{ "id": 239898, "title": "💥💥 Winning Solutions of Similar Past Kaggle Competitions 🔥🔥", "url": "/competitions/siim-covid19-detection/discussion/239898", "author_name": "", "post_date": "2021-05-18T01:50:57.732096600Z", "votes": 72, "comment_count": 12, "views": 0, "content": "

In this post, I want to share some winner solution ideas and their codes of similar past Kaggle competitions. I publish another post to summarize key takeaways of the solutions including preprocessing, model zoo, and post-processing. Please check it also 😃

\n

1. VinBigData Chest X-ray Abnormalities Detection

\n\n

2. RSNA Pneumonia Detection Challenge

\n\n

3. SIIM-ACR Pneumothorax Segmentation

\n", "messages": [ { "id": "1312321", "postDate": "05/18/2021 01:50:57", "content": "

In this post, I want to share some winner solution ideas and their codes of similar past Kaggle competitions. I publish another post to summarize key takeaways of the solutions including preprocessing, model zoo, and post-processing. Please check it also 😃

\n

1. VinBigData Chest X-ray Abnormalities Detection

\n\n

2. RSNA Pneumonia Detection Challenge

\n\n

3. SIIM-ACR Pneumothorax Segmentation

\n", "rawMarkdown": "In this post, I want to share some winner solution ideas and their codes of similar past Kaggle competitions. I publish another post to summarize [key takeaways of the solutions](https://www.kaggle.com/c/siim-covid19-detection/discussion/240015) including preprocessing, model zoo, and post-processing. Please check it also 😃\n\n## 1. VinBigData Chest X-ray Abnormalities Detection\n\n- **1st Place** Solution + Code by @fatihozturk and @morizin -> ❤️ **code available** ❤️\nApproach: Ensembling Detectron2 + YOLOv5\n[Solution](https://www.kaggle.com/c/vinbigdata-chest-xray-abnormalities-detection/discussion/229724) by @fatihozturk, [Solution](https://www.kaggle.com/c/vinbigdata-chest-xray-abnormalities-detection/discussion/231511) by @morizin, [Code](https://www.kaggle.com/c/vinbigdata-chest-xray-abnormalities-detection/discussion/232393)\n- **2nd Place** Solution + Code by @zfturbo -> ❤️ **code available** ❤️\nApproach: Ensembling YOLOv5, ResNet101, ResNet152, and HourGlass\n[Solution](https://www.kaggle.com/c/vinbigdata-chest-xray-abnormalities-detection/discussion/229696), [Code](https://github.com/ZFTurbo/2nd-place-solution-for-VinBigData-Chest-X-ray-Abnormalities-Detection)\n- **5th Place** Solution by @aerdem4\nApproach: EfficientDet\n[Solution](https://www.kaggle.com/c/vinbigdata-chest-xray-abnormalities-detection/discussion/229711)\n- **6th Place** Solution by @wowfattie\nApproach: Preprocessing images, Efficientdet-d5 with 5 fold\n[Solution](https://www.kaggle.com/c/vinbigdata-chest-xray-abnormalities-detection/discussion/229770)\n- **7th Place** Solution by @cdeotte\nApproach: VFNet, Yolov5, EfficientDet4\n[Solution](https://www.kaggle.com/c/vinbigdata-chest-xray-abnormalities-detection/discussion/229637)\n- **Best EDA + Visualization** -> ❤️ **code available** ❤️\n[Notebook](https://www.kaggle.com/c/vinbigdata-chest-xray-abnormalities-detection/code?competitionId=24800&sortBy=voteCount)\n\n##2. RSNA Pneumonia Detection Challenge\n\n- **1st Place** Solution + Code -> ❤️ **code available** ❤️\nApproach: A classification-detection pipeline with 10-fold CV ensembled from RetinaNet, Deformable R-FCN, and Deformable Relation Networks \n[Solution](https://www.kaggle.com/c/rsna-pneumonia-detection-challenge/discussion/70421), [Code](https://github.com/i-pan/kaggle-rsna18)\n- **2nd Place** Solution\nApproach: Downscale the images, data augmentation, RetinaNet, some insightful postprocessing\n[Solution](https://www.kaggle.com/c/rsna-pneumonia-detection-challenge/discussion/70427)\n- **3rd Place** Solution + Code + Released Models -> ❤️ **code available** ❤️\nApproach: RetinaNet with Focal Loss \n[Solution ](https://www.kaggle.com/c/rsna-pneumonia-detection-challenge/discussion/70632), [Code](https://github.com/pmcheng/rsna-pneumonia), [Released Models](https://github.com/pmcheng/rsna-pneumonia/releases)\n- **6th Place** Solution + Code -> ❤️ **code available** ❤️\nApproach: Semantic segmentation, using U-net-like network architecture\n[Solution](https://www.kaggle.com/c/rsna-pneumonia-detection-challenge/discussion/70650), [Code](https://github.com/pfnet-research/pfneumonia)\n- **Best EDA + Visualization**\n[Notebook](https://www.kaggle.com/gpreda/rsna-pneumonia-detection-eda)\n\n\n## 3. SIIM-ACR Pneumothorax Segmentation\n\n- **1st Place** Solution + Code -> ❤️ **code available** ❤️\nApproach: Ensembling Resnet34, Resnet50, and SeResnext50\n[Solution](https://www.kaggle.com/c/siim-acr-pneumothorax-segmentation/discussion/107824), [Code](https://github.com/sneddy/kaggle-pneumathorax)\n- **2nd Place** Solution + Code -> ❤️ **code available** ❤️\nApproach: Two parts, i.e. classification and segmentation\n[Solution](https://www.kaggle.com/c/siim-acr-pneumothorax-segmentation/discussion/108009), [Code](https://github.com/yelanlan/Pneumothorax-Segmentation-2nd-place-solution)\n- **3rd Place** Solution + Code -> ❤️ **code available** ❤️\nApproach: UNet Resnet34 and SeResnext50\n[Solution](https://www.kaggle.com/c/siim-acr-pneumothorax-segmentation/discussion/107981), [Code](https://github.com/bestfitting/kaggle/tree/master/siim_acr)\n- **4th Place** Solution + Code -> ❤️ **code available** ❤️\nApproach: UNet ResNet34 backbone with frozen batch-normalization\n[Solution](https://www.kaggle.com/c/siim-acr-pneumothorax-segmentation/discussion/108397), [Code](https://github.com/amirassov/kaggle-pneumothorax)\n- **5th Place** Solution + Code -> ❤️ **code available** ❤️\nApproach: UNet with Aspp using SeResnext50 & SeResnext101\n[Solution](https://www.kaggle.com/c/siim-acr-pneumothorax-segmentation/discussion/107603), [Code](https://github.com/earhian/SIIM-ACR-Pneumothorax-Segmentation-5th)\n- **Best EDA + Visualization**\n[Notebook](https://www.kaggle.com/jesperdramsch/intro-chest-xray-dicom-viz-u-nets-full-data)", "votes": null }, { "id": "1314776", "postDate": "05/19/2021 11:28:41", "content": "

Thank you, I hope you get a good score too👍

", "rawMarkdown": "Thank you, I hope you get a good score too👍", "votes": null }, { "id": "1315760", "postDate": "05/20/2021 05:34:36", "content": "

Thank you

", "rawMarkdown": "Thank you", "votes": null }, { "id": "1320139", "postDate": "05/23/2021 19:20:40", "content": "

Thanks for these amazing resources

", "rawMarkdown": "Thanks for these amazing resources", "votes": null }, { "id": "1320152", "postDate": "05/23/2021 19:41:17", "content": "

Thanks for the Summary

", "rawMarkdown": "Thanks for the Summary", "votes": null }, { "id": "1320735", "postDate": "05/24/2021 09:40:13", "content": "

Thanks a lot!

", "rawMarkdown": "Thanks a lot!", "votes": null }, { "id": "1325559", "postDate": "05/27/2021 20:36:31", "content": "

Thanks a lot, this is very helpful.

", "rawMarkdown": "Thanks a lot, this is very helpful.", "votes": null }, { "id": "1326677", "postDate": "05/28/2021 15:51:10", "content": "

Thanks for posting, very useful.

", "rawMarkdown": "Thanks for posting, very useful.", "votes": null }, { "id": "1331149", "postDate": "06/01/2021 09:56:25", "content": "

thank you. its good.

", "rawMarkdown": "thank you. its good.", "votes": null }, { "id": "1331594", "postDate": "06/01/2021 15:11:07", "content": "

thank you for compiling all the resources!

", "rawMarkdown": "thank you for compiling all the resources!", "votes": null }, { "id": "1333900", "postDate": "06/03/2021 06:15:52", "content": "

Thank you for the resources😄

", "rawMarkdown": "Thank you for the resources😄", "votes": null }, { "id": "1343070", "postDate": "06/10/2021 02:29:21", "content": "

thank you very much. It will be very useful :)

", "rawMarkdown": "thank you very much. It will be very useful :)", "votes": null }, { "id": "1397737", "postDate": "07/23/2021 13:05:00", "content": "

Thanks for sharing @mhilmiasyrofi

", "rawMarkdown": "Thanks for sharing @mhilmiasyrofi", "votes": null } ], "comments": [ { "id": 1314776, "author_name": "sungzz", "author_url": "", "post_date": "05/19/2021 11:28:41", "content": "

Thank you, I hope you get a good score too👍

", "votes": null, "replies": [] }, { "id": 1315760, "author_name": "gautamgottipati", "author_url": "", "post_date": "05/20/2021 05:34:36", "content": "

Thank you

", "votes": null, "replies": [] }, { "id": 1320139, "author_name": "saketkattuboina", "author_url": "", "post_date": "05/23/2021 19:20:40", "content": "

Thanks for these amazing resources

", "votes": null, "replies": [] }, { "id": 1320152, "author_name": "cv13j0", "author_url": "", "post_date": "05/23/2021 19:41:17", "content": "

Thanks for the Summary

", "votes": null, "replies": [] }, { "id": 1320735, "author_name": "yoheiyamasaki", "author_url": "", "post_date": "05/24/2021 09:40:13", "content": "

Thanks a lot!

", "votes": null, "replies": [] }, { "id": 1325559, "author_name": "mohamedbakrey", "author_url": "", "post_date": "05/27/2021 20:36:31", "content": "

Thanks a lot, this is very helpful.

", "votes": null, "replies": [] }, { "id": 1326677, "author_name": "stevelangs", "author_url": "", "post_date": "05/28/2021 15:51:10", "content": "

Thanks for posting, very useful.

", "votes": null, "replies": [] }, { "id": 1331149, "author_name": "tensorchoko", "author_url": "", "post_date": "06/01/2021 09:56:25", "content": "

thank you. its good.

", "votes": null, "replies": [] }, { "id": 1331594, "author_name": "perkymaster", "author_url": "", "post_date": "06/01/2021 15:11:07", "content": "

thank you for compiling all the resources!

", "votes": null, "replies": [] }, { "id": 1333900, "author_name": "abhisheksabnis", "author_url": "", "post_date": "06/03/2021 06:15:52", "content": "

Thank you for the resources😄

", "votes": null, "replies": [] }, { "id": 1343070, "author_name": "shangweichen", "author_url": "", "post_date": "06/10/2021 02:29:21", "content": "

thank you very much. It will be very useful :)

", "votes": null, "replies": [] }, { "id": 1397737, "author_name": "amritpal333", "author_url": "", "post_date": "07/23/2021 13:05:00", "content": "

Thanks for sharing @mhilmiasyrofi

", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "1312321": "In this post, I want to share some winner solution ideas and their codes of similar past Kaggle competitions. I publish another post to summarize [key takeaways of the solutions](https://www.kaggle.com/c/siim-covid19-detection/discussion/240015) including preprocessing, model zoo, and post-processing. Please check it also 😃\n\n## 1. VinBigData Chest X-ray Abnormalities Detection\n\n- **1st Place** Solution + Code by @fatihozturk and @morizin -> ❤️ **code available** ❤️\nApproach: Ensembling Detectron2 + YOLOv5\n[Solution](https://www.kaggle.com/c/vinbigdata-chest-xray-abnormalities-detection/discussion/229724) by @fatihozturk, [Solution](https://www.kaggle.com/c/vinbigdata-chest-xray-abnormalities-detection/discussion/231511) by @morizin, [Code](https://www.kaggle.com/c/vinbigdata-chest-xray-abnormalities-detection/discussion/232393)\n- **2nd Place** Solution + Code by @zfturbo -> ❤️ **code available** ❤️\nApproach: Ensembling YOLOv5, ResNet101, ResNet152, and HourGlass\n[Solution](https://www.kaggle.com/c/vinbigdata-chest-xray-abnormalities-detection/discussion/229696), [Code](https://github.com/ZFTurbo/2nd-place-solution-for-VinBigData-Chest-X-ray-Abnormalities-Detection)\n- **5th Place** Solution by @aerdem4\nApproach: EfficientDet\n[Solution](https://www.kaggle.com/c/vinbigdata-chest-xray-abnormalities-detection/discussion/229711)\n- **6th Place** Solution by @wowfattie\nApproach: Preprocessing images, Efficientdet-d5 with 5 fold\n[Solution](https://www.kaggle.com/c/vinbigdata-chest-xray-abnormalities-detection/discussion/229770)\n- **7th Place** Solution by @cdeotte\nApproach: VFNet, Yolov5, EfficientDet4\n[Solution](https://www.kaggle.com/c/vinbigdata-chest-xray-abnormalities-detection/discussion/229637)\n- **Best EDA + Visualization** -> ❤️ **code available** ❤️\n[Notebook](https://www.kaggle.com/c/vinbigdata-chest-xray-abnormalities-detection/code?competitionId=24800&sortBy=voteCount)\n\n##2. RSNA Pneumonia Detection Challenge\n\n- **1st Place** Solution + Code -> ❤️ **code available** ❤️\nApproach: A classification-detection pipeline with 10-fold CV ensembled from RetinaNet, Deformable R-FCN, and Deformable Relation Networks \n[Solution](https://www.kaggle.com/c/rsna-pneumonia-detection-challenge/discussion/70421), [Code](https://github.com/i-pan/kaggle-rsna18)\n- **2nd Place** Solution\nApproach: Downscale the images, data augmentation, RetinaNet, some insightful postprocessing\n[Solution](https://www.kaggle.com/c/rsna-pneumonia-detection-challenge/discussion/70427)\n- **3rd Place** Solution + Code + Released Models -> ❤️ **code available** ❤️\nApproach: RetinaNet with Focal Loss \n[Solution ](https://www.kaggle.com/c/rsna-pneumonia-detection-challenge/discussion/70632), [Code](https://github.com/pmcheng/rsna-pneumonia), [Released Models](https://github.com/pmcheng/rsna-pneumonia/releases)\n- **6th Place** Solution + Code -> ❤️ **code available** ❤️\nApproach: Semantic segmentation, using U-net-like network architecture\n[Solution](https://www.kaggle.com/c/rsna-pneumonia-detection-challenge/discussion/70650), [Code](https://github.com/pfnet-research/pfneumonia)\n- **Best EDA + Visualization**\n[Notebook](https://www.kaggle.com/gpreda/rsna-pneumonia-detection-eda)\n\n\n## 3. SIIM-ACR Pneumothorax Segmentation\n\n- **1st Place** Solution + Code -> ❤️ **code available** ❤️\nApproach: Ensembling Resnet34, Resnet50, and SeResnext50\n[Solution](https://www.kaggle.com/c/siim-acr-pneumothorax-segmentation/discussion/107824), [Code](https://github.com/sneddy/kaggle-pneumathorax)\n- **2nd Place** Solution + Code -> ❤️ **code available** ❤️\nApproach: Two parts, i.e. classification and segmentation\n[Solution](https://www.kaggle.com/c/siim-acr-pneumothorax-segmentation/discussion/108009), [Code](https://github.com/yelanlan/Pneumothorax-Segmentation-2nd-place-solution)\n- **3rd Place** Solution + Code -> ❤️ **code available** ❤️\nApproach: UNet Resnet34 and SeResnext50\n[Solution](https://www.kaggle.com/c/siim-acr-pneumothorax-segmentation/discussion/107981), [Code](https://github.com/bestfitting/kaggle/tree/master/siim_acr)\n- **4th Place** Solution + Code -> ❤️ **code available** ❤️\nApproach: UNet ResNet34 backbone with frozen batch-normalization\n[Solution](https://www.kaggle.com/c/siim-acr-pneumothorax-segmentation/discussion/108397), [Code](https://github.com/amirassov/kaggle-pneumothorax)\n- **5th Place** Solution + Code -> ❤️ **code available** ❤️\nApproach: UNet with Aspp using SeResnext50 & SeResnext101\n[Solution](https://www.kaggle.com/c/siim-acr-pneumothorax-segmentation/discussion/107603), [Code](https://github.com/earhian/SIIM-ACR-Pneumothorax-Segmentation-5th)\n- **Best EDA + Visualization**\n[Notebook](https://www.kaggle.com/jesperdramsch/intro-chest-xray-dicom-viz-u-nets-full-data)", "1314776": "Thank you, I hope you get a good score too👍", "1315760": "Thank you", "1320139": "Thanks for these amazing resources", "1320152": "Thanks for the Summary", "1320735": "Thanks a lot!", "1325559": "Thanks a lot, this is very helpful.", "1326677": "Thanks for posting, very useful.", "1331149": "thank you. its good.", "1331594": "thank you for compiling all the resources!", "1333900": "Thank you for the resources😄", "1343070": "thank you very much. It will be very useful :)", "1397737": "Thanks for sharing @mhilmiasyrofi" }, "source": "meta" }