![\"FIN-PRINT-workflow-including-1-dorsal-fin-saddle-patch-detection-2-extraction-of.png\"](\"https://i.postimg.cc/g2fnbs0K/FIN-PRINT-workflow-including-1-dorsal-fin-saddle-patch-detection-2-extraction-of.png\")
FIN-PRINT - a fully-automated multi-stage deep-learning-based framework for the individual recognition of killer whales
\nCode - https://github.com/ChristianBergler/FIN-PRINT
\nPaper - https://www.nature.com/articles/s41598-021-02506-6.pdf
EDIT - No update here, but for those interested, I am sharing 1500 annotations, sampling from all species proportionally - https://www.kaggle.com/bsridatta/happierwhale-1500-manual-annotations
", "messages": [ { "id": "1677877", "postDate": "02/06/2022 04:54:45", "content": "FIN-PRINT - a fully-automated multi-stage deep-learning-based framework for the individual recognition of killer whales
\nCode - https://github.com/ChristianBergler/FIN-PRINT
\nPaper - https://www.nature.com/articles/s41598-021-02506-6.pdf
EDIT - No update here, but for those interested, I am sharing 1500 annotations, sampling from all species proportionally - https://www.kaggle.com/bsridatta/happierwhale-1500-manual-annotations
", "rawMarkdown": "FIN-PRINT - a fully-automated multi-stage deep-learning-based framework for the individual recognition of killer whales\n\nCode - https://github.com/ChristianBergler/FIN-PRINT\nPaper - https://www.nature.com/articles/s41598-021-02506-6.pdf\n\n[](https://postimg.cc/hX1S99TJ)\n\n\nEDIT - No update here, but for those interested, I am sharing 1500 annotations, sampling from all species proportionally - https://www.kaggle.com/bsridatta/happierwhale-1500-manual-annotations", "votes": null }, { "id": "1678758", "postDate": "02/06/2022 19:07:36", "content": "The data or the weights for this paper doesn't seem to exist yet. I have created an issue in the repo and shall update here if anything comes up.
\nP.S There seems to be another efficient way to detect and crop fins called Photo-ID Ninja that has been used in other related papers. However, the tool seems to be taken down :(
expired-site: https://photoid.ninja/
\npresentation: link
I think it's very important to extract the fins to not lose subtle markers when scaling down the image for the CNN. If you find a easier way than annotating a thousand images, please share with the rest of us here :)
", "rawMarkdown": "The data or the weights for this paper doesn't seem to exist yet. I have created an issue in the repo and shall update here if anything comes up.\n\nP.S There seems to be another efficient way to detect and crop fins called `Photo-ID Ninja` that has been used in other related papers. However, the tool seems to be taken down :( \n\nexpired-site: https://photoid.ninja/\npresentation: [link](https://drive.google.com/file/d/1kiuWckr9nTYULBp8YjHQAXLPUf3vgRB7/view)\n\nI think it's very important to extract the fins to not lose subtle markers when scaling down the image for the CNN. If you find a easier way than annotating a thousand images, please share with the rest of us here :)", "votes": null }, { "id": "1684099", "postDate": "02/10/2022 09:12:48", "content": "thanks for sharing!
", "rawMarkdown": "thanks for sharing!", "votes": null }, { "id": "1701833", "postDate": "02/23/2022 05:09:28", "content": "No update here, but for those interested, I am sharing 1500 annotations, sampling from all species proportionally - https://www.kaggle.com/bsridatta/happierwhale-1500-manual-annotations
", "rawMarkdown": "No update here, but for those interested, I am sharing 1500 annotations, sampling from all species proportionally - https://www.kaggle.com/bsridatta/happierwhale-1500-manual-annotations", "votes": null }, { "id": "1702142", "postDate": "02/23/2022 11:13:39", "content": "Thank you very much for this work @bsridatta
\nI wish you all the best in your future work 👍
Thanks a lot @datascientistfp !
", "rawMarkdown": "Thanks a lot @datascientistfp !", "votes": null } ], "comments": [ { "id": 1678758, "author_name": "bsridatta", "author_url": "", "post_date": "02/06/2022 19:07:36", "content": "The data or the weights for this paper doesn't seem to exist yet. I have created an issue in the repo and shall update here if anything comes up.
\nP.S There seems to be another efficient way to detect and crop fins called Photo-ID Ninja that has been used in other related papers. However, the tool seems to be taken down :(
expired-site: https://photoid.ninja/
\npresentation: link
I think it's very important to extract the fins to not lose subtle markers when scaling down the image for the CNN. If you find a easier way than annotating a thousand images, please share with the rest of us here :)
", "votes": null, "replies": [] }, { "id": 1684099, "author_name": "maricinnamon", "author_url": "", "post_date": "02/10/2022 09:12:48", "content": "thanks for sharing!
", "votes": null, "replies": [] }, { "id": 1701833, "author_name": "bsridatta", "author_url": "", "post_date": "02/23/2022 05:09:28", "content": "No update here, but for those interested, I am sharing 1500 annotations, sampling from all species proportionally - https://www.kaggle.com/bsridatta/happierwhale-1500-manual-annotations
", "votes": null, "replies": [] }, { "id": 1702142, "author_name": "datascientistfp", "author_url": "", "post_date": "02/23/2022 11:13:39", "content": "Thank you very much for this work @bsridatta
\nI wish you all the best in your future work 👍
Thanks a lot @datascientistfp !
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "1677877": "FIN-PRINT - a fully-automated multi-stage deep-learning-based framework for the individual recognition of killer whales\n\nCode - https://github.com/ChristianBergler/FIN-PRINT\nPaper - https://www.nature.com/articles/s41598-021-02506-6.pdf\n\n[](https://postimg.cc/hX1S99TJ)\n\n\nEDIT - No update here, but for those interested, I am sharing 1500 annotations, sampling from all species proportionally - https://www.kaggle.com/bsridatta/happierwhale-1500-manual-annotations", "1678758": "The data or the weights for this paper doesn't seem to exist yet. I have created an issue in the repo and shall update here if anything comes up.\n\nP.S There seems to be another efficient way to detect and crop fins called `Photo-ID Ninja` that has been used in other related papers. However, the tool seems to be taken down :( \n\nexpired-site: https://photoid.ninja/\npresentation: [link](https://drive.google.com/file/d/1kiuWckr9nTYULBp8YjHQAXLPUf3vgRB7/view)\n\nI think it's very important to extract the fins to not lose subtle markers when scaling down the image for the CNN. If you find a easier way than annotating a thousand images, please share with the rest of us here :)", "1684099": "thanks for sharing!", "1701833": "No update here, but for those interested, I am sharing 1500 annotations, sampling from all species proportionally - https://www.kaggle.com/bsridatta/happierwhale-1500-manual-annotations", "1702142": "Thank you very much for this work @bsridatta \nI wish you all the best in your future work 👍", "1702166": "Thanks a lot @datascientistfp !" }, "source": "meta" }