Per the Competition Rules, External Data is permitted in this competition, but must be posted to this forum thread no later than the Entry Deadline (one week before competition close).
\n\nOnce someone posts an external dataset to this thread, you do not need to re-post it if you are using the same one.
", "messages": [ { "id": 628916, "postDate": "2019-09-18T05:43:52.870Z", "content": "Per the Competition Rules, External Data is permitted in this competition, but must be posted to this forum thread no later than the Entry Deadline (one week before competition close).
\n\nOnce someone posts an external dataset to this thread, you do not need to re-post it if you are using the same one.
", "rawMarkdown": "Per the [Competition Rules](https://www.kaggle.com/c/rsna-intracranial-hemorrhage-detection/rules), External Data is permitted in this competition, but must be posted to this forum thread no later than the Entry Deadline (one week before competition close).\n\nOnce someone posts an external dataset to this thread, you do not need to re-post it if you are using the same one.", "votes": 12 }, { "id": 630583, "postDate": "2019-09-20T12:29:22.553Z", "content": "I will possibly use this public dataset:\nhttp://headctstudy.qure.ai/dataset
", "rawMarkdown": "I will possibly use this public dataset:\nhttp://headctstudy.qure.ai/dataset", "votes": 8, "replies": [ { "id": 649282, "postDate": "2019-10-15T07:10:54.183Z", "content": "I don't think the licence allows using this dataset for a kaggle competition, if you win any kind of compensation for a model that is based on the dataset, then it might become illegal. see https://creativecommons.org/licenses/by-nc-sa/4.0/
", "rawMarkdown": "I don't think the licence allows using this dataset for a kaggle competition, if you win any kind of compensation for a model that is based on the dataset, then it might become illegal. see https://creativecommons.org/licenses/by-nc-sa/4.0/", "votes": 2 }, { "id": 650140, "postDate": "2019-10-16T06:26:01.797Z", "content": "That's a good point. @juliaelliott could you please clarify this one?
", "rawMarkdown": "That's a good point. @juliaelliott could you please clarify this one?", "votes": 1 }, { "id": 652391, "postDate": "2019-10-18T19:50:24.590Z", "content": "Image-level (Bbx) annotations for qure.ai CQ500 dataset:\nhttps://www.kaggle.com/c/rsna-intracranial-hemorrhage-detection/discussion/113339#latest-652389
\n\nhttps://public.md.ai/annotator/project/Y2qr6vqv
", "rawMarkdown": "Image-level (Bbx) annotations for qure.ai CQ500 dataset:\nhttps://www.kaggle.com/c/rsna-intracranial-hemorrhage-detection/discussion/113339#latest-652389\n\nhttps://public.md.ai/annotator/project/Y2qr6vqv" }, { "id": 652500, "postDate": "2019-10-18T23:52:13.393Z", "content": "We are excited to see the progress that the Kaggle teams are making! Eduardo Reis, MD and his group from Hospital Israelita Albert Einstein, São Paulo, BR, have annotated the qure.ai CQ500 dataset with bounding boxes for the different types of hemorrhage. They annotated the thick sliced series within each exam and extrapolated the boxes to the thinner sliced series to expand the available data. They are in the process of writing up their findings for publication. The dataset is made available to the Kaggle community and can be viewed on the MD.ai platform: https://public.md.ai/annotator/project/Y2qr6vqv Eduardo's annotations are on labelgroup 4 - BrainHemX and the bounding boxes can be downloaded using the link above. The original images are hosted by qure.ai at http://headctstudy.qure.ai/dataset licensed under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License
", "rawMarkdown": "We are excited to see the progress that the Kaggle teams are making! Eduardo Reis, MD and his group from Hospital Israelita Albert Einstein, São Paulo, BR, have annotated the qure.ai CQ500 dataset with bounding boxes for the different types of hemorrhage. They annotated the thick sliced series within each exam and extrapolated the boxes to the thinner sliced series to expand the available data. They are in the process of writing up their findings for publication. The dataset is made available to the Kaggle community and can be viewed on the MD.ai platform: https://public.md.ai/annotator/project/Y2qr6vqv Eduardo's annotations are on labelgroup 4 - BrainHemX and the bounding boxes can be downloaded using the link above. The original images are hosted by qure.ai at http://headctstudy.qure.ai/dataset licensed under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License", "votes": 3 }, { "id": 652678, "postDate": "2019-10-19T08:11:42.603Z", "content": "That's nice that you duplicate your post here. But could you actually address the license question? :)
", "rawMarkdown": "That's nice that you duplicate your post here. But could you actually address the license question? :)", "votes": 2 } ] }, { "id": 665117, "postDate": "2019-11-04T16:39:52.697Z", "content": "Publicly available dataset:\nComputed Tomography Images for Intracranial Hemorrhage Detection and Segmentation\nhttps://physionet.org/content/ct-ich/1.0.0/\nhttps://github.com/Murtadha44/-Intracranial-Hemorrhage-Segmentation-Using-Deep-Convolutional-Model-U-Net-\nThe dataset will be updated soon with CT scans in NIfTI format.
", "rawMarkdown": "Publicly available dataset:\n[Computed Tomography Images for Intracranial Hemorrhage Detection and Segmentation](https://physionet.org/content/ct-ich/1.0.0/)\nhttps://physionet.org/content/ct-ich/1.0.0/\nhttps://github.com/Murtadha44/-Intracranial-Hemorrhage-Segmentation-Using-Deep-Convolutional-Model-U-Net-\nThe dataset will be updated soon with CT scans in NIfTI format.", "votes": 1 }, { "id": 660021, "postDate": "2019-10-28T15:07:41.117Z", "content": "https://github.com/Cadene/pretrained-models.pytorch
", "rawMarkdown": "https://github.com/Cadene/pretrained-models.pytorch", "votes": 1 }, { "id": 660735, "postDate": "2019-10-29T14:28:09.823Z", "content": "https://github.com/osmr/imgclsmob
", "rawMarkdown": "https://github.com/osmr/imgclsmob", "votes": 2, "replies": [ { "id": 661222, "postDate": "2019-10-30T03:14:00.313Z", "content": "Wow how I had never heard of this one before! :)
", "rawMarkdown": "Wow how I had never heard of this one before! :) " } ] }, { "id": 658985, "postDate": "2019-10-26T21:52:05.833Z", "content": "Pretrained models from:
\n\n\n\nData / models from:
\n\nApologies - the data at https://mrbrains18.isi.uu.nl/ actually can not be used for the competition, since the ToS for that data say they can only be used for the MRBrainS18 competition, and not for any other purpose.
", "rawMarkdown": "Apologies - the data at https://mrbrains18.isi.uu.nl/ actually can not be used for the competition, since the ToS for that data say they can only be used for the MRBrainS18 competition, and not for any other purpose." } ] }, { "id": 654011, "postDate": "2019-10-21T09:50:12.730Z", "content": "https://github.com/MGH-LMIC/graynet_keras
", "rawMarkdown": "https://github.com/MGH-LMIC/graynet_keras", "votes": 2 }, { "id": 659417, "postDate": "2019-10-27T16:19:57.207Z", "content": "Using Appian's open source code:\nhttps://github.com/appian42/kaggle-rsna-intracranial-hemorrhage
", "rawMarkdown": "Using Appian's open source code:\nhttps://github.com/appian42/kaggle-rsna-intracranial-hemorrhage" }, { "id": 631802, "postDate": "2019-09-22T17:51:52.390Z", "content": "https://www.kaggle.com/felipekitamura/head-ct-hemorrhage has 100 healthy and 100 pathological brain CTs, in PNG format.
\n\nDo we have any leads on healthy brain CTs in dicom format?
", "rawMarkdown": "https://www.kaggle.com/felipekitamura/head-ct-hemorrhage has 100 healthy and 100 pathological brain CTs, in PNG format.\n\nDo we have any leads on healthy brain CTs in dicom format?", "votes": 2 }, { "id": 665143, "postDate": "2019-11-04T17:13:07.123Z", "content": "InceptionResNetv2\nEfficientnets\nResnext\nseresnext\ninceptionv3\nresnet34, 50, 18
\n\nall pretrained on Imagenet
", "rawMarkdown": "InceptionResNetv2\nEfficientnets\nResnext\nseresnext\ninceptionv3\nresnet34, 50, 18\n\nall pretrained on Imagenet" }, { "id": 664287, "postDate": "2019-11-03T13:02:24.867Z", "content": "Using EfficientNet-B3 from:\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nInstall by: pip install efficientnet_pytorch
", "rawMarkdown": "Using EfficientNet-B3 from:\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nInstall by: pip install efficientnet_pytorch" }, { "id": 662002, "postDate": "2019-10-31T01:34:51.420Z", "content": "Using Appian's open source code:\nhttps://github.com/appian42/kaggle-rsna-intracranial-hemorrhage
", "rawMarkdown": "Using Appian's open source code:\nhttps://github.com/appian42/kaggle-rsna-intracranial-hemorrhage" }, { "id": 659722, "postDate": "2019-10-28T06:58:30.557Z", "content": "Pretrained models from: https://github.com/facebookresearch/semi-supervised-ImageNet1K-models
", "rawMarkdown": "Pretrained models from: https://github.com/facebookresearch/semi-supervised-ImageNet1K-models\n" }, { "id": 659581, "postDate": "2019-10-27T23:57:06.793Z", "content": "https://github.com/qubvel/efficientnet\nhttps://github.com/qubvel/classification_models
", "rawMarkdown": "https://github.com/qubvel/efficientnet\nhttps://github.com/qubvel/classification_models\n" }, { "id": 658870, "postDate": "2019-10-26T16:56:13.270Z", "content": "Pretrained model from: https://github.com/Tencent/MedicalNet
", "rawMarkdown": "Pretrained model from: https://github.com/Tencent/MedicalNet" }, { "id": 658005, "postDate": "2019-10-25T15:24:03.097Z", "content": "Might use pretrained models from: \nhttps://github.com/Cadene/pretrained-models.pytorch\nhttps://github.com/facebookresearch/WSL-Images\nhttps://github.com/facebookresearch/FixRes\nhttps://pytorch.org/hub/pytorch_vision_resnext/
", "rawMarkdown": "Might use pretrained models from: \nhttps://github.com/Cadene/pretrained-models.pytorch\nhttps://github.com/facebookresearch/WSL-Images\nhttps://github.com/facebookresearch/FixRes\nhttps://pytorch.org/hub/pytorch_vision_resnext/" }, { "id": 657753, "postDate": "2019-10-25T12:07:13.683Z", "content": "Hi, I may possibly use the pretrained models from https://github.com/kenshohara/3D-ResNets-PyTorch
", "rawMarkdown": "Hi, I may possibly use the pretrained models from https://github.com/kenshohara/3D-ResNets-PyTorch" }, { "id": 656995, "postDate": "2019-10-24T20:32:27.680Z", "content": "https://pytorch.org/hub/research-models\nhttps://github.com/lukemelas/EfficientNet-PyTorch
", "rawMarkdown": "https://pytorch.org/hub/research-models\nhttps://github.com/lukemelas/EfficientNet-PyTorch" }, { "id": 656204, "postDate": "2019-10-24T02:30:53.140Z", "content": "keras efficientnet
\n\nhttps://github.com/titu1994/keras-efficientnets
", "rawMarkdown": "keras efficientnet\n\nhttps://github.com/titu1994/keras-efficientnets" }, { "id": 655441, "postDate": "2019-10-23T03:55:15.380Z", "content": "http://headctstudy.qure.ai\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/rwightman/pytorch-image-models/\n Data from: https://mrbrains18.isi.uu.nl/
", "rawMarkdown": "http://headctstudy.qure.ai\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/rwightman/pytorch-image-models/\n Data from: https://mrbrains18.isi.uu.nl/" }, { "id": 654760, "postDate": "2019-10-22T09:20:48.330Z", "content": "@anoukstein Thanks, could you tell me how to match your annotations with images?
", "rawMarkdown": "@anoukstein Thanks, could you tell me how to match your annotations with images?", "replies": [ { "id": 655003, "postDate": "2019-10-22T15:25:52.203Z", "content": "Hi @qiuzhongwei The annotations on qure.ai by Eduardo Reis et al are associated with the Study, Series, and SOP Instance UIDs which can be read from the Dicom.
", "rawMarkdown": "Hi @qiuzhongwei The annotations on qure.ai by Eduardo Reis et al are associated with the Study, Series, and SOP Instance UIDs which can be read from the Dicom." } ] }, { "id": 654694, "postDate": "2019-10-22T07:27:22.727Z", "content": "Pretrained models from: https://github.com/creafz/pytorch-cnn-finetune
", "rawMarkdown": "Pretrained models from: https://github.com/creafz/pytorch-cnn-finetune" }, { "id": 652018, "postDate": "2019-10-18T08:27:24.893Z", "content": "https://github.com/lukemelas/EfficientNet-PyTorch
", "rawMarkdown": "https://github.com/lukemelas/EfficientNet-PyTorch" }, { "id": 651724, "postDate": "2019-10-17T21:10:44.267Z", "content": "fastai_dev pretrainedmodels\nfastai pretrained model\npretrained-models.pytorch
", "rawMarkdown": "fastai_dev pretrainedmodels\nfastai pretrained model\npretrained-models.pytorch" }, { "id": 648657, "postDate": "2019-10-14T13:06:07.127Z", "content": "Keras pertained models with imagnet weights:\nXception\nVGG16\nVGG19\nResNet, ResNetV2\nInceptionV3\nInceptionResNetV2\nDenseNet\nNASNet\nhttps://keras.io/applications/
", "rawMarkdown": "Keras pertained models with imagnet weights:\nXception\nVGG16\nVGG19\nResNet, ResNetV2\nInceptionV3\nInceptionResNetV2\nDenseNet\nNASNet\nhttps://keras.io/applications/" }, { "id": 648565, "postDate": "2019-10-14T11:15:58.347Z", "content": "Pretrained models from:\n1. torchvision.models\n2. https://github.com/osmr/imgclsmob/tree/master/pytorch/pytorchcv\n3. https://github.com/facebookresearch/WSL-Images
", "rawMarkdown": "Pretrained models from:\n1. torchvision.models\n2. https://github.com/osmr/imgclsmob/tree/master/pytorch/pytorchcv\n3. https://github.com/facebookresearch/WSL-Images" }, { "id": 637002, "postDate": "2019-09-30T13:55:31.957Z", "content": "Pretrained model from https://github.com/rgeirhos/texture-vs-shape
", "rawMarkdown": "Pretrained model from https://github.com/rgeirhos/texture-vs-shape" }, { "id": 636991, "postDate": "2019-09-30T13:39:56.007Z", "content": "Are we supposed to share external datasets only or pretrained models as well?
", "rawMarkdown": "Are we supposed to share external datasets only or pretrained models as well?" }, { "id": 630982, "postDate": "2019-09-21T05:41:28.007Z", "content": "Any of the pretrained models and model architectures from any of the following locations:
\n\nfastai_dev pretrainedmodels\nfastai pretrained model\npretrained-models.pytorch
", "rawMarkdown": "Any of the pretrained models and model architectures from any of the following locations:\n\n[fastai_dev pretrainedmodels](https://github.com/fastai/fastai_dev)\n[fastai pretrained model](https://github.com/fastai/fastai)\n[pretrained-models.pytorch](https://github.com/Cadene/pretrained-models.pytorch)" }, { "id": 650113, "postDate": "2019-10-16T05:45:02.613Z", "rawMarkdown": "", "isDeleted": true } ], "comments": [ { "id": 630583, "author_name": "Alexandre Cadrin-Chênevert", "author_url": "", "post_date": "2019-09-20T12:29:22.553000", "content": "I will possibly use this public dataset:\nhttp://headctstudy.qure.ai/dataset
", "votes": 8, "replies": [ { "id": 649282, "author_name": "pantoine", "author_url": "", "post_date": "2019-10-15T07:10:54.183000", "content": "I don't think the licence allows using this dataset for a kaggle competition, if you win any kind of compensation for a model that is based on the dataset, then it might become illegal. see https://creativecommons.org/licenses/by-nc-sa/4.0/
", "votes": 2, "replies": [] }, { "id": 650140, "author_name": "Dmytro Panchenko", "author_url": "", "post_date": "2019-10-16T06:26:01.797000", "content": "That's a good point. @juliaelliott could you please clarify this one?
", "votes": 1, "replies": [] }, { "id": 652391, "author_name": "Eduardo Pontes Reis", "author_url": "", "post_date": "2019-10-18T19:50:24.590000", "content": "Image-level (Bbx) annotations for qure.ai CQ500 dataset:\nhttps://www.kaggle.com/c/rsna-intracranial-hemorrhage-detection/discussion/113339#latest-652389
\n\nhttps://public.md.ai/annotator/project/Y2qr6vqv
", "votes": 0, "replies": [] }, { "id": 652500, "author_name": "Anouk Stein, MD", "author_url": "", "post_date": "2019-10-18T23:52:13.393000", "content": "We are excited to see the progress that the Kaggle teams are making! Eduardo Reis, MD and his group from Hospital Israelita Albert Einstein, São Paulo, BR, have annotated the qure.ai CQ500 dataset with bounding boxes for the different types of hemorrhage. They annotated the thick sliced series within each exam and extrapolated the boxes to the thinner sliced series to expand the available data. They are in the process of writing up their findings for publication. The dataset is made available to the Kaggle community and can be viewed on the MD.ai platform: https://public.md.ai/annotator/project/Y2qr6vqv Eduardo's annotations are on labelgroup 4 - BrainHemX and the bounding boxes can be downloaded using the link above. The original images are hosted by qure.ai at http://headctstudy.qure.ai/dataset licensed under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License
", "votes": 3, "replies": [] }, { "id": 652678, "author_name": "Dmytro Panchenko", "author_url": "", "post_date": "2019-10-19T08:11:42.603000", "content": "That's nice that you duplicate your post here. But could you actually address the license question? :)
", "votes": 2, "replies": [] } ] }, { "id": 665117, "author_name": "Murtadha Hssayeni", "author_url": "", "post_date": "2019-11-04T16:39:52.697000", "content": "Publicly available dataset:\nComputed Tomography Images for Intracranial Hemorrhage Detection and Segmentation\nhttps://physionet.org/content/ct-ich/1.0.0/\nhttps://github.com/Murtadha44/-Intracranial-Hemorrhage-Segmentation-Using-Deep-Convolutional-Model-U-Net-\nThe dataset will be updated soon with CT scans in NIfTI format.
", "votes": 1, "replies": [] }, { "id": 660021, "author_name": "Appian", "author_url": "", "post_date": "2019-10-28T15:07:41.117000", "content": "https://github.com/Cadene/pretrained-models.pytorch
", "votes": 1, "replies": [] }, { "id": 660735, "author_name": "hengck23", "author_url": "", "post_date": "2019-10-29T14:28:09.823000", "content": "https://github.com/osmr/imgclsmob
", "votes": 2, "replies": [ { "id": 661222, "author_name": "Jeremy Howard", "author_url": "", "post_date": "2019-10-30T03:14:00.313000", "content": "Wow how I had never heard of this one before! :)
", "votes": 0, "replies": [] } ] }, { "id": 658985, "author_name": "Jeremy Howard", "author_url": "", "post_date": "2019-10-26T21:52:05.833000", "content": "Pretrained models from:
\n\n\n\nData / models from:
\n\nApologies - the data at https://mrbrains18.isi.uu.nl/ actually can not be used for the competition, since the ToS for that data say they can only be used for the MRBrainS18 competition, and not for any other purpose.
", "votes": 0, "replies": [] } ] }, { "id": 654011, "author_name": "Neuron Engineer", "author_url": "", "post_date": "2019-10-21T09:50:12.730000", "content": "https://github.com/MGH-LMIC/graynet_keras
", "votes": 2, "replies": [] }, { "id": 659417, "author_name": "Yiheng Wang", "author_url": "", "post_date": "2019-10-27T16:19:57.207000", "content": "Using Appian's open source code:\nhttps://github.com/appian42/kaggle-rsna-intracranial-hemorrhage
", "votes": 0, "replies": [] }, { "id": 631802, "author_name": "Tom H.", "author_url": "", "post_date": "2019-09-22T17:51:52.390000", "content": "https://www.kaggle.com/felipekitamura/head-ct-hemorrhage has 100 healthy and 100 pathological brain CTs, in PNG format.
\n\nDo we have any leads on healthy brain CTs in dicom format?
", "votes": 2, "replies": [] }, { "id": 665143, "author_name": "CE Kan", "author_url": "", "post_date": "2019-11-04T17:13:07.123000", "content": "InceptionResNetv2\nEfficientnets\nResnext\nseresnext\ninceptionv3\nresnet34, 50, 18
\n\nall pretrained on Imagenet
", "votes": 0, "replies": [] }, { "id": 664287, "author_name": "Nic Ma", "author_url": "", "post_date": "2019-11-03T13:02:24.867000", "content": "Using EfficientNet-B3 from:\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nInstall by: pip install efficientnet_pytorch
", "votes": 0, "replies": [] }, { "id": 662002, "author_name": "Nic Ma", "author_url": "", "post_date": "2019-10-31T01:34:51.420000", "content": "Using Appian's open source code:\nhttps://github.com/appian42/kaggle-rsna-intracranial-hemorrhage
", "votes": 0, "replies": [] }, { "id": 659722, "author_name": "Mindy X", "author_url": "", "post_date": "2019-10-28T06:58:30.557000", "content": "Pretrained models from: https://github.com/facebookresearch/semi-supervised-ImageNet1K-models
", "votes": 0, "replies": [] }, { "id": 659581, "author_name": "Sems Kurtoglu", "author_url": "", "post_date": "2019-10-27T23:57:06.793000", "content": "https://github.com/qubvel/efficientnet\nhttps://github.com/qubvel/classification_models
", "votes": 0, "replies": [] }, { "id": 658870, "author_name": "Tomasz Gilewicz", "author_url": "", "post_date": "2019-10-26T16:56:13.270000", "content": "Pretrained model from: https://github.com/Tencent/MedicalNet
", "votes": 0, "replies": [] }, { "id": 658005, "author_name": "Daniel Souza", "author_url": "", "post_date": "2019-10-25T15:24:03.097000", "content": "Might use pretrained models from: \nhttps://github.com/Cadene/pretrained-models.pytorch\nhttps://github.com/facebookresearch/WSL-Images\nhttps://github.com/facebookresearch/FixRes\nhttps://pytorch.org/hub/pytorch_vision_resnext/
", "votes": 0, "replies": [] }, { "id": 657753, "author_name": "Hadar", "author_url": "", "post_date": "2019-10-25T12:07:13.683000", "content": "Hi, I may possibly use the pretrained models from https://github.com/kenshohara/3D-ResNets-PyTorch
", "votes": 0, "replies": [] }, { "id": 656995, "author_name": "Felipe Loque", "author_url": "", "post_date": "2019-10-24T20:32:27.680000", "content": "https://pytorch.org/hub/research-models\nhttps://github.com/lukemelas/EfficientNet-PyTorch
", "votes": 0, "replies": [] }, { "id": 656204, "author_name": "outrunner", "author_url": "", "post_date": "2019-10-24T02:30:53.140000", "content": "keras efficientnet
\n\nhttps://github.com/titu1994/keras-efficientnets
", "votes": 0, "replies": [] }, { "id": 655441, "author_name": "Hilal Shaath", "author_url": "", "post_date": "2019-10-23T03:55:15.380000", "content": "http://headctstudy.qure.ai\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/rwightman/pytorch-image-models/\n Data from: https://mrbrains18.isi.uu.nl/
", "votes": 0, "replies": [] }, { "id": 654760, "author_name": "qzw", "author_url": "", "post_date": "2019-10-22T09:20:48.330000", "content": "@anoukstein Thanks, could you tell me how to match your annotations with images?
", "votes": 0, "replies": [ { "id": 655003, "author_name": "Anouk Stein, MD", "author_url": "", "post_date": "2019-10-22T15:25:52.203000", "content": "Hi @qiuzhongwei The annotations on qure.ai by Eduardo Reis et al are associated with the Study, Series, and SOP Instance UIDs which can be read from the Dicom.
", "votes": 0, "replies": [] } ] }, { "id": 654694, "author_name": "Nic Ma", "author_url": "", "post_date": "2019-10-22T07:27:22.727000", "content": "Pretrained models from: https://github.com/creafz/pytorch-cnn-finetune
", "votes": 0, "replies": [] }, { "id": 652018, "author_name": "Dmytro Panchenko", "author_url": "", "post_date": "2019-10-18T08:27:24.893000", "content": "https://github.com/lukemelas/EfficientNet-PyTorch
", "votes": 0, "replies": [] }, { "id": 651724, "author_name": "Hilal Shaath", "author_url": "", "post_date": "2019-10-17T21:10:44.267000", "content": "fastai_dev pretrainedmodels\nfastai pretrained model\npretrained-models.pytorch
", "votes": 0, "replies": [] }, { "id": 648657, "author_name": "Josh Myers", "author_url": "", "post_date": "2019-10-14T13:06:07.127000", "content": "Keras pertained models with imagnet weights:\nXception\nVGG16\nVGG19\nResNet, ResNetV2\nInceptionV3\nInceptionResNetV2\nDenseNet\nNASNet\nhttps://keras.io/applications/
", "votes": 0, "replies": [] }, { "id": 648565, "author_name": "yuval reina", "author_url": "", "post_date": "2019-10-14T11:15:58.347000", "content": "Pretrained models from:\n1. torchvision.models\n2. https://github.com/osmr/imgclsmob/tree/master/pytorch/pytorchcv\n3. https://github.com/facebookresearch/WSL-Images
", "votes": 0, "replies": [] }, { "id": 637002, "author_name": "Ruslan Baynazarov", "author_url": "", "post_date": "2019-09-30T13:55:31.957000", "content": "Pretrained model from https://github.com/rgeirhos/texture-vs-shape
", "votes": 0, "replies": [] }, { "id": 636991, "author_name": "Gurgel", "author_url": "", "post_date": "2019-09-30T13:39:56.007000", "content": "Are we supposed to share external datasets only or pretrained models as well?
", "votes": 0, "replies": [] }, { "id": 630982, "author_name": "Radek Osmulski", "author_url": "", "post_date": "2019-09-21T05:41:28.007000", "content": "Any of the pretrained models and model architectures from any of the following locations:
\n\nfastai_dev pretrainedmodels\nfastai pretrained model\npretrained-models.pytorch
", "votes": 0, "replies": [] }, { "id": 650113, "author_name": "", "author_url": "", "post_date": "2019-10-16T05:45:02.613000", "content": "", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "628916": "Per the [Competition Rules](https://www.kaggle.com/c/rsna-intracranial-hemorrhage-detection/rules), External Data is permitted in this competition, but must be posted to this forum thread no later than the Entry Deadline (one week before competition close).\n\nOnce someone posts an external dataset to this thread, you do not need to re-post it if you are using the same one.", "630583": "I will possibly use this public dataset:\nhttp://headctstudy.qure.ai/dataset", "665117": "Publicly available dataset:\n[Computed Tomography Images for Intracranial Hemorrhage Detection and Segmentation](https://physionet.org/content/ct-ich/1.0.0/)\nhttps://physionet.org/content/ct-ich/1.0.0/\nhttps://github.com/Murtadha44/-Intracranial-Hemorrhage-Segmentation-Using-Deep-Convolutional-Model-U-Net-\nThe dataset will be updated soon with CT scans in NIfTI format.", "660021": "https://github.com/Cadene/pretrained-models.pytorch", "660735": "https://github.com/osmr/imgclsmob", "658985": "Pretrained models from:\n\n- https://github.com/rwightman/pytorch-image-models/\n- https://github.com/lessw2020/res2net-plus\n\nData / models from:\n\n- https://github.com/fastai/imagenette\n- ~~mrbrains18.isi.uu.nl~~ (*correction: not used, see below*)\n- https://www.nih.gov/news-events/news-releases/nih-clinical-center-releases-dataset-32000-ct-images\n- https://github.com/rsummers11/CADLab/tree/master/LesaNet\n- http://medicaldecathlon.com/", "654011": "https://github.com/MGH-LMIC/graynet_keras", "659417": "Using Appian's open source code:\nhttps://github.com/appian42/kaggle-rsna-intracranial-hemorrhage", "631802": "https://www.kaggle.com/felipekitamura/head-ct-hemorrhage has 100 healthy and 100 pathological brain CTs, in PNG format.\n\nDo we have any leads on healthy brain CTs in dicom format?", "665143": "InceptionResNetv2\nEfficientnets\nResnext\nseresnext\ninceptionv3\nresnet34, 50, 18\n\nall pretrained on Imagenet", "664287": "Using EfficientNet-B3 from:\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nInstall by: pip install efficientnet_pytorch", "662002": "Using Appian's open source code:\nhttps://github.com/appian42/kaggle-rsna-intracranial-hemorrhage", "659722": "Pretrained models from: https://github.com/facebookresearch/semi-supervised-ImageNet1K-models\n", "659581": "https://github.com/qubvel/efficientnet\nhttps://github.com/qubvel/classification_models\n", "658870": "Pretrained model from: https://github.com/Tencent/MedicalNet", "658005": "Might use pretrained models from: \nhttps://github.com/Cadene/pretrained-models.pytorch\nhttps://github.com/facebookresearch/WSL-Images\nhttps://github.com/facebookresearch/FixRes\nhttps://pytorch.org/hub/pytorch_vision_resnext/", "657753": "Hi, I may possibly use the pretrained models from https://github.com/kenshohara/3D-ResNets-PyTorch", "656995": "https://pytorch.org/hub/research-models\nhttps://github.com/lukemelas/EfficientNet-PyTorch", "656204": "keras efficientnet\n\nhttps://github.com/titu1994/keras-efficientnets", "655441": "http://headctstudy.qure.ai\nhttps://github.com/lukemelas/EfficientNet-PyTorch\nhttps://github.com/rwightman/pytorch-image-models/\n Data from: https://mrbrains18.isi.uu.nl/", "654760": "@anoukstein Thanks, could you tell me how to match your annotations with images?", "654694": "Pretrained models from: https://github.com/creafz/pytorch-cnn-finetune", "652018": "https://github.com/lukemelas/EfficientNet-PyTorch", "651724": "fastai_dev pretrainedmodels\nfastai pretrained model\npretrained-models.pytorch", "648657": "Keras pertained models with imagnet weights:\nXception\nVGG16\nVGG19\nResNet, ResNetV2\nInceptionV3\nInceptionResNetV2\nDenseNet\nNASNet\nhttps://keras.io/applications/", "648565": "Pretrained models from:\n1. torchvision.models\n2. https://github.com/osmr/imgclsmob/tree/master/pytorch/pytorchcv\n3. https://github.com/facebookresearch/WSL-Images", "637002": "Pretrained model from https://github.com/rgeirhos/texture-vs-shape", "636991": "Are we supposed to share external datasets only or pretrained models as well?", "630982": "Any of the pretrained models and model architectures from any of the following locations:\n\n[fastai_dev pretrainedmodels](https://github.com/fastai/fastai_dev)\n[fastai pretrained model](https://github.com/fastai/fastai)\n[pretrained-models.pytorch](https://github.com/Cadene/pretrained-models.pytorch)", "650113": "" } }