{
  "id": 109458,
  "title": "How to Remove False Positives?",
  "url": "/competitions/understanding_cloud_organization/discussion/109458",
  "author_name": "0DD1",
  "post_date": "2019-09-19T10:37:28.107000",
  "votes": 4,
  "comment_count": 0,
  "views": 0,
  "content": "<p>I want to remove the cases of False Positives (no pattern in test image but my model predicted a bounding box). Is training another binary classifier first to predict pattern presence and then using the Image Segmentation model to predict bounding boxes only on the true cases of the previous model a good idea? What else can be done to resolve this issue?  </p>",
  "messages": [
    {
      "id": 629861,
      "postDate": "2019-09-19T10:37:28.107Z",
      "content": "<p>I want to remove the cases of False Positives (no pattern in test image but my model predicted a bounding box). Is training another binary classifier first to predict pattern presence and then using the Image Segmentation model to predict bounding boxes only on the true cases of the previous model a good idea? What else can be done to resolve this issue?  </p>",
      "rawMarkdown": "I want to remove the cases of False Positives (no pattern in test image but my model predicted a bounding box). Is training another binary classifier first to predict pattern presence and then using the Image Segmentation model to predict bounding boxes only on the true cases of the previous model a good idea? What else can be done to resolve this issue?  ",
      "votes": 4
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "629861": "I want to remove the cases of False Positives (no pattern in test image but my model predicted a bounding box). Is training another binary classifier first to predict pattern presence and then using the Image Segmentation model to predict bounding boxes only on the true cases of the previous model a good idea? What else can be done to resolve this issue?  "
  }
}