In rate-distortion optimization, the encoder settings are determined by maximizing a reconstruction quality measure subject to constraint on bit rate. One of main challenges this approach is define that can be computed with low computational cost and which correlates well perceptual quality. While several measures fulfil these two criteria have been developed for images video, no such one exists 3D point clouds. We address limitation video-based cloud compression (V-PCC) standard proposing...

With the widespread adoption of multidevice communication, such as telecommuting, screen content images (SCIs) have become more closely and frequently related to our daily lives. For SCIs, tasks accurate visual quality assessment, high-efficiency compression, suitable contrast enhancement thus currently attracted increased attention. In particular, evaluation SCIs is important due its good ability for instruction optimization in various processing systems. Hence, this paper, we develop a new...

Research on screen content images (SCIs) becomes important as they are increasingly used in multi-device communication applications. In this paper, we present a study perceptual quality assessment of distorted SCIs subjectively and objectively. We construct large-scale image database (SIQAD) consisting 20 source 980 SCIs. order to get the subjective scores investigate, which part (text or picture) contributes more overall visual quality, single stimulus methodology with 11 point numerical...

Achieving subjective and objective quality assessment of underwater images is high significance in visual perception image/video processing. However, the development image (UIQA) limited for lack publicly available datasets with human scores reliable UIQA metrics. To address this issue, we establish a large-scale dataset, dubbed UID2021, evaluating no-reference (NR) The constructed dataset contains 60 multiply degraded collected from various sources, covering six common scenes (i.e., bluish...

We propose the first joint audio-video generation framework that brings engaging watching and listening experiences simultaneously, towards high-quality realistic videos. To generate pairs, we a novel Multi-Modal Diffusion model (i.e., MM-Diffusion), with two-coupled denoising autoencoders. In contrast to existing single-modal diffusion models, MM-Diffusion consists of sequential multi-modal U-Net for process by design. Two subnets audio video learn gradually aligned pairs from Gaussian...

Recently, 3D point cloud is becoming popular due to its capability represent the real world for advanced content modality in modern communication systems. In view of wide applications, especially immersive towards human perception, quality metrics clouds are essential. Existing evaluations rely on a full or certain portion original cloud, which severely limits their applications. To overcome this problem, we propose novel deep learning-based no reference assessment method, namely PQA-Net....

Legume cultivars affect N uptake, component crop growth, and soil physical chemical characteristics in maize–legume intercropping systems. However, how belowground interactions mediate root fixation, nodulation of different legumes to uptake is still unclear. Hence, a two-year experiment was conducted with five planting patterns, i.e., maize–soybean strip (IMS), maize–peanut (IMP), corresponding monocultures (monoculture maize (MM), monoculture soybean (MS), peanut (MP)), two application...

Image enhancement and restoration is among the most investigated topics in field of underwater machine vision. The objective image quality assessment a fundamental part optimizing technologies. However, no-reference (NR) metrics are not specifically designed for assessment. Moreover, since reference (undegraded) images available scenes, classical full-reference (FR) cannot be used to evaluate methods. In this paper, we first design an synthesis algorithm (UISA), which depending on real-world...

In this paper, we propose a novel method to address the nighttime single image dehazing problem. Estimation of ambient illumination map and transmission are key steps modern approaches. For hazy scenes at night, is usually not globally isotropic as scene typically contains multiple light sources. Frequently, Light source regions non-light exhibit distinct color features. However, existing methods have been attempting process these two based on identical prior assumptions. Moreover,...

In contemporary society full of stereoscopic images, how to assess visual quality 3D images has attracted an increasing attention in field Stereoscopic Image Quality Assessment (SIQA). Compared with 2D-IQA, SIQA is more challenging because some complicated features Human Visual System (HVS), such as binocular interaction and fusion, must be considered. this paper, considering both fusion mechanisms the HVS, a hierarchical no-reference image assessment network (StereoIF-Net) proposed simulate...

An important scenario for image quality assessment (IQA) is to evaluate restoration (IR) algorithms. The state-of-the-art approaches adopt a full-reference paradigm that compares restored images with their corresponding pristine-quality images. However, are usually unavailable in blind tasks and real-world scenarios. In this paper, we propose practical solution named degraded-reference IQA (DR-IQA), which exploits the inputs of IR models, degraded images, as references. Specifically, extract...

With the increasing demand of compressing and streaming 3D point clouds under constrained bandwidth, it has become ever more important to accurately efficiently determine quality compressed clouds, so as assess optimize quality-of-experience (QoE) end users. Here we make one first attempts developing a bitstream-based no-reference (NR) model for perceptual assessment without resorting full decoding data stream. Specifically, establish relationship between texture complexity bitrate...

How to incentivize strategic workers using limited budget is a very fundamental problem for crowdsensing systems; nevertheless, since the sensing abilities of may not always be known as prior knowledge due diversities their sensor devices and behaviors, it difficult properly select pay unknown workers. Although uncertainties can addressed by standard <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Combinatorial Multi-Armed Bandit</i> (CMAB)...

Point clouds are rapidly gaining popularity in many practical applications, and point cloud quality assessment (PCQA) is an important research topic that helps us measure improve the visual experience applications using clouds. Research on full-reference (FR) PCQAs has recently made impressive progress, no-reference (NR) also gradually increased. However, performance of prior NR PCQA methods still suffers from weak generalization ability lower accuracy than FR metrics general. In this work,...

Appearance change of moving targets is a challenging problem in visual tracking. In this paper, we present novel object tracking algorithm based on the observation dependent hidden Markov model (OD-HMM) framework. The dependency computed by structure complexity coefficients (SCC) which defined to predict target appearance change. Unlike conventional methods addressing investigating different online models, handle fundamental reason motion -related during Based analysis motion-related change,...

Research on Screen Content Images (SCIs) becomes important as they are increasingly used in multi-device communication applications. In this paper, we present a study of subjective quality assessment for distorted SCIs, and investigate which part (text or picture) contributes more to the overall visual quality. We construct large-scale Image Quality Assessment Database (SIQAD) consisting 20 source 980 SCIs. The 11-category Absolute Category Rating (ACR) is employed obtain three scores...

Many recent applications require text segmentation for born-digital compound images. To this end, we propose a coarse-to-fine framework segmenting texts of arbitrary scales and orientations in In the coarse stage, local image activity measure is designed based upon variation distribution characters, to highlight difference between textual pictorial regions. This stage outputs layer including regions as well few with high activity. fine connected component (TCC) refinement proposed eliminate...

Automatic spine and vertebra segmentation from X-ray images is a critical challenging problem in many computer-aid spinal image analysis disease diagnosis applications. In this paper, two-stage automatic framework for proposed, which can firstly locate the regions (including backbone, sacrum ilium) coarse stage then identify eighteen vertebrae (i.e., cervical 7, thoracic 1-12 lumbar 1-5) with isolate clear boundary fine stage. A novel Attention Gate based dual-pathway Network (AGNet)...

Semantic segmentation is a key technology for remote sensing image analysis widely used in land cover classification, natural disaster monitoring, and other fields. Unlike traditional segmentation, there are various targets images, with large feature difference between the targets. As result, more difficult, existing models retain low accuracy inaccurate edge when images. This paper proposes multi-attention-based semantic network images order to address these problems. Specifically, we...