Masked auto-encoding and scatter-decoupling transformer for polarimetric SAR image classification

Pattern Recognition, Journal Year: 2025, Volume and Issue: 166, P. 111660 - 111660

Published: April 12, 2025

Language: Английский

---

Hierarchical Context Learning of object components for unsupervised semantic segmentation

Pattern Recognition, Journal Year: 2025, Volume and Issue: unknown, P. 111713 - 111713

Published: April 1, 2025

Language: Английский

---

Triple Graph Convolutional Network for Hyperspectral Image Feature Fusion and Classification

Remote Sensing, Journal Year: 2025, Volume and Issue: 17(9), P. 1623 - 1623

Published: May 3, 2025

Most graph-based networks utilize superpixel generation methods as a preprocessing step, considering superpixels graph nodes. In the case of hyperspectral images having high variability in spectral features, an image region node may degrade class discrimination ability for pixel-based classification. Moreover, most focus on global feature extraction, while both local and information are important To deal with these challenges, superpixel-based graphs overruled this work, Graph-based Feature Fusion (GF2) method relying three different is proposed instead. A patch considered around each pixel under test, at same time, anchors highest informational content selected from entire scene. While first explores relationships between neighboring pixels anchors, second third use nodes, respectively. These processed using convolutional networks, their results fused cross-attention mechanism. The experiments benchmark datasets show that GF2 network has classification performance compared to state-of-the-art methods, imposing reasonable number learnable parameters.

Language: Английский

---

Underwater image captioning via attention mechanism based fusion of visual and textual information

Information Fusion, Journal Year: 2025, Volume and Issue: 123, P. 103269 - 103269

Published: May 5, 2025

Language: Английский

---