A5027370351- Generative Adversarial Networks and Image Synthesis
- Video Analysis and Summarization
- Remote-Sensing Image Classification
- Multimodal Machine Learning Applications
- Synthetic Aperture Radar (SAR) Applications and Techniques
- Advanced Image Processing Techniques
- Computer Graphics and Visualization Techniques
- Domain Adaptation and Few-Shot Learning
- Image Retrieval and Classification Techniques
- Advanced SAR Imaging Techniques
Yonsei University
2021-2024
We present a novel unsupervised framework for instance-level image-to-image translation. Although recent advances have been made by incorporating additional object annotations, existing methods often fail to handle images with multiple disparate objects. The main cause is that, during inference, they apply global style the whole image and do not consider large discrepancy between instance background, or within instances. To address this problem, we propose class-aware memory network that...
Synthetic aperture radar (SAR) building segmentation, which is one of the fundamental tasks in remote sensing community, has been achieved remarkable performance using convolutional neural networks (CNNs). Since most methods do not consider distinctive characteristics SAR images, they tend to be biased towards simple and large buildings while ignoring small complex-shaped ones. To build a general powerful segmentation model, this letter, we introduce semi-supervised learning (SSL) framework...
Automatic building footprint extraction from SAR imagery is one of the critical tasks in remote sensing community. CNN has been recently explored and achieved improved performance. However, due to scarcity training data, it suffers overfitting problem. This paper presents a novel knowledge distillation based framework consisting teacher student networks. Regarding EO image as privileged information, network learns extract rich pair features. The then estimate footprints only images on...
We present a novel unsupervised framework for instance-level image-to-image translation. Although recent advances have been made by incorporating additional object annotations, existing methods often fail to handle images with multiple disparate objects. The main cause is that, during inference, they apply global style the whole image and do not consider large discrepancy between instance background, or within instances. To address this problem, we propose class-aware memory network that...