A5008925597- Image Enhancement Techniques
- Video Surveillance and Tracking Methods
- Advanced Neural Network Applications
- Visual Attention and Saliency Detection
- X-ray Diffraction in Crystallography
- Crystallization and Solubility Studies
- Origins and Evolution of Life
- Protein Interaction Studies and Fluorescence Analysis
- Earth Systems and Cosmic Evolution
- Remote Sensing and LiDAR Applications
- Drug Transport and Resistance Mechanisms
- Human Pose and Action Recognition
- Industrial Vision Systems and Defect Detection
- Metal-Organic Frameworks: Synthesis and Applications
- Metal complexes synthesis and properties
- Photonic and Optical Devices
- Ionic liquids properties and applications
- Fluorine in Organic Chemistry
- Space Science and Extraterrestrial Life
- Advanced Chemical Physics Studies
- Advanced Image Fusion Techniques
- Multimodal Machine Learning Applications
- Porphyrin Metabolism and Disorders
- Automated Road and Building Extraction
- Hemoglobin structure and function
University of Science and Technology of China
2023-2025
Qufu Normal University
2021
Jining University
2007-2015
Nankai University
1997-1999
Recent methods using diffusion models have made significant progress in human image generation with various control signals such as pose priors. However, existing efforts are still struggling to generate high-quality images consistent alignment, resulting unsatisfactory output. In this paper, we propose a framework that delves into the graph relations of priors provide information for generation. The main idea is establish topological structure between and latent representation capture...
Lane segmentation at night is a challenging problem in autonomous driving perception, which beneficial to improve the robustness of application. Existing methods has shown great performance benchmark dataset, however, they do not consider bad lighting scenes practical applications, for example, lane algorithm will be greatly affected night. In this paper, we propose novel multi-modal nighttime algorithm, utilizes camera and LiDAR complementary information. We illustrate role image entropy...
Recent methods using diffusion models have made significant progress in human image generation with various additional controls such as pose priors. However, existing approaches still struggle to generate high-quality images consistent alignment, resulting unsatisfactory outputs. In this paper, we propose a framework delving into the graph relations of priors provide control information for generation. The main idea is establish topological structure between and latent representation capture...
The goal of low-light image enhancement is to restore the color and details great significance for high-level visual tasks in autonomous driving. However, it difficult lost dark area by relying only on RGB domain. In this paper we introduce frequency as a new clue into network propose novel DCT-driven transformer (DEFormer). First, learnable branch (LFB) contains DCT processing curvature-based (CFE). CFE calculates curvature each channel represent detail richness different bands, then...
Traditional machine learning algorithms have been used in forestry to identify pests, but the results can not adapt complex environment. In order realize rapid and accurate pest detection on devices with limited performance, an improved lightweight model ds-YOLOv3-tiny was proposed by us. Firstly, feature extraction network is changed DenseNet improve capability of enable extract high semantic features. We change activation function from Leaky ReLU Swish cover potential information loss...
In nature, the most possible reason that helix is chosen as basic structure of life molecule based on its simplest chiral three-dimensional structure. The process conversion from chemical molecules to double helical completed by topology effect which belongs way form helix, and no external power needed; moreover, energy has fixed drive direction [1]. dual-branch loop (II)—the transition state many uses, for example, it can be turned may broken into two fragments a b construct more...
Current object detection models have achieved good results on many benchmark datasets, detecting objects in dark conditions remains a large challenge. To address this issue, we propose pyramid enhanced network (PENet) and joint it with YOLOv3 to build framework named PE-YOLO. Firstly, PENet decomposes the image into four components of different resolutions using Laplacian pyramid. Specifically detail processing module (DPM) enhance images, which consists context branch edge branch. In...