The paper presents a novel method, Zero-Reference Deep Curve Estimation (Zero-DCE), which formulates light enhancement as task of image-specific curve estimation with deep network. Our method trains lightweight network, DCE-Net, to estimate pixel-wise and high-order curves for dynamic range adjustment given image. is specially designed, considering pixel value range, monotonicity, differentiability. Zero-DCE appealing in its relaxed assumption on reference images, i.e., it does not require...

Underwater image enhancement has been attracting much attention due to its significance in marine engineering and aquatic robotics. Numerous underwater algorithms have proposed the last few years. However, these are mainly evaluated using either synthetic datasets or selected real-world images. It is thus unclear how would perform on images acquired wild we could gauge progress field. To bridge this gap, present first comprehensive perceptual study analysis of large-scale In paper, construct...

Images captured under water are usually degraded due to the effects of absorption and scattering. Degraded underwater images show some limitations when they used for display analysis. For example, with low contrast color cast decrease accuracy rate object detection marine biology recognition. To overcome those limitations, a systematic image enhancement method, which includes an dehazing algorithm algorithm, is proposed. Built on minimum information loss principle, effective proposed restore...

Underwater images suffer from color casts and low contrast due to wavelength- distance-dependent attenuation scattering. To solve these two degradation issues, we present an underwater image enhancement network via medium transmission-guided multi-color space embedding, called Ucolor. Concretely, first propose a encoder network, which enriches the diversity of feature representations by incorporating characteristics different spaces into unified structure. Coupled with attention mechanism,...

Deep convolutional neural networks have achieved competitive performance in salient object detection, which how to learn effective and comprehensive features plays a critical role. Most of the previous works mainly adopted multiple-level feature integration yet ignored gap between different features. Besides, there also exists dilution process high-level as they passed on top-down pathway. To remedy these issues, we propose novel network named GCPANet effectively integrate low-level...

Visual saliency detection model simulates the human visual system to perceive scene, and has been widely used in many vision tasks. With acquisition technology development, more comprehensive information, such as depth cue, inter-image correspondence, or temporal relationship, is available extend image RGBD detection, co-saliency video detection. focuses on extracting salient regions from images by combining information. Co-saliency introduces correspondence constraint discover common object...

Stereoscopic perception is an important part of human visual system that allows the brain to perceive depth. However, depth information has not been well explored in existing saliency detection models. In this letter, a novel method for stereoscopic images proposed. First, we propose measure evaluate reliability map, and use it reduce influence poor map on detection. Then, input image represented as graph, introduced into graph construction. After that, new definition compactness using color...

Despite single image dehazing has been made promising progress with Convolutional Neural Networks (CNNs), the inherent equivariance and locality of convolution still bottleneck deharing performance. Though Transformer occupied various computer vision tasks, directly leveraging for is challenging: 1) it tends to result in ambiguous coarse details that are undesired reconstruction; 2) previous position embedding provided logic or spatial order neglects variational haze densities, which results...

Salient object detection from RGB-D images is an important yet challenging vision task, which aims at detecting the most distinctive objects in a scene by combining color information and depth constraints. Unlike prior fusion manners, we propose attention steered interweave network (ASIF-Net) to detect salient objects, progressively integrates cross-modal cross-level complementarity RGB image corresponding map via steering of mechanism. Specifically, complementary features are jointly...

Despite the remarkable advances in visual saliency analysis for natural scene images (NSIs), salient object detection (SOD) optical remote sensing (RSIs) still remains an open and challenging problem. In this paper, we propose end-to-end Dense Attention Fluid Network (DAFNet) SOD RSIs. A Global Context-aware (GCA) module is proposed to adaptively capture long-range semantic context relationships, further embedded a (DAF) structure that enables shallow attention cues flow into deep layers...

There are two main issues in RGB-D salient object detection: (1) how to effectively integrate the complementarity from cross-modal data; (2) prevent contamination effect unreliable depth map. In fact, these problems linked and intertwined, but previous methods tend focus only on first problem ignore consideration of map quality, which may yield model fall into sub-optimal state. this paper, we address a holistic synergistically, propose novel network named DPANet explicitly potentiality...

Rapid development of affordable and portable consumer depth cameras facilitates the use information in many computer vision tasks such as intelligent vehicles 3D reconstruction. However, map captured by low-cost sensors (e.g., Kinect) usually suffers from low spatial resolution, which limits its potential applications. In this paper, we propose a novel deep network for super-resolution (SR), called DepthSR-Net. The proposed DepthSR-Net automatically infers high-resolution (HR) low-resolution...

Depth information has been demonstrated to be useful for saliency detection. However, the existing methods RGBD detection mainly focus on designing straightforward and comprehensive models, while ignoring transferable ability of RGB models. In this article, we propose a novel depth-guided transformation model (DTM) going from saliency. The proposed includes three components, that is: 1) multilevel initialization; 2) refinement; 3) optimization with depth constraints. explicit feature is...

Due to the attenuation and scattering of light by water, there are many quality defects in raw underwater images such as color casts, decreased visibility, reduced contrast, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">et al.</i> . Many different image enhancement (UIE) algorithms have been proposed enhance quality. However, how fairly compare performance among UIE remains a challenging problem. So far, lack comprehensive human subjective...

Salient object detection (SOD) for optical remote sensing images (RSIs) aims at locating and extracting visually distinctive objects/regions from the RSIs. Despite some saliency models were proposed to solve intrinsic problem of RSIs (such as complex background scale-variant objects), accuracy completeness are still unsatisfactory. To this end, we propose a relational reasoning network with parallel multi-scale attention SOD in paper. The module that integrates spatial channel dimensions is...

Focusing on the issue of how to effectively capture and utilize cross-modality information in RGB-D salient object detection (SOD) task, we present a convolutional neural network (CNN) model, named CIR-Net, based novel interaction refinement. For interaction, 1) progressive attention guided integration unit is proposed sufficiently integrate feature representations encoder stage, 2) convergence aggregation structure proposed, which flows RGB depth decoding features into corresponding streams...

Optical remote sensing images (RSIs) have been widely used in many applications, and one of the interesting issues about optical RSIs is salient object detection (SOD). However, due to diverse types, various scales, numerous orientations, cluttered backgrounds RSIs, performance existing SOD models often degrade largely. Meanwhile, cutting-edge targeting typically focus on suppressing backgrounds, while they neglect importance edge information which crucial for obtaining precise saliency...

Due to the light absorption and scattering induced by water medium, underwater images usually suffer from some degradation problems, such as low contrast, color distortion, blurring details, which aggravate difficulty of downstream understanding tasks. Therefore, how obtain clear visually pleasant has become a common concern people, task image enhancement (UIE) also emerged times require. Among existing UIE methods, Generative Adversarial Networks (GANs) based methods perform well in visual...

In recent years, various neural network architectures for computer vision have been devised, such as the visual transformer and multilayer perceptron (MLP). A based on an attention mechanism can outperform a traditional convolutional network. Compared with transformer, MLP introduces less inductive bias achieves stronger generalization. addition, shows exponential increase in inference, training, debugging times. Considering wave function representation, we propose WaveNet architecture that...

Recently, deep convolutional neural networks (CNNs) have provided us an effective tool for automated polyp segmentation in colonoscopy images. However, most CNN-based methods do not fully consider the feature interaction among different layers and often cannot provide satisfactory performance. In this article, a novel attention-guided pyramid context network (APCNet) is proposed accurate robust Specifically, considering that represent aspects, APCNet first extracts multilayer features...

Recently, camouflaged object detection (COD), which suffers from numerous challenges such as low contrast between objects and background large variations of appearances, has received more concerns. However, the performance existing methods is still unsatisfactory, especially when dealing with complex scenes. Therefore, in this paper, we propose a novel Decoupling Integration Network (DINet) to detect objects. Here, depiction can be regarded iterative decoupling integration body features...

During recent years, we have witnessed a rapid development of wireless network technologies which revolutionized the way people take and share multimedia content. However, images captured in outdoor scenes usually suffer from limited visibility due to suspended atmospheric particles, directly affects quality photos. Despite progress image dehazing methods, visual dehazed results still needs further improvement. In this paper, propose deep convolutional neural (CNN) for single called PDR-Net,...

Co-saliency detection aims at extracting the common salient regions from an image group containing two or more relevant images. It is a newly emerging topic in computer vision community. Different most existing co-saliency methods focusing on RGB images, this paper proposes novel model for RGBD which utilizes depth information to enhance identification of co-saliency. First, intra saliency map each generated by single model, while inter calculated based multi-constraint feature matching,...