{ "id": 606917, "title": "Data Update 2", "url": "/competitions/rsna-intracranial-aneurysm-detection/discussion/606917", "author_name": "Ryan Holbrook", "post_date": "2025-09-10T20:19:50.264000", "votes": 27, "comment_count": 25, "views": 0, "content": "

Hi everyone,

\n

We are posting another data update that fixes a couple of issues. First, there are a few corrections to the localizers in the training set. Second, the 'PerFrameFunctionalGroupsSequence' tag has been restored to the multiframe DICOMs in the hidden test set.

\n

(The training set with the new localizers is still uploading as of this posting, but should be available within an hour or two.)

", "messages": [ { "id": 3287067, "postDate": "2025-09-10T20:19:50.263Z", "content": "

Hi everyone,

\n

We are posting another data update that fixes a couple of issues. First, there are a few corrections to the localizers in the training set. Second, the 'PerFrameFunctionalGroupsSequence' tag has been restored to the multiframe DICOMs in the hidden test set.

\n

(The training set with the new localizers is still uploading as of this posting, but should be available within an hour or two.)

", "rawMarkdown": "Hi everyone,\n\nWe are posting another data update that fixes a couple of issues. First, there are a few corrections to the localizers in the training set. Second, the 'PerFrameFunctionalGroupsSequence' tag [has been restored](https://www.kaggle.com/competitions/rsna-intracranial-aneurysm-detection/discussion/603094) to the multiframe DICOMs in the hidden test set.\n\n(The training set with the new localizers is still uploading as of this posting, but should be available within an hour or two.)", "votes": 27 }, { "id": 3287151, "postDate": "2025-09-11T00:05:21.053Z", "content": "

Hi @ryanholbrook @evancalabrese , Did you only update the train_localizers.csv file, or updated something else, for fixing the localizers?

", "rawMarkdown": "Hi @ryanholbrook @evancalabrese , Did you only update the `train_localizers.csv` file, or updated something else, for fixing the localizers?", "votes": 9 }, { "id": 3287907, "postDate": "2025-09-12T15:18:47.760Z", "content": "

Hello everyone.

\n

The following adjustments have been made to the original 'train.csv' and 'train_localizers.csv' files:

\n\n

We hope this clarifies details for this update and will continue to monitor and review any future issues. Happy Kaggling!

", "rawMarkdown": "Hello everyone.\n\nThe following adjustments have been made to the original 'train.csv' and 'train_localizers.csv' files:\n\n- Aneurysm localizer data location and coordinates have been fixed for a number of series where incorrectly registered.\n\n- Series affected by localizer issues have had their label sets adjusted accordingly in 'train.csv' master file.\n\n- Frame range indexing for **multi-frame** DICOM localizers have been set to begin at 0.\n\n- Certain series with mismatched localizer UID values have been fixed.\n\nWe hope this clarifies details for this update and will continue to monitor and review any future issues. Happy Kaggling!", "votes": 5, "replies": [ { "id": 3289571, "postDate": "2025-09-16T08:51:11.353Z", "content": "

Thank you to the organizer for the effort in providing us with accurate labels. I hope the next update will be the final version. Some of us, including myself, have to rent rather expensive GPUs externally, so I’m quite exhausted from repeatedly preprocessing the data and retraining the models (we have to reproduce our solutions if we want to win). I just want to finish this competition and take a break for a while. Competing with a large dataset (~310GB) is truly tiring.

", "rawMarkdown": "Thank you to the organizer for the effort in providing us with accurate labels. I hope the next update will be the final version. Some of us, including myself, have to rent rather expensive GPUs externally, so I’m quite exhausted from repeatedly preprocessing the data and retraining the models (we have to reproduce our solutions if we want to win). I just want to finish this competition and take a break for a while. Competing with a large dataset (~310GB) is truly tiring.", "votes": 4, "replies": [ { "id": 3289846, "postDate": "2025-09-16T16:03:15.770Z", "content": "

Understandably so, and we are grateful for your persistence in continuing to engage with us and this challenge. We are cognizant of the individual resources required to perform well in the competition, and will ensure our best efforts to make any adjustments efficient to the success of our participants.

", "rawMarkdown": "Understandably so, and we are grateful for your persistence in continuing to engage with us and this challenge. We are cognizant of the individual resources required to perform well in the competition, and will ensure our best efforts to make any adjustments efficient to the success of our participants.", "votes": 2 } ] } ] }, { "id": 3294640, "postDate": "2025-09-26T13:25:35.477Z", "content": "

Is SharedFunctionalGroupsSequence also available for multiframe DICOMs? This allows us to get the pixel spacing and image orientation information.

", "rawMarkdown": "Is `SharedFunctionalGroupsSequence` also available for multiframe DICOMs? This allows us to get the pixel spacing and image orientation information.", "votes": 4, "replies": [ { "id": 3294727, "postDate": "2025-09-26T15:39:34.920Z", "content": "

@ryanholbrook can comment when a pending update to the test set to restore the 'PerFrameFunctionalGroupsSequence' tag is complete. It was originally restored in the second update, but the dataset was reverted when other issues were identified with the update.

", "rawMarkdown": "@ryanholbrook can comment when a pending update to the test set to restore the 'PerFrameFunctionalGroupsSequence' tag is complete. It was originally restored in the second update, but the dataset was reverted when other issues were identified with the update.", "votes": 2, "replies": [ { "id": 3295058, "postDate": "2025-09-27T14:45:31.047Z", "content": "

Ok thanks, SharedFunctionalGroupsSequence is a separate tag that would also be helpful to have.

", "rawMarkdown": "Ok thanks, `SharedFunctionalGroupsSequence` is a separate tag that would also be helpful to have.", "votes": 1 } ] } ] }, { "id": 3289804, "postDate": "2025-09-16T15:06:54.320Z", "content": "

Could you please share a list of the series where there have been changes in either DICOM or NIfTI? That way, I can skip downloading the full dataset and just pull the updated ones and the CSVs.

", "rawMarkdown": "Could you please share a list of the series where there have been changes in either DICOM or NIfTI? That way, I can skip downloading the full dataset and just pull the updated ones and the CSVs.", "votes": 3 }, { "id": 3288090, "postDate": "2025-09-13T04:03:10.460Z", "content": "

can you tell me more about which file you update? you update localizer csv file or any other file?

", "rawMarkdown": "can you tell me more about which file you update? you update localizer csv file or any other file? ", "votes": 4 }, { "id": 3287186, "postDate": "2025-09-11T03:35:19.827Z", "content": "

Hi @ryanholbrook ,

\n

Thank you for keeping the dataset in order.

\n

I noticed the diagnoses in train.csv also changed:

\n
-.....,,Female,MRA,,,,,\n+.....,,Female,MRA,,,,,\n...\n-.....,,Male,CTA,,,,,\n+.....,,Male,CTA,,,,,\n
\n

The first patient has one location less, the second patient's only location was removed.
\nIs the change correct?

\n

Also, the following series have different sets of localized targets, but the targets in train.csv are unchanged:

\n
1.2.826.0.1.3680043.8.498.40499156229797942175102861717242994172\ntrain.csv:  [0100000]\nlocalizers: [0000000] (was left artery, now right)\n\n\n1.2.826.0.1.3680043.8.498.75798029534455454939797323020706657426\ntrain.csv:  [0000000]\nlocalizers: [0000000] (basilar tip no longer localized)\n
\n

Which targets are correct - train.csv or localizers?

", "rawMarkdown": "Hi @ryanholbrook ,\n\nThank you for keeping the dataset in order.\n\nI noticed the diagnoses in train.csv also changed:\n```\n-1.2.826.0.1.3680043.8.498.11292203154407642658894712229998766945,81,Female,MRA,0,0,1,0,0,0,0,0,0,0,1,1,0,1\n+1.2.826.0.1.3680043.8.498.11292203154407642658894712229998766945,81,Female,MRA,0,0,1,1,0,0,0,0,0,0,1,1,0,1\n...\n-1.2.826.0.1.3680043.8.498.74390569791112039529514861261033590424,60,Male,CTA,0,0,1,0,0,0,0,0,0,0,0,0,0,1\n+1.2.826.0.1.3680043.8.498.74390569791112039529514861261033590424,60,Male,CTA,0,0,0,0,0,0,0,0,0,0,0,0,0,0\n```\nThe first patient has one location less, the second patient's only location was removed.\nIs the change correct?\n\nAlso, the following series have different sets of localized targets, but the targets in train.csv are unchanged:\n```\n1.2.826.0.1.3680043.8.498.40499156229797942175102861717242994172\ntrain.csv: [0 0 1 0 0 0 0 0 0 0 0 0 0]\nlocalizers: [0 0 0 1 0 0 0 0 0 0 0 0 0] (was left artery, now right)\n\n\n1.2.826.0.1.3680043.8.498.75798029534455454939797323020706657426\ntrain.csv: [0 0 0 0 0 0 0 0 0 1 0 1 0]\nlocalizers: [0 0 0 0 0 0 0 0 0 1 0 0 0] (basilar tip no longer localized)\n```\nWhich targets are correct - train.csv or localizers?\n", "votes": 4 }, { "id": 3288967, "postDate": "2025-09-15T07:23:12.377Z", "content": "

@shosys

\n
\n

unique_train_series_with_aneurysm = train[train['Aneurysm Present'] == 1].SeriesInstanceUID.unique()\nunique_localization_sequences = train_localization.SeriesInstanceUID.unique()\nset(unique_train_sequences_with_aneurysm) - set(unique_localization_sequences)

\n
\n

results in two series whose Aneurysm Present is true, yet there is no localization information for these two ( checked after data update 2 )

\n
\n

{'1.2.826.0.1.3680043.8.498.12937082136541515013380696257898978214',\n '1.2.826.0.1.3680043.8.498.86840850085811129970747331978337342341'}

\n
\n

Is it an intended aspect of the dataset that certain series with aneurysm present lack corresponding localization data, or should all such series have this information?

", "rawMarkdown": "@shosys \n> `unique_train_series_with_aneurysm = train[train['Aneurysm Present'] == 1].SeriesInstanceUID.unique()\nunique_localization_sequences = train_localization.SeriesInstanceUID.unique()\nset(unique_train_sequences_with_aneurysm) - set(unique_localization_sequences)`\n\nresults in two series whose `Aneurysm Present` is true, yet there is no localization information for these two ( checked after data update 2 )\n\n> `{'1.2.826.0.1.3680043.8.498.12937082136541515013380696257898978214',\n '1.2.826.0.1.3680043.8.498.86840850085811129970747331978337342341'}`\n\nIs it an intended aspect of the dataset that certain series with aneurysm present lack corresponding localization data, or should all such series have this information?", "votes": 1, "replies": [ { "id": 3296277, "postDate": "2025-09-30T14:56:31.933Z", "content": "

Hello,

\n

The Kaggle staff has rectified certain issues affecting the 2nd dataset update, which should resolve this within the localizer CSV.

", "rawMarkdown": "Hello,\n\nThe Kaggle staff has rectified certain issues affecting the 2nd dataset update, which should resolve this within the localizer CSV." } ] }, { "id": 3287370, "postDate": "2025-09-11T13:55:19.667Z", "content": "

does anyone successfully submit? almost all the our subs time-out today.

", "rawMarkdown": "does anyone successfully submit? almost all the our subs time-out today. ", "votes": 1, "replies": [ { "id": 3287408, "postDate": "2025-09-11T14:59:35.773Z", "content": "

I submitted one, and it worked …

", "rawMarkdown": "I submitted one, and it worked ...", "replies": [ { "id": 3288774, "postDate": "2025-09-14T18:18:45.563Z", "content": "

which one?

", "rawMarkdown": "which one?\n" } ] }, { "id": 3287462, "postDate": "2025-09-11T17:18:11.453Z", "content": "

Looks like there were a handful of submissions that might have failed during the transition. I will rerun and post new scores soon.

", "rawMarkdown": "Looks like there were a handful of submissions that might have failed during the transition. I will rerun and post new scores soon.", "votes": 2, "replies": [ { "id": 3287625, "postDate": "2025-09-11T23:24:02.110Z", "content": "

@ryanholbrook I also want ot point out that the submission time becomes much longer, our original submission before update taking 5hr to finish, but needs 11hr after updated.

", "rawMarkdown": "@ryanholbrook I also want ot point out that the submission time becomes much longer, our original submission before update taking 5hr to finish, but needs 11hr after updated. ", "votes": 1 }, { "id": 3287872, "postDate": "2025-09-12T13:36:06.863Z", "content": "

That should have just been temporary. If it persists, please let me know and I'll investigate.

", "rawMarkdown": "That should have just been temporary. If it persists, please let me know and I'll investigate." }, { "id": 3288086, "postDate": "2025-09-13T03:31:13.863Z", "content": "

@ryanholbrook yeah Ryan the problem still exists

", "rawMarkdown": "@ryanholbrook yeah Ryan the problem still exists" }, { "id": 3288584, "postDate": "2025-09-14T10:53:50.407Z", "content": "

Problem solved, thanks!

", "rawMarkdown": "Problem solved, thanks!" } ] } ] }, { "id": 3287110, "postDate": "2025-09-10T22:10:11.127Z", "content": "

Thanks for this update!!

", "rawMarkdown": "Thanks for this update!!" }, { "id": 3289404, "postDate": "2025-09-16T02:44:13.160Z", "rawMarkdown": "", "isDeleted": true }, { "id": 3287164, "postDate": "2025-09-11T01:29:48.143Z", "rawMarkdown": "", "votes": 1, "isDeleted": true }, { "id": 3288996, "postDate": "2025-09-15T08:35:04.643Z", "content": "

Thank you for the update!

", "rawMarkdown": "Thank you for the update!" }, { "id": 3288799, "postDate": "2025-09-14T20:12:12.500Z", "content": "

Thank you for the update!

", "rawMarkdown": "Thank you for the update!" } ], "comments": [ { "id": 3287151, "author_name": "ForcewithMe", "author_url": "", "post_date": "2025-09-11T00:05:21.053000", "content": "

Hi @ryanholbrook @evancalabrese , Did you only update the train_localizers.csv file, or updated something else, for fixing the localizers?

", "votes": 9, "replies": [] }, { "id": 3287907, "author_name": "Jason Sho", "author_url": "", "post_date": "2025-09-12T15:18:47.760000", "content": "

Hello everyone.

\n

The following adjustments have been made to the original 'train.csv' and 'train_localizers.csv' files:

\n\n

We hope this clarifies details for this update and will continue to monitor and review any future issues. Happy Kaggling!

", "votes": 5, "replies": [ { "id": 3289571, "author_name": "HoangHuyen", "author_url": "", "post_date": "2025-09-16T08:51:11.353000", "content": "

Thank you to the organizer for the effort in providing us with accurate labels. I hope the next update will be the final version. Some of us, including myself, have to rent rather expensive GPUs externally, so I’m quite exhausted from repeatedly preprocessing the data and retraining the models (we have to reproduce our solutions if we want to win). I just want to finish this competition and take a break for a while. Competing with a large dataset (~310GB) is truly tiring.

", "votes": 4, "replies": [ { "id": 3289846, "author_name": "Jason Sho", "author_url": "", "post_date": "2025-09-16T16:03:15.770000", "content": "

Understandably so, and we are grateful for your persistence in continuing to engage with us and this challenge. We are cognizant of the individual resources required to perform well in the competition, and will ensure our best efforts to make any adjustments efficient to the success of our participants.

", "votes": 2, "replies": [] } ] } ] }, { "id": 3294640, "author_name": "Ian Pan", "author_url": "", "post_date": "2025-09-26T13:25:35.477000", "content": "

Is SharedFunctionalGroupsSequence also available for multiframe DICOMs? This allows us to get the pixel spacing and image orientation information.

", "votes": 4, "replies": [ { "id": 3294727, "author_name": "JeffRudie", "author_url": "", "post_date": "2025-09-26T15:39:34.920000", "content": "

@ryanholbrook can comment when a pending update to the test set to restore the 'PerFrameFunctionalGroupsSequence' tag is complete. It was originally restored in the second update, but the dataset was reverted when other issues were identified with the update.

", "votes": 2, "replies": [ { "id": 3295058, "author_name": "Ian Pan", "author_url": "", "post_date": "2025-09-27T14:45:31.047000", "content": "

Ok thanks, SharedFunctionalGroupsSequence is a separate tag that would also be helpful to have.

", "votes": 1, "replies": [] } ] } ] }, { "id": 3289804, "author_name": "hongree", "author_url": "", "post_date": "2025-09-16T15:06:54.320000", "content": "

Could you please share a list of the series where there have been changes in either DICOM or NIfTI? That way, I can skip downloading the full dataset and just pull the updated ones and the CSVs.

", "votes": 3, "replies": [] }, { "id": 3288090, "author_name": "Muhammad Hamza", "author_url": "", "post_date": "2025-09-13T04:03:10.460000", "content": "

can you tell me more about which file you update? you update localizer csv file or any other file?

", "votes": 4, "replies": [] }, { "id": 3287186, "author_name": "vsemionov", "author_url": "", "post_date": "2025-09-11T03:35:19.827000", "content": "

Hi @ryanholbrook ,

\n

Thank you for keeping the dataset in order.

\n

I noticed the diagnoses in train.csv also changed:

\n
-.....,,Female,MRA,,,,,\n+.....,,Female,MRA,,,,,\n...\n-.....,,Male,CTA,,,,,\n+.....,,Male,CTA,,,,,\n
\n

The first patient has one location less, the second patient's only location was removed.
\nIs the change correct?

\n

Also, the following series have different sets of localized targets, but the targets in train.csv are unchanged:

\n
1.2.826.0.1.3680043.8.498.40499156229797942175102861717242994172\ntrain.csv:  [0100000]\nlocalizers: [0000000] (was left artery, now right)\n\n\n1.2.826.0.1.3680043.8.498.75798029534455454939797323020706657426\ntrain.csv:  [0000000]\nlocalizers: [0000000] (basilar tip no longer localized)\n
\n

Which targets are correct - train.csv or localizers?

", "votes": 4, "replies": [] }, { "id": 3288967, "author_name": "Amruthsagar1024", "author_url": "", "post_date": "2025-09-15T07:23:12.377000", "content": "

@shosys

\n
\n

unique_train_series_with_aneurysm = train[train['Aneurysm Present'] == 1].SeriesInstanceUID.unique()\nunique_localization_sequences = train_localization.SeriesInstanceUID.unique()\nset(unique_train_sequences_with_aneurysm) - set(unique_localization_sequences)

\n
\n

results in two series whose Aneurysm Present is true, yet there is no localization information for these two ( checked after data update 2 )

\n
\n

{'1.2.826.0.1.3680043.8.498.12937082136541515013380696257898978214',\n '1.2.826.0.1.3680043.8.498.86840850085811129970747331978337342341'}

\n
\n

Is it an intended aspect of the dataset that certain series with aneurysm present lack corresponding localization data, or should all such series have this information?

", "votes": 1, "replies": [ { "id": 3296277, "author_name": "Jason Sho", "author_url": "", "post_date": "2025-09-30T14:56:31.933000", "content": "

Hello,

\n

The Kaggle staff has rectified certain issues affecting the 2nd dataset update, which should resolve this within the localizer CSV.

", "votes": 0, "replies": [] } ] }, { "id": 3287370, "author_name": "Tom", "author_url": "", "post_date": "2025-09-11T13:55:19.667000", "content": "

does anyone successfully submit? almost all the our subs time-out today.

", "votes": 1, "replies": [ { "id": 3287408, "author_name": "Arunodhayan", "author_url": "", "post_date": "2025-09-11T14:59:35.773000", "content": "

I submitted one, and it worked …

", "votes": 0, "replies": [ { "id": 3288774, "author_name": "Muhammad Hamza", "author_url": "", "post_date": "2025-09-14T18:18:45.563000", "content": "

which one?

", "votes": 0, "replies": [] } ] }, { "id": 3287462, "author_name": "Ryan Holbrook", "author_url": "", "post_date": "2025-09-11T17:18:11.453000", "content": "

Looks like there were a handful of submissions that might have failed during the transition. I will rerun and post new scores soon.

", "votes": 2, "replies": [ { "id": 3287625, "author_name": "Tom", "author_url": "", "post_date": "2025-09-11T23:24:02.110000", "content": "

@ryanholbrook I also want ot point out that the submission time becomes much longer, our original submission before update taking 5hr to finish, but needs 11hr after updated.

", "votes": 1, "replies": [] }, { "id": 3287872, "author_name": "Ryan Holbrook", "author_url": "", "post_date": "2025-09-12T13:36:06.863000", "content": "

That should have just been temporary. If it persists, please let me know and I'll investigate.

", "votes": 0, "replies": [] }, { "id": 3288086, "author_name": "Tom", "author_url": "", "post_date": "2025-09-13T03:31:13.863000", "content": "

@ryanholbrook yeah Ryan the problem still exists

", "votes": 0, "replies": [] }, { "id": 3288584, "author_name": "Tom", "author_url": "", "post_date": "2025-09-14T10:53:50.407000", "content": "

Problem solved, thanks!

", "votes": 0, "replies": [] } ] } ] }, { "id": 3287110, "author_name": "Evan Calabrese", "author_url": "", "post_date": "2025-09-10T22:10:11.127000", "content": "

Thanks for this update!!

", "votes": 0, "replies": [] }, { "id": 3289404, "author_name": "", "author_url": "", "post_date": "2025-09-16T02:44:13.160000", "content": "", "votes": 0, "replies": [] }, { "id": 3287164, "author_name": "", "author_url": "", "post_date": "2025-09-11T01:29:48.143000", "content": "", "votes": 1, "replies": [] }, { "id": 3288996, "author_name": "Khushi Yadav", "author_url": "", "post_date": "2025-09-15T08:35:04.643000", "content": "

Thank you for the update!

", "votes": 0, "replies": [] }, { "id": 3288799, "author_name": "ABDELHAK OUANZOUGUI", "author_url": "", "post_date": "2025-09-14T20:12:12.500000", "content": "

Thank you for the update!

", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "3287067": "Hi everyone,\n\nWe are posting another data update that fixes a couple of issues. First, there are a few corrections to the localizers in the training set. Second, the 'PerFrameFunctionalGroupsSequence' tag [has been restored](https://www.kaggle.com/competitions/rsna-intracranial-aneurysm-detection/discussion/603094) to the multiframe DICOMs in the hidden test set.\n\n(The training set with the new localizers is still uploading as of this posting, but should be available within an hour or two.)", "3287151": "Hi @ryanholbrook @evancalabrese , Did you only update the `train_localizers.csv` file, or updated something else, for fixing the localizers?", "3287907": "Hello everyone.\n\nThe following adjustments have been made to the original 'train.csv' and 'train_localizers.csv' files:\n\n- Aneurysm localizer data location and coordinates have been fixed for a number of series where incorrectly registered.\n\n- Series affected by localizer issues have had their label sets adjusted accordingly in 'train.csv' master file.\n\n- Frame range indexing for **multi-frame** DICOM localizers have been set to begin at 0.\n\n- Certain series with mismatched localizer UID values have been fixed.\n\nWe hope this clarifies details for this update and will continue to monitor and review any future issues. Happy Kaggling!", "3294640": "Is `SharedFunctionalGroupsSequence` also available for multiframe DICOMs? This allows us to get the pixel spacing and image orientation information.", "3289804": "Could you please share a list of the series where there have been changes in either DICOM or NIfTI? That way, I can skip downloading the full dataset and just pull the updated ones and the CSVs.", "3288090": "can you tell me more about which file you update? you update localizer csv file or any other file? ", "3287186": "Hi @ryanholbrook ,\n\nThank you for keeping the dataset in order.\n\nI noticed the diagnoses in train.csv also changed:\n```\n-1.2.826.0.1.3680043.8.498.11292203154407642658894712229998766945,81,Female,MRA,0,0,1,0,0,0,0,0,0,0,1,1,0,1\n+1.2.826.0.1.3680043.8.498.11292203154407642658894712229998766945,81,Female,MRA,0,0,1,1,0,0,0,0,0,0,1,1,0,1\n...\n-1.2.826.0.1.3680043.8.498.74390569791112039529514861261033590424,60,Male,CTA,0,0,1,0,0,0,0,0,0,0,0,0,0,1\n+1.2.826.0.1.3680043.8.498.74390569791112039529514861261033590424,60,Male,CTA,0,0,0,0,0,0,0,0,0,0,0,0,0,0\n```\nThe first patient has one location less, the second patient's only location was removed.\nIs the change correct?\n\nAlso, the following series have different sets of localized targets, but the targets in train.csv are unchanged:\n```\n1.2.826.0.1.3680043.8.498.40499156229797942175102861717242994172\ntrain.csv: [0 0 1 0 0 0 0 0 0 0 0 0 0]\nlocalizers: [0 0 0 1 0 0 0 0 0 0 0 0 0] (was left artery, now right)\n\n\n1.2.826.0.1.3680043.8.498.75798029534455454939797323020706657426\ntrain.csv: [0 0 0 0 0 0 0 0 0 1 0 1 0]\nlocalizers: [0 0 0 0 0 0 0 0 0 1 0 0 0] (basilar tip no longer localized)\n```\nWhich targets are correct - train.csv or localizers?\n", "3288967": "@shosys \n> `unique_train_series_with_aneurysm = train[train['Aneurysm Present'] == 1].SeriesInstanceUID.unique()\nunique_localization_sequences = train_localization.SeriesInstanceUID.unique()\nset(unique_train_sequences_with_aneurysm) - set(unique_localization_sequences)`\n\nresults in two series whose `Aneurysm Present` is true, yet there is no localization information for these two ( checked after data update 2 )\n\n> `{'1.2.826.0.1.3680043.8.498.12937082136541515013380696257898978214',\n '1.2.826.0.1.3680043.8.498.86840850085811129970747331978337342341'}`\n\nIs it an intended aspect of the dataset that certain series with aneurysm present lack corresponding localization data, or should all such series have this information?", "3287370": "does anyone successfully submit? almost all the our subs time-out today. ", "3287110": "Thanks for this update!!", "3289404": "", "3287164": "", "3288996": "Thank you for the update!", "3288799": "Thank you for the update!" } }