Summary of Multisource Collaborative Domain Generalization For Cross-scene Remote Sensing Image Classification, by Zhu Han et al.
Multisource Collaborative Domain Generalization for Cross-Scene Remote Sensing Image Classification
by Zhu Han, Ce Zhang, Lianru Gao, Zhiqiang Zeng, Michael K. Ng, Bing Zhang, Jocelyn Chanussot
First submitted to arxiv on: 5 Dec 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Machine Learning (cs.LG)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The proposed multi-source collaborative domain generalization framework (MS-CDG) tackles cross-scene image classification in remote sensing by leveraging multiple sources of data, incorporating homogeneity and heterogeneity characteristics. The approach combines data-aware adversarial augmentation and model-aware multi-level diversification to enhance performance. Key innovations include an adversary neural network with semantic guide for generating realistic samples across domains and a class-wise prototype and kernel mixture module for addressing domain discrepancies. Evaluation on three public datasets demonstrates the method’s superiority over state-of-the-art approaches. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Cross-scene image classification in remote sensing aims to use prior knowledge of ground materials to label different regions without needing as much human effort. Current methods only work well when there’s not too big a difference between the training data and new, unseen areas. But this isn’t always the case. To solve this problem, researchers propose a new way to combine multiple sources of remote sensing data. They use an “adversary” neural network that can create fake samples by changing channel and distribution details from one domain to another. This helps the model learn how to generalize across different regions. The team also introduces a module that transforms shared spatial-channel features into class-wise prototypes and kernel mixtures, making it better at handling domain differences and clustering classes together effectively. In the end, they test their method on three public datasets and show that it outperforms existing approaches. |
Keywords
» Artificial intelligence » Clustering » Domain generalization » Image classification » Neural network
