A5101771714- Image Enhancement Techniques
- Advanced Neural Network Applications
- Infrared Target Detection Methodologies
- Photoacoustic and Ultrasonic Imaging
- Remote-Sensing Image Classification
- Advanced Image Fusion Techniques
- Metaheuristic Optimization Algorithms Research
- Neural Networks and Applications
- Advanced Clustering Algorithms Research
- Artificial Immune Systems Applications
- Target Tracking and Data Fusion in Sensor Networks
- Video Surveillance and Tracking Methods
- Aerospace and Aviation Technology
Nanjing University of Science and Technology
2023-2024
Beihang University
2016
Data augmentation methods offer a cost-effective and efficient alternative to the acquisition of additional data, significantly enhancing data diversity model generalization, making them particularly favored in object detection tasks. However, existing techniques primarily focus on visible spectrum are directly applied RGB-T tasks, overlooking inherent differences image between two Visible images capture rich color texture information during daytime, while infrared capable imaging under...
In harsh weather conditions, the infrared modality can supplement or even replace visible modality. However, lack of a large-scale dataset for features hinders generation robust pre-training model. Most existing object-detection algorithms rely on models from modality, which accelerate network convergence but also limit performance due to differences. order provide more reliable feature representation cross-modality object detection and enhance its performance, this paper investigates impact...
Existing infrared colorization methods based on image-to-image translation in a GAN framework require paired image data for network learning and are limited by quality. They also restricted their ability to perform spectral between bands, specifically, long-wave cannot be directly applied the near-infrared modality. This paper proposes novel algorithm that leverages similarities high-frequency features visible images. The method achieves through frequency domain feature decoupling...