{
  "id": 103492,
  "title": "Evaluating the ICDAR2013 gender identification scores",
  "url": "/competitions/icdar2013-gender-prediction-from-handwriting/discussion/103492",
  "author_name": "EvyaIL",
  "post_date": "2019-08-09T14:51:47.033000",
  "votes": 0,
  "comment_count": 0,
  "views": 0,
  "content": "<p>Hi,\nRecently I have started to engage with this field and built a CNN model in order to predict a gender through offline handwriting.\nI would like to evaluate my performance on the QUWI subset, used in the competition. My output contains a list of the probability of a form to be male (actually the softmax outputed by the network) for each QUWI test form. Apparently, my results seem to be very good and i'm suspicious about it.</p>\n\n<p>There is any source to get an output example with the male scores of one of the methods used in the competition?  I wish for an example of full output in order to compare the calculation process. I can also upload my output here.\nThanks in advance,\nEvya.</p>",
  "messages": [
    {
      "id": 595694,
      "postDate": "2019-08-09T14:51:47.033Z",
      "content": "<p>Hi,\nRecently I have started to engage with this field and built a CNN model in order to predict a gender through offline handwriting.\nI would like to evaluate my performance on the QUWI subset, used in the competition. My output contains a list of the probability of a form to be male (actually the softmax outputed by the network) for each QUWI test form. Apparently, my results seem to be very good and i'm suspicious about it.</p>\n\n<p>There is any source to get an output example with the male scores of one of the methods used in the competition?  I wish for an example of full output in order to compare the calculation process. I can also upload my output here.\nThanks in advance,\nEvya.</p>",
      "rawMarkdown": "Hi,\nRecently I have started to engage with this field and built a CNN model in order to predict a gender through offline handwriting.\nI would like to evaluate my performance on the QUWI subset, used in the competition. My output contains a list of the probability of a form to be male (actually the softmax outputed by the network) for each QUWI test form. Apparently, my results seem to be very good and i'm suspicious about it.\n\nThere is any source to get an output example with the male scores of one of the methods used in the competition?  I wish for an example of full output in order to compare the calculation process. I can also upload my output here.\nThanks in advance,\nEvya."
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "595694": "Hi,\nRecently I have started to engage with this field and built a CNN model in order to predict a gender through offline handwriting.\nI would like to evaluate my performance on the QUWI subset, used in the competition. My output contains a list of the probability of a form to be male (actually the softmax outputed by the network) for each QUWI test form. Apparently, my results seem to be very good and i'm suspicious about it.\n\nThere is any source to get an output example with the male scores of one of the methods used in the competition?  I wish for an example of full output in order to compare the calculation process. I can also upload my output here.\nThanks in advance,\nEvya."
  }
}