Hope these papers give you a better understanding of how the problem is addressed in various categories and settings.
\nAnd here's a bonus paper to all you curious Kagglers on view aware contrastive learning. Might give the inspiration to handle the dramatic differences in the views for the same individuals. In short, the loss between samples of the same individual from various views hurts the mode. Learning the view and having a different loss for similar and dissimilar views improves the metric learning. Paper
\nMight update here again.
", "messages": [ { "id": 1689127, "postDate": "2022-02-14T04:01:55.133Z", "content": "Hope these papers give you a better understanding of how the problem is addressed in various categories and settings.
\nAnd here's a bonus paper to all you curious Kagglers on view aware contrastive learning. Might give the inspiration to handle the dramatic differences in the views for the same individuals. In short, the loss between samples of the same individual from various views hurts the mode. Learning the view and having a different loss for similar and dissimilar views improves the metric learning. Paper
\nMight update here again.
", "rawMarkdown": "Hope these papers give you a better understanding of how the problem is addressed in various categories and settings.\n\n- Paper with general Insight into the problem [The Animal ID Problem: Continual Curation](https://arxiv.org/pdf/2106.10377v1.pdf)\n- Multiple papers and approaches from WACV2020\n[Summary from Deep Learning for Animal Re-Identification Workshop](https://sites.google.com/view/wacv2020animalreid/)\n- Surveys the related works across different animals [Perspectives on individual animal identification from biology and computer vision](https://arxiv.org/pdf/2103.00560.pdf)\n- Bird ID via classification, tries to handle new birds [Deep learning-based methods for individual recognition in small birds\n](https://besjournals.onlinelibrary.wiley.com/doi/full/10.1111/2041-210x.13436)\n\n\n\nAnd here's a bonus paper to all you curious Kagglers on view aware contrastive learning. Might give the inspiration to handle the dramatic differences in the views for the same individuals. In short, the loss between samples of the same individual from various views hurts the mode. Learning the view and having a different loss for similar and dissimilar views improves the metric learning. [Paper](https://arxiv.org/abs/1910.04104v1)\n\n\n*Might* update here again.\n", "votes": 12 } ], "comments": [], "raw_markdown_by_id": { "1689127": "Hope these papers give you a better understanding of how the problem is addressed in various categories and settings.\n\n- Paper with general Insight into the problem [The Animal ID Problem: Continual Curation](https://arxiv.org/pdf/2106.10377v1.pdf)\n- Multiple papers and approaches from WACV2020\n[Summary from Deep Learning for Animal Re-Identification Workshop](https://sites.google.com/view/wacv2020animalreid/)\n- Surveys the related works across different animals [Perspectives on individual animal identification from biology and computer vision](https://arxiv.org/pdf/2103.00560.pdf)\n- Bird ID via classification, tries to handle new birds [Deep learning-based methods for individual recognition in small birds\n](https://besjournals.onlinelibrary.wiley.com/doi/full/10.1111/2041-210x.13436)\n\n\n\nAnd here's a bonus paper to all you curious Kagglers on view aware contrastive learning. Might give the inspiration to handle the dramatic differences in the views for the same individuals. In short, the loss between samples of the same individual from various views hurts the mode. Learning the view and having a different loss for similar and dissimilar views improves the metric learning. [Paper](https://arxiv.org/abs/1910.04104v1)\n\n\n*Might* update here again.\n" } }