Summary of Training Deep Learning Models with Hybrid Datasets For Robust Automatic Target Detection on Real Sar Images, by Benjamin Camus et al.
Training Deep Learning Models with Hybrid Datasets for Robust Automatic Target Detection on real SAR images
by Benjamin Camus, Théo Voillemin, Corentin Le Barbu, Jean-Christophe Louvigné, Carole Belloni, Emmanuel Vallée
First submitted to arxiv on: 15 May 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Signal Processing (eess.SP)
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 This paper tackles challenges in developing Automatic Target Detection (ATD) algorithms for ground targets in SAR images by proposing a Deep Learning approach that combines synthetic target signatures with real backgrounds. The authors define an incrustation pipeline to generate hybrid datasets, which are then used to train ATD models tailored to bridge the domain gap between synthetic and real data. The approach relies on massive physics-based data augmentation techniques and Adversarial Training of two deep-learning detection architectures. The models are tested on several datasets, achieving up to 90% Average Precision on real data while exclusively using synthetic targets for training. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper helps improve Automatic Target Detection (ATD) algorithms by creating fake target signatures and combining them with real backgrounds. This allows the algorithm to learn from both synthetic and real data. The authors test their approach on different types of images and find that it can accurately detect targets while only using fake data for training. |
Keywords
» Artificial intelligence » Data augmentation » Deep learning » Precision
