Feng Cai

Feng Cai

Auto-generated profile, based on publicly available information

Feng Cai

Key Laboratory for Information Science of Electromagnetic Waves (MoE), Fudan University, Shanghai, China

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
30 August 2024 in IEEE Transactions on Geoscience and Remote Sensing
Verified
Journal Article
Closed access
ZLZhengxin Lei
Zhengxin Lei
Feng XuFeng Xu
Feng Xu
JWJiangtao Wei
Jiangtao Wei
FCFeng Cai
Feng Cai
FWFeng Wang
Feng Wang
YJYa-Qiu Jin
Ya-Qiu Jin
SAR images are highly sensitive to observation configurations, and exhibit significant variations across different viewing angles, making it challenging to represent and learn their anisotropic features. As a result, deep learning methods often generalize poorly across different view angles. Inspired by the concept of neural radiance fields (NeRF), this study combines SAR imaging mechanisms with neural networks to propose a novel NeRF model for SAR image generation. Following the mapping and projection principles, a set of SAR images is modeled implicitly as a function of attenuation coefficients and scattering intensities in the 3D imaging space through a differentiable rendering equation. SAR-NeRF is then constructed to learn the distribution of attenuation coefficients and scattering intensities of voxels, where the vectorized form of the 3D voxel SAR rendering equation and the sampling relationship between the 3D space voxels and the 2D view ray grids are analytically derived. Through quantitative experiments on various datasets, we thoroughly assess the multi-view representation and generalization capabilities of SAR-NeRF. Additionally, this paper includes few-shot classification performance improvement as a metric for generation performance. The study found that using 12 images per class resulted in an accuracy improvement of nearly 10% for the classification algorithm.

5 February 2024 in IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Verified
Journal Article
Open access
KWKeyu Wu
Keyu Wu
FCFeng Cai
Feng Cai
HWHaipeng Wang
Haipeng Wang
Airports represent essential infrastructure, offering substantial research and application potential. However, extracting airports from complex Synthetic Aperture Radar (SAR) scenes is challenging due to the cluttered background and fine structure of airports. This necessitates the integration of global and local information for fine-grained extraction. To tackle this issue, this paper introduces a novel framework for fine-grained extraction of airports from large-scale SAR images. First, a CNN-Transformer hybrid semantic segmentation network with multi-scale contextual fusion is proposed, named CTMANet. In this network, the encoder combines CNNs and Transformers to capture local and global information, while the Multi-Scale Context Aggregation (MCA) block fuses multi-scale contextual information. Skip connections between the encoder and decoder are established to minimize the loss of detailed information and fuse low-level features with high-level semantic features. Moreover, a Category Balance block is designed to address class imbalance. Experimental results on the GF-3 dataset demonstrate that CTMANet outperforms state-of-the-art methods, proving its superior suitability for fine-grained airport extraction in large-scale scenarios.