{ "id": 180723, "title": "Council GAN - Breaking the Cycle - Colleagues are all you need ", "url": "/competitions/gan-getting-started/discussion/180723", "author_name": "", "post_date": "2020-09-06T08:45:35.210638400Z", "votes": 3, "comment_count": 1, "views": 0, "content": "

This paper proposes a novel approach to performing image-to-image translation between unpaired domains. Rather than relying on a cycle constraint, our method takes advantage of collaboration between various GANs.
\nThis results in a multi-modal method, in which multiple optional and diverse images are produced for a given image. Our model addresses some of the shortcomings of classical GANs:

\n
    \n
  1. It is able to remove large objects, such as glasses.
    2. \n
  2. Since it does not need to support the cycle constraint, no irrelevant traces of the input are left on the generated image.
    3. \n
  3. It manages to translate between domains that require large shape modifications.
    4. \n
\n

Link to Paper

\n

Link to Code

\n

Complete Project details here

\n

\"Image

", "messages": [ { "id": "1000056", "postDate": "09/06/2020 08:45:35", "content": "

This paper proposes a novel approach to performing image-to-image translation between unpaired domains. Rather than relying on a cycle constraint, our method takes advantage of collaboration between various GANs.
\nThis results in a multi-modal method, in which multiple optional and diverse images are produced for a given image. Our model addresses some of the shortcomings of classical GANs:

\n
    \n
  1. It is able to remove large objects, such as glasses.
    2. \n
  2. Since it does not need to support the cycle constraint, no irrelevant traces of the input are left on the generated image.
    3. \n
  3. It manages to translate between domains that require large shape modifications.
    4. \n
\n

Link to Paper

\n

Link to Code

\n

Complete Project details here

\n

\"Image

", "rawMarkdown": "This paper proposes a novel approach to performing image-to-image translation between unpaired domains. Rather than relying on a cycle constraint, our method takes advantage of collaboration between various GANs. \nThis results in a multi-modal method, in which multiple optional and diverse images are produced for a given image. Our model addresses some of the shortcomings of classical GANs:\n1. It is able to remove large objects, such as glasses.\n2. Since it does not need to support the cycle constraint, no irrelevant traces of the input are left on the generated image.\n3. It manages to translate between domains that require large shape modifications. \n\n***[Link to Paper](https://openaccess.thecvf.com/content_CVPR_2020/papers/Nizan_Breaking_the_Cycle_-_Colleagues_Are_All_You_Need_CVPR_2020_paper.pdf)***\n\n***[Link to Code](https://github.com/Onr/Council-GAN)***\n\n***[Complete Project details here](https://onr.github.io/Council_web/)***\n\n![Image from the Paper](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F743946%2F77bffaad5cec7bfed8e3c538dfff6450%2Ftest.png?generation=1599381899195063&alt=media)", "votes": null }, { "id": "1003935", "postDate": "09/09/2020 11:17:37", "content": "

i tried it 15000 iter but result look worse. Maybe because main of this competition is problem with color

", "rawMarkdown": "i tried it 15000 iter but result look worse. Maybe because main of this competition is problem with color", "votes": null } ], "comments": [ { "id": 1003935, "author_name": "doanquanvietnamca", "author_url": "", "post_date": "09/09/2020 11:17:37", "content": "

i tried it 15000 iter but result look worse. Maybe because main of this competition is problem with color

", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "1000056": "This paper proposes a novel approach to performing image-to-image translation between unpaired domains. Rather than relying on a cycle constraint, our method takes advantage of collaboration between various GANs. \nThis results in a multi-modal method, in which multiple optional and diverse images are produced for a given image. Our model addresses some of the shortcomings of classical GANs:\n1. It is able to remove large objects, such as glasses.\n2. Since it does not need to support the cycle constraint, no irrelevant traces of the input are left on the generated image.\n3. It manages to translate between domains that require large shape modifications. \n\n***[Link to Paper](https://openaccess.thecvf.com/content_CVPR_2020/papers/Nizan_Breaking_the_Cycle_-_Colleagues_Are_All_You_Need_CVPR_2020_paper.pdf)***\n\n***[Link to Code](https://github.com/Onr/Council-GAN)***\n\n***[Complete Project details here](https://onr.github.io/Council_web/)***\n\n![Image from the Paper](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F743946%2F77bffaad5cec7bfed8e3c538dfff6450%2Ftest.png?generation=1599381899195063&alt=media)", "1003935": "i tried it 15000 iter but result look worse. Maybe because main of this competition is problem with color" }, "source": "meta" }