{
  "id": 243431,
  "title": "【Albumentations】Bounding boxes augmentation for object detection",
  "url": "/competitions/siim-covid19-detection/discussion/243431",
  "author_name": "",
  "post_date": "2021-06-02T14:14:44.210461400Z",
  "votes": 14,
  "comment_count": 6,
  "views": 0,
  "content": "<p>If you want to do data augmentation in <strong>object detection</strong> competitions, <br>\n<a href=\"https://albumentations.ai/docs/getting_started/bounding_boxes_augmentation/\" target=\"_blank\">Albumentations</a> <br>\nwould be one of the best libraries to use.</p>\n<p>It transforms the image while preserving the coordinate information.<br>\nIt supports four coordinates formats: </p>\n<ul>\n<li>pascal_voc,</li>\n<li>albumentations,</li>\n<li>coco, </li>\n<li>yolo</li>\n</ul>\n<p><img src=\"https://albumentations.ai/docs/images/getting_started/augmenting_bboxes/bbox_augmentation_example_2.jpg\" alt=\"image\"></p>\n<p>I'm glad if It could help you.<br>\nThanks☺️</p>",
  "messages": [
    {
      "id": "1333157",
      "postDate": "06/02/2021 14:14:44",
      "content": "<p>If you want to do data augmentation in <strong>object detection</strong> competitions, <br>\n<a href=\"https://albumentations.ai/docs/getting_started/bounding_boxes_augmentation/\" target=\"_blank\">Albumentations</a> <br>\nwould be one of the best libraries to use.</p>\n<p>It transforms the image while preserving the coordinate information.<br>\nIt supports four coordinates formats: </p>\n<ul>\n<li>pascal_voc,</li>\n<li>albumentations,</li>\n<li>coco, </li>\n<li>yolo</li>\n</ul>\n<p><img src=\"https://albumentations.ai/docs/images/getting_started/augmenting_bboxes/bbox_augmentation_example_2.jpg\" alt=\"image\"></p>\n<p>I'm glad if It could help you.<br>\nThanks☺️</p>",
      "rawMarkdown": "If you want to do data augmentation in **object detection** competitions, \n[Albumentations](https://albumentations.ai/docs/getting_started/bounding_boxes_augmentation/) \nwould be one of the best libraries to use.\n\nIt transforms the image while preserving the coordinate information.\nIt supports four coordinates formats: \n- pascal_voc,\n- albumentations,\n- coco, \n- yolo\n\n![image](https://albumentations.ai/docs/images/getting_started/augmenting_bboxes/bbox_augmentation_example_2.jpg)\n\nI'm glad if It could help you.\nThanks☺️",
      "votes": null
    },
    {
      "id": "1334313",
      "postDate": "06/03/2021 12:09:37",
      "content": "<p>In addition to this(Albumentations),<br>\n<a href=\"https://www.kaggle.com/c/siim-covid19-detection/discussion/242760\" target=\"_blank\">https://www.kaggle.com/c/siim-covid19-detection/discussion/242760</a><br>\ndescribes more in detail about augmentation.<br>\nSo please check it.</p>",
      "rawMarkdown": "In addition to this(Albumentations),\nhttps://www.kaggle.com/c/siim-covid19-detection/discussion/242760\ndescribes more in detail about augmentation.\nSo please check it.",
      "votes": null
    },
    {
      "id": "1334576",
      "postDate": "06/03/2021 15:51:30",
      "content": "<p>Thanks, very helpful. Upvoted.</p>",
      "rawMarkdown": "Thanks, very helpful. Upvoted.",
      "votes": null
    },
    {
      "id": "1335525",
      "postDate": "06/04/2021 09:13:00",
      "content": "<p><a href=\"https://www.kaggle.com/shoku-pan\" target=\"_blank\">@shoku-pan</a>  Great post! Upvoted. I like how you have images, it's so much easier to understand the flow. May I ask, what exactly does \"transformed\" do to the image? Does it just flip the image vertically? Is is dependant on the parameters? Sorry, I am new to data science 😓 Thanks!</p>",
      "rawMarkdown": "shoku-pan  Great post! Upvoted. I like how you have images, it's so much easier to understand the flow. May I ask, what exactly does \"transformed\" do to the image? Does it just flip the image vertically? Is is dependant on the parameters? Sorry, I am new to data science 😓 Thanks!",
      "votes": null
    },
    {
      "id": "1341768",
      "postDate": "06/09/2021 01:31:59",
      "content": "<p>Thank you☺️<br>\nAnd sorry for my late reply.</p>\n<p>You can use not only VerticalFlip but also HorizontalFlip, ShiftScaleRotate, Blur and many other methods.</p>\n<p>For more detail, please check this☺️<br>\n<a href=\"https://albumentations.ai/docs/examples/example/\" target=\"_blank\">https://albumentations.ai/docs/examples/example/</a></p>",
      "rawMarkdown": "Thank you☺️\nAnd sorry for my late reply.\n\nYou can use not only VerticalFlip but also HorizontalFlip, ShiftScaleRotate, Blur and many other methods.\n\nFor more detail, please check this☺️\nhttps://albumentations.ai/docs/examples/example/",
      "votes": null
    },
    {
      "id": "1341770",
      "postDate": "06/09/2021 01:32:29",
      "content": "<p>Thank you too!</p>",
      "rawMarkdown": "Thank you too!",
      "votes": null
    },
    {
      "id": "1342517",
      "postDate": "06/09/2021 14:21:19",
      "content": "<p>This too☺️</p>\n<p><a href=\"https://www.kaggle.com/boltcoder/siim-covid19-simple-data-augmentation-techniques\" target=\"_blank\">https://www.kaggle.com/boltcoder/siim-covid19-simple-data-augmentation-techniques</a></p>",
      "rawMarkdown": "This too☺️\n\nhttps://www.kaggle.com/boltcoder/siim-covid19-simple-data-augmentation-techniques",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1334313,
      "author_name": "yhirakawa",
      "author_url": "",
      "post_date": "06/03/2021 12:09:37",
      "content": "<p>In addition to this(Albumentations),<br>\n<a href=\"https://www.kaggle.com/c/siim-covid19-detection/discussion/242760\" target=\"_blank\">https://www.kaggle.com/c/siim-covid19-detection/discussion/242760</a><br>\ndescribes more in detail about augmentation.<br>\nSo please check it.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1342517,
          "author_name": "yhirakawa",
          "author_url": "",
          "post_date": "06/09/2021 14:21:19",
          "content": "<p>This too☺️</p>\n<p><a href=\"https://www.kaggle.com/boltcoder/siim-covid19-simple-data-augmentation-techniques\" target=\"_blank\">https://www.kaggle.com/boltcoder/siim-covid19-simple-data-augmentation-techniques</a></p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1334576,
      "author_name": "stevelangs",
      "author_url": "",
      "post_date": "06/03/2021 15:51:30",
      "content": "<p>Thanks, very helpful. Upvoted.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1341770,
          "author_name": "yhirakawa",
          "author_url": "",
          "post_date": "06/09/2021 01:32:29",
          "content": "<p>Thank you too!</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1335525,
      "author_name": "yohmori02",
      "author_url": "",
      "post_date": "06/04/2021 09:13:00",
      "content": "<p><a href=\"https://www.kaggle.com/shoku-pan\" target=\"_blank\">@shoku-pan</a>  Great post! Upvoted. I like how you have images, it's so much easier to understand the flow. May I ask, what exactly does \"transformed\" do to the image? Does it just flip the image vertically? Is is dependant on the parameters? Sorry, I am new to data science 😓 Thanks!</p>",
      "votes": null,
      "replies": [
        {
          "id": 1341768,
          "author_name": "yhirakawa",
          "author_url": "",
          "post_date": "06/09/2021 01:31:59",
          "content": "<p>Thank you☺️<br>\nAnd sorry for my late reply.</p>\n<p>You can use not only VerticalFlip but also HorizontalFlip, ShiftScaleRotate, Blur and many other methods.</p>\n<p>For more detail, please check this☺️<br>\n<a href=\"https://albumentations.ai/docs/examples/example/\" target=\"_blank\">https://albumentations.ai/docs/examples/example/</a></p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1333157": "If you want to do data augmentation in **object detection** competitions, \n[Albumentations](https://albumentations.ai/docs/getting_started/bounding_boxes_augmentation/) \nwould be one of the best libraries to use.\n\nIt transforms the image while preserving the coordinate information.\nIt supports four coordinates formats: \n- pascal_voc,\n- albumentations,\n- coco, \n- yolo\n\n![image](https://albumentations.ai/docs/images/getting_started/augmenting_bboxes/bbox_augmentation_example_2.jpg)\n\nI'm glad if It could help you.\nThanks☺️",
    "1334313": "In addition to this(Albumentations),\nhttps://www.kaggle.com/c/siim-covid19-detection/discussion/242760\ndescribes more in detail about augmentation.\nSo please check it.",
    "1334576": "Thanks, very helpful. Upvoted.",
    "1335525": "shoku-pan  Great post! Upvoted. I like how you have images, it's so much easier to understand the flow. May I ask, what exactly does \"transformed\" do to the image? Does it just flip the image vertically? Is is dependant on the parameters? Sorry, I am new to data science 😓 Thanks!",
    "1341768": "Thank you☺️\nAnd sorry for my late reply.\n\nYou can use not only VerticalFlip but also HorizontalFlip, ShiftScaleRotate, Blur and many other methods.\n\nFor more detail, please check this☺️\nhttps://albumentations.ai/docs/examples/example/",
    "1341770": "Thank you too!",
    "1342517": "This too☺️\n\nhttps://www.kaggle.com/boltcoder/siim-covid19-simple-data-augmentation-techniques"
  },
  "source": "meta"
}