{
  "id": 411606,
  "title": "Missing ground truth for phototourism dataset",
  "url": "/competitions/image-matching-challenge-2023/discussion/411606",
  "author_name": "",
  "post_date": "2023-05-20T03:54:39.111792200Z",
  "votes": 2,
  "comment_count": 5,
  "views": 0,
  "content": "<p>Hello,<br>\nThe <code>train_labels.csv</code> file does not contain any entries for images from the phototourism dataset. This omission might be an error since it is not straightforward to retrieve this information from the data in the sfm directory. Can any of the organisers please take a look into this issue.</p>",
  "messages": [
    {
      "id": "2266417",
      "postDate": "05/20/2023 03:54:39",
      "content": "<p>Hello,<br>\nThe <code>train_labels.csv</code> file does not contain any entries for images from the phototourism dataset. This omission might be an error since it is not straightforward to retrieve this information from the data in the sfm directory. Can any of the organisers please take a look into this issue.</p>",
      "rawMarkdown": "Hello,\nThe `train_labels.csv` file does not contain any entries for images from the phototourism dataset. This omission might be an error since it is not straightforward to retrieve this information from the data in the sfm directory. Can any of the organisers please take a look into this issue.",
      "votes": null
    },
    {
      "id": "2266468",
      "postDate": "05/20/2023 05:24:49",
      "content": "<p>Phototourism dataset is from IMC2022.I do not know the use of this dataset.</p>",
      "rawMarkdown": "Phototourism dataset is from IMC2022.I do not know the use of this dataset.",
      "votes": null
    },
    {
      "id": "2268167",
      "postDate": "05/21/2023 14:25:45",
      "content": "<p>Yes, Phototourism dataset is not used.</p>",
      "rawMarkdown": "Yes, Phototourism dataset is not used.",
      "votes": null
    },
    {
      "id": "2268991",
      "postDate": "05/22/2023 06:41:03",
      "content": "<p>The data is there, in the sfm subfolders, but it requires a little bit of preprocessing. We expect people will focus on the new datasets, but provide also the one from last year if you want to run more experiments/tuning on the training data.</p>",
      "rawMarkdown": "The data is there, in the sfm subfolders, but it requires a little bit of preprocessing. We expect people will focus on the new datasets, but provide also the one from last year if you want to run more experiments/tuning on the training data.",
      "votes": null
    },
    {
      "id": "2269162",
      "postDate": "05/22/2023 08:50:51",
      "content": "<p>Ah sorry, I missed the <code>NAME</code> field in images.txt. With that it is easy to map provided data to specific image.</p>",
      "rawMarkdown": "Ah sorry, I missed the `NAME` field in images.txt. With that it is easy to map provided data to specific image.",
      "votes": null
    },
    {
      "id": "2269237",
      "postDate": "05/22/2023 10:21:42",
      "content": "<p>Yes. Please note that all images in the dataset have an entry in the images.txt file, but not the other way around (we removed some images, for various reasons).</p>",
      "rawMarkdown": "Yes. Please note that all images in the dataset have an entry in the images.txt file, but not the other way around (we removed some images, for various reasons).",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2266468,
      "author_name": "distiller",
      "author_url": "",
      "post_date": "05/20/2023 05:24:49",
      "content": "<p>Phototourism dataset is from IMC2022.I do not know the use of this dataset.</p>",
      "votes": null,
      "replies": [
        {
          "id": 2268167,
          "author_name": "fabiobellavia",
          "author_url": "",
          "post_date": "05/21/2023 14:25:45",
          "content": "<p>Yes, Phototourism dataset is not used.</p>",
          "votes": null,
          "replies": [
            {
              "id": 2268991,
              "author_name": "eduardtrulls",
              "author_url": "",
              "post_date": "05/22/2023 06:41:03",
              "content": "<p>The data is there, in the sfm subfolders, but it requires a little bit of preprocessing. We expect people will focus on the new datasets, but provide also the one from last year if you want to run more experiments/tuning on the training data.</p>",
              "votes": null,
              "replies": [
                {
                  "id": 2269162,
                  "author_name": "python467",
                  "author_url": "",
                  "post_date": "05/22/2023 08:50:51",
                  "content": "<p>Ah sorry, I missed the <code>NAME</code> field in images.txt. With that it is easy to map provided data to specific image.</p>",
                  "votes": null,
                  "replies": [
                    {
                      "id": 2269237,
                      "author_name": "eduardtrulls",
                      "author_url": "",
                      "post_date": "05/22/2023 10:21:42",
                      "content": "<p>Yes. Please note that all images in the dataset have an entry in the images.txt file, but not the other way around (we removed some images, for various reasons).</p>",
                      "votes": null,
                      "replies": []
                    }
                  ]
                }
              ]
            }
          ]
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "2266417": "Hello,\nThe `train_labels.csv` file does not contain any entries for images from the phototourism dataset. This omission might be an error since it is not straightforward to retrieve this information from the data in the sfm directory. Can any of the organisers please take a look into this issue.",
    "2266468": "Phototourism dataset is from IMC2022.I do not know the use of this dataset.",
    "2268167": "Yes, Phototourism dataset is not used.",
    "2268991": "The data is there, in the sfm subfolders, but it requires a little bit of preprocessing. We expect people will focus on the new datasets, but provide also the one from last year if you want to run more experiments/tuning on the training data.",
    "2269162": "Ah sorry, I missed the `NAME` field in images.txt. With that it is easy to map provided data to specific image.",
    "2269237": "Yes. Please note that all images in the dataset have an entry in the images.txt file, but not the other way around (we removed some images, for various reasons)."
  },
  "source": "meta"
}