body collapses due to <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">self-motion</i> . Accordingly, VR experience is accompanied by unavoidable sickness referred as visually induced motion (VIMS). this article, our primary purpose simultaneously estimate VIMS score referring content and calculate...

Conventional stereoscopic 3D (S3D) displays do not provide accommodation depth cues of the image or video contents being viewed. The sense content depths is thus limited to supplied by motion parallax (for video), vergence created presenting left and right views respective eyes, other contextual perspective cues. absence can induce two kinds mismatches (AVM) at fixation peripheral points, which result in severe visual discomfort. With aim alleviating discomfort arising from AVM, we propose a...

Image inpainting methods leverage the similarity of adjacent pixels to create alternative content. However, as invisible region becomes larger, completed in deeper hole are difficult infer from surrounding pixel signal, which is more prone visual artifacts. To help fill this void, we adopt an progressive hole-filling scheme that hierarchically fills corrupted feature and image spaces. This technique allows us utilize reliable contextual information pixels, even for large samples, then...

Bridging distant space-time interactions is important for high-quality video inpainting with large moving masks. Most existing technologies exploit patch similarities within the frames, or leaverage large-scale training data to fill hole along spatial and temporal dimensions. Recent works introduce promissing Transformer architecture into deep escape from dominanace of nearby achieve superior performance than their baselines. However, such methods still struggle complete larger holes...

To maximize the presence experienced by humans, visual content has evolved to achieve a higher in series of high definition (HD), ultra HD (UHD), 8K UHD, and stereoscopic 3D (S3D). Several studies have introduced delivered from when viewing UHD S3D analysis perspective. Nevertheless, no clear been presented for presence, only subjective evaluation relied upon. The main reason this is that there limitation defining via use information itself. In paper, we define each environment, investigate...

Visual saliency on stereoscopic 3D (S3D) images has been shown to be heavily influenced by image quality. Hence, this dependency is an important factor in quality prediction, restoration and discomfort reduction, but it still very difficult predict such a nonlinear relation images. In addition, most algorithms specialized detecting visual pristine may unsurprisingly fail when facing distorted paper, we investigate deep learning scheme named Deep Saliency (DeepVS) achieve more accurate...

Single-image 3-D reconstruction has long been a challenging problem. Recent deep learning approaches have introduced to this area, but the ability generate point clouds still remains limited due inefficient and expensive representations, dependency between output number of model parameters, or lack suitable computing operation. In article, we present novel deep-learning-based method reconstruct cloud an object from single image. The proposed can be decomposed into two steps: feature fusion...

Most prior approaches to the problem of stereoscopic 3D (S3D) visual discomfort prediction (VDP) have focused on extraction perceptually meaningful handcrafted features based models perception and natural depth statistics. Toward advancing performance this problem, we developed a deep learning-based VDP model named predictor (DeepVDP). The DeepVDP uses convolutional neural network (CNN) learn that are highly predictive experienced discomfort. Since large amount reference data is needed train...

Compressive sensing (CS) makes it possible to more naturally create compact representations of data with respect a desired rate. Through wavelet decomposition, smooth and piecewise signals can be represented as sparse compressible coefficients. These coefficients then effectively compressed via the CS. Since transform divides image information into layered blockwise over spatial frequency domains, visual improvement attained by an appropriate perceptually weighted CS scheme. We introduce...

Multiple object tracking (MOT) is a fundamental task in vision, but MOT techniques for plenoptic video are scarce. Almost all 2D algorithms that show high performance mostly use the detection-based method which has disadvantage of operating only specific object. To enable arbitrary desired objects, this paper introduces groundbreaking detection-free videos. The proposed deviates from traditional methods, emphasizing challenges targets with occlusions. presents specialized exploit multifocal...

This letter proposes a wavelet denoising method in the presence of missing data. approach is based on coupling shrinkage and hierarchical (or h)-likelihood method. The h-likelihood provides an effective imputation methodology data to give estimators for signals motivates fast simple algorithm. can be easily extended other settings, such as image denoising. Simulation studies demonstrate empirical properties proposed

현실세계에서 관찰되는 시그널(signal)은 다양한 주파수(frequency)들의 시그널로 혼합되어 있는 경우가 많다. 예를 들어 태양 흑점 자료의 경우 약 11년 주기와 85년 주기로 변동한다는 사실은 널리 알려져 있다. 또한 경제 시계열 경우는 통상적으로 계절요인(seasonal component), 순환요인(cyclic component) 그리고 장기적인 추세요인(long-term trend)으로 분해하여 분석한다. 이러한 자료를 구성요소별로 분해하는 것은 오래된 주제중 하나이다. 전통적인 시계열자료 분석기법으로 스펙트럴 분석기법 등이 사용되고 있으나 자료들이 비정상(nonstationary)일 경우에는 적용하기 어렵다. Huang et. al(1998)은 경험적 모드분해법(empirical mode decomposition)이라고 하는 자료적응적인(data-adaptive) 방법을 제안하였는데, 비정상성(nonstationarity)에 대한...

Over the past 20 years, research on quality of experience (QoE) has been actively expanded even to cover aesthetic, emotional and psychological experiences.QoE an important topic in determining perceptual factors that are essential users keeping with emergence new display technologies.In this paper, we provide in-depth reviews recent assessment studies field.Compared previous reviews, our examines human observed over various displays their associated methods.In study, first a comprehensive...

A hybrid architecture composed of a convolutional neural network (CNN) and Transformer is the new trend in realizing various vision tasks while pushing limits learning representation. From perspective mechanisms CNN Transformer, functional combination them suitable for image quality assessment (IQA) since which requires leveraging both local distortion perception global aggregation, however, there has been scarce study employing such an approach. This paper presents end-to-end...