{
  "id": 74144,
  "title": "Confusion about the filters and annotations",
  "url": "/competitions/human-protein-atlas-image-classification/discussion/74144",
  "author_name": "",
  "post_date": "2018-12-09T05:25:15.658246100Z",
  "votes": null,
  "comment_count": 2,
  "views": 0,
  "content": "<p>I checked the green filter of an image annotated as solely \"microtubules,\" then I compare it to the red filter.  The results are different, and why? Also, if yellow channel stands for ER, and there is a class called ER that we are supposed to predict, how come not every image is classified with the label ER?</p>",
  "messages": [
    {
      "id": "435938",
      "postDate": "12/09/2018 05:25:15",
      "content": "<p>I checked the green filter of an image annotated as solely \"microtubules,\" then I compare it to the red filter.  The results are different, and why? Also, if yellow channel stands for ER, and there is a class called ER that we are supposed to predict, how come not every image is classified with the label ER?</p>",
      "rawMarkdown": "I checked the green filter of an image annotated as solely \"microtubules,\" then I compare it to the red filter.  The results are different, and why? Also, if yellow channel stands for ER, and there is a class called ER that we are supposed to predict, how come not every image is classified with the label ER?",
      "votes": null
    },
    {
      "id": "435952",
      "postDate": "12/09/2018 06:20:59",
      "content": "<p>We are predicting the location of the protein, and the blue, red, yellow filters serve as landmarks. I think if an image's green filter overlaps with its yellow filter, then the label of that image should include ER, otherwise, it shouldn't.\nI learned a lot from Tilii's wonderful post. <a href=\"https://www.kaggle.com/c/human-protein-atlas-image-classification/discussion/72895\">https://www.kaggle.com/c/human-protein-atlas-image-classification/discussion/72895</a></p>",
      "rawMarkdown": "We are predicting the location of the protein, and the blue, red, yellow filters serve as landmarks. I think if an image's green filter overlaps with its yellow filter, then the label of that image should include ER, otherwise, it shouldn't.\nI learned a lot from Tilii's wonderful post. https://www.kaggle.com/c/human-protein-atlas-image-classification/discussion/72895",
      "votes": null
    },
    {
      "id": "436075",
      "postDate": "12/09/2018 14:20:29",
      "content": "<p>thanks!</p>",
      "rawMarkdown": "thanks!",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 435952,
      "author_name": "shisususu",
      "author_url": "",
      "post_date": "12/09/2018 06:20:59",
      "content": "<p>We are predicting the location of the protein, and the blue, red, yellow filters serve as landmarks. I think if an image's green filter overlaps with its yellow filter, then the label of that image should include ER, otherwise, it shouldn't.\nI learned a lot from Tilii's wonderful post. <a href=\"https://www.kaggle.com/c/human-protein-atlas-image-classification/discussion/72895\">https://www.kaggle.com/c/human-protein-atlas-image-classification/discussion/72895</a></p>",
      "votes": null,
      "replies": [
        {
          "id": 436075,
          "author_name": "alexanderliao",
          "author_url": "",
          "post_date": "12/09/2018 14:20:29",
          "content": "<p>thanks!</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "435938": "I checked the green filter of an image annotated as solely \"microtubules,\" then I compare it to the red filter.  The results are different, and why? Also, if yellow channel stands for ER, and there is a class called ER that we are supposed to predict, how come not every image is classified with the label ER?",
    "435952": "We are predicting the location of the protein, and the blue, red, yellow filters serve as landmarks. I think if an image's green filter overlaps with its yellow filter, then the label of that image should include ER, otherwise, it shouldn't.\nI learned a lot from Tilii's wonderful post. https://www.kaggle.com/c/human-protein-atlas-image-classification/discussion/72895",
    "436075": "thanks!"
  },
  "source": "meta"
}