A5084126948- Advanced Neural Network Applications
- Visual Attention and Saliency Detection
- Remote-Sensing Image Classification
- Video Surveillance and Tracking Methods
- Industrial Vision Systems and Defect Detection
- Advanced Image and Video Retrieval Techniques
- Infrared Target Detection Methodologies
- Advanced Image Fusion Techniques
- Advanced Computational Techniques and Applications
- Evolution and Genetic Dynamics
- Time Series Analysis and Forecasting
- Face Recognition and Perception
- Advanced Algorithms and Applications
- Autonomous Vehicle Technology and Safety
- Image Enhancement Techniques
- Olfactory and Sensory Function Studies
- Data-Driven Disease Surveillance
- Advanced Measurement and Detection Methods
- Parasites and Host Interactions
- Land Use and Ecosystem Services
- Alkaline Phosphatase Research Studies
- Surface Roughness and Optical Measurements
University of Electronic Science and Technology of China
2019-2025
Change detection plays a fundamental role in Earth observation for analyzing temporal iterations over time. However, recent studies have largely neglected the utilization of multimodal data that presents significant practical and technical advantages compared to single-modal approaches. This research focuses on leveraging pre-event digital surface model (DSM) post-event aerial images captured at different times detecting change beyond 2D. We observe current methods struggle with multitask...
In real-life complex traffic environments, vehicles are often occluded by extraneous background objects and other vehicles, leading to severe degradation of object detector performance. To address this issue, we propose a method named YOLO-OVD (YOLO for vehicle detection) dataset effectively handling occlusion in various scenarios. highlight the model attention unobstructed region design novel grouped orthogonal (GOA) module achieve maximum information extraction between channels. We utilize...
Abstract Despite the significant advancements made by deep visual networks in detecting surface defects at a regional level, challenge of achieving high-quality pixel-wise defect detection persists due to varied appearances and limited availability data. To address over-reliance on appearance enhance accuracy segmentation, we proposed Transformer-based Siamese network with change awareness, which formulates segmentation under complex background as mimic human inspection process....
The majority of modern object detectors rely on a set pre-defined anchor boxes, which enhances detection performance dramatically. Nevertheless, the strategy suffers some drawbacks, especially complex hyper-parameters anchors, seriously affecting performance. In this paper, we propose feature-guided generation method named dynamic anchor. Dynamic mainly includes two structures: generator and feature enhancement module. leverages semantic features to predict optimized shapes at locations...
Change detection plays a fundamental role in Earth observation for analyzing temporal iterations over time. However, recent studies have largely neglected the utilization of multimodal data that presents significant practical and technical advantages compared to single-modal approaches. This research focuses on leveraging {pre-event} digital surface model (DSM) {post-event} aerial images captured at different times detecting change beyond 2D. We observe current methods struggle with...
The deep neural network method has been widely used in the field of video object detection. However, detection performance such detectors are very sensitive to scaling change targets. To improve detecting performance, it's necessary use a large number sample data at different scales train model, but generally we could not get real multi-scale one scene. Aiming Faster-RCNN, state-of-the-art detector, this paper proposes based on Poisson fusion, its augment samples using method, which extracts...
In order to reduce computational redundancy and increase the speed of image analysis, Saliency Object Detection (SOD) is one outstanding methods for Very High Resolution (VHR) remote sensing analysis. However, Remote images (RSIs) have characteristics diverse spatial resolutions cluttered backgrounds, leading direct use SOD natural scenes generally not achieving satisfactory results. this paper, combining advantages Convolutional Sparse Coding (CSC) deep neural networks, a CSC network model...