- Image and Signal Denoising Methods
- Medical Image Segmentation Techniques
- Advanced Image Processing Techniques
- Advanced Image Fusion Techniques
- Image Enhancement Techniques
- 3D Shape Modeling and Analysis
- Image and Object Detection Techniques
- Advanced Vision and Imaging
- Image Retrieval and Classification Techniques
- Advanced Image and Video Retrieval Techniques
- Image Processing Techniques and Applications
- Sparse and Compressive Sensing Techniques
- Face recognition and analysis
- Biometric Identification and Security
- Distributed and Parallel Computing Systems
- Advanced Numerical Analysis Techniques
- Computer Graphics and Visualization Techniques
- Image and Video Quality Assessment
- Cloud Computing and Resource Management
- Generative Adversarial Networks and Image Synthesis
- Peer-to-Peer Network Technologies
- Remote-Sensing Image Classification
- Mobile Ad Hoc Networks
- Visual Attention and Saliency Detection
- stochastic dynamics and bifurcation
Qingdao University
2014-2024
Qingdao University of Science and Technology
2017-2023
Underwater captured images are usually degraded by low contrast, hazy, and blurry due to absorbing scattering, which limits their analyses applications. To address these problems, a red channel prior guided variational framework is proposed based on the complete underwater image formation model (UIFM). Unlike most of existing methods that only consider direct transmission backscattering components, we additionally include forward scattering component into UIFM. In framework, successfully...
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...
High‐order variational models are powerful methods for image processing and analysis, but they can lead to complicated high‐order nonlinear partial differential equations that difficult discretise solve computationally. In this paper, we present some representative provide detailed descretisation of these numerical implementation the split Bregman algorithm solving using fast Fourier transform. We demonstrate advantages disadvantages in context denoising through extensive experiments. The...
In this study, a novel underwater colour image enhancement approach based on hue preserving is presented by combining hue–saturation–intensity (HSI) and HS–value (HSV) models. the proposed wavelet‐domain filtering (WDF) constrained histogram stretching (CHS) algorithms are operated HSI HSV models, respectively. The degraded first converted from red–green–blue model into model, wherein component H preserved WDF algorithm executed S I components. Similarly, further kept invariant as well CHS...
The problem of routing security in vehicular ad hoc networks has become a major concern for researchers. Compared with cryptography-based solutions, trust-based solutions are considered to be more acceptable as promising approach, which mainly define two operations: trust computing and application. In this paper we first study properties construct novel inference model, where attributes named subjective recommendation trust, selected quantify the level specific vehicle. SCGM(1,1)-weighted...
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...
This paper advocates a novel learning solution to the modeling of long-term spatial-temporal saliency consistency in order boost accuracy for video detection. Conventional methods typically utilize "slack" model locally ensure smoothness computed saliency, yet they could easily encounter performance tradeoff dilemma (i.e., detection' and integrity). In contrast, our approach proposes bilevel strategy globally exploit while overcoming aforementioned difficulty. Our method first starts with...
This article proposes a multimode medical image fusion with CNN and supervised learning, in order to solve the problem of practical diagnosis. It can implement different types multimodal problems batch processing mode effectively overcome that traditional only be solved by single fusion. To certain extent, it greatly improves effect, detail clarity, time efficiency new method. The experimental results indicate proposed method exhibits state-of-the-art performance terms visual quality variety...
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...
Recently, there has been growing interest in hyperspectral images (HSIs) classification tasks, with both Graph Neural Networks (GNN) and Convolutional (CNN) proving to be effective means of analysis. GNN can better capture the spatial structure HSIs large target irregular regions through superpixel segmentation, while CNN refine tasks by processing pixel-level features small regular regions. However, neither nor models alone simultaneously consider superpixel-level cover To fully utilize...
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...
In hyperspectral images (HSIs), both local and non-local features play crucial roles in classification tasks. Vision Transformer (VIT) can extract through attention mechanisms, while Convolutional Neural Networks (CNN) excel at handling components. However, traditional dual-branch models based on VIT CNN, there is a lack of interaction during feature processing, leading to potential compatibility issues when merging the two types features. this article, we propose HyperSINet, Synergetic...
Simulation speed is crucial for virtual reality simulators involving real-time interactive cutting of deformable objects, such as surgical simulators. Previous efforts to accelerate these simulations resulted in significant increases during non-cutting periods, but only moderate ones periods. This paper aims further increase the latter. Three novel methods are proposed: (1) GPU-based update mass and stiffness matrices composite finite elements. (2) collision processing between tools objects....