Welcome to the 3D Car Pose Estimation Challenge on Autonomous Driving. We (Peking University/Baidu and Kaggle) are so excited for to host this competition, which provides a very challenging task to estimate the vehicle pose from a single frame. For this challenge, we provide about 60,000 labeled 3D car instances from 5,324 real-world images, based on industry-grade CAD car models.
\n\nIn the training dataset, we provide the RGB image, camera intrinsic parameters, CAD car models, ignore mask and corresponding car poses in each frame. For challenge competition, you are asked to give the 3D pose for each vehicle in the image. To get the best results, you are encouraged to explore the CAD models for providing you some prior information of the vehicle.
\n\nBy the way, please refer the following website for more challenges on Autonomous Driving. http://apolloscape.auto/
\n\nPlease feel free to ask your questions in this thread.
\n\nGood luck!
", "messages": [ { "id": "678193", "postDate": "11/21/2019 04:44:16", "content": "Welcome to the 3D Car Pose Estimation Challenge on Autonomous Driving. We (Peking University/Baidu and Kaggle) are so excited for to host this competition, which provides a very challenging task to estimate the vehicle pose from a single frame. For this challenge, we provide about 60,000 labeled 3D car instances from 5,324 real-world images, based on industry-grade CAD car models.
\n\nIn the training dataset, we provide the RGB image, camera intrinsic parameters, CAD car models, ignore mask and corresponding car poses in each frame. For challenge competition, you are asked to give the 3D pose for each vehicle in the image. To get the best results, you are encouraged to explore the CAD models for providing you some prior information of the vehicle.
\n\nBy the way, please refer the following website for more challenges on Autonomous Driving. http://apolloscape.auto/
\n\nPlease feel free to ask your questions in this thread.
\n\nGood luck!
", "rawMarkdown": "Welcome to the 3D Car Pose Estimation Challenge on Autonomous Driving. We (Peking University/Baidu and Kaggle) are so excited for to host this competition, which provides a very challenging task to estimate the vehicle pose from a single frame. For this challenge, we provide about 60,000 labeled 3D car instances from 5,324 real-world images, based on industry-grade CAD car models.\n\nIn the training dataset, we provide the RGB image, camera intrinsic parameters, CAD car models, ignore mask and corresponding car poses in each frame. For challenge competition, you are asked to give the 3D pose for each vehicle in the image. To get the best results, you are encouraged to explore the CAD models for providing you some prior information of the vehicle.\n\nBy the way, please refer the following website for more challenges on Autonomous Driving. http://apolloscape.auto/\n\nPlease feel free to ask your questions in this thread.\n\nGood luck!", "votes": null }, { "id": "678211", "postDate": "11/21/2019 05:13:04", "content": "Hi @dingfuzhou, it looks like leaderboard's scores are wrong. Please see discussions here: https://www.kaggle.com/c/pku-autonomous-driving/discussion/116661
\n\nCan you provide the evaluation code? Thanks.
", "rawMarkdown": "Hi @dingfuzhou, it looks like leaderboard's scores are wrong. Please see discussions here: https://www.kaggle.com/c/pku-autonomous-driving/discussion/116661\n\nCan you provide the evaluation code? Thanks.", "votes": null }, { "id": "678244", "postDate": "11/21/2019 06:28:01", "content": "Hi, @dingfuzhou \nbroken images\nFive pictures seem to be broken.Is there any approach to solve this problem
", "rawMarkdown": "Hi, @dingfuzhou \n[broken images](https://www.kaggle.com/c/pku-autonomous-driving/discussion/117621#latest-677894)\nFive pictures seem to be broken.Is there any approach to solve this problem", "votes": null }, { "id": "679656", "postDate": "11/23/2019 05:35:12", "content": "@dingfuzhou I hope you could answer Bo‘s question, which is very important for competitors, it will determine how we optimize the code
", "rawMarkdown": "dingfuzhou I hope you could answer Bo‘s question, which is very important for competitors, it will determine how we optimize the code", "votes": null }, { "id": "683178", "postDate": "11/28/2019 07:08:55", "content": "@boliu0 @hesene Thanks for your feedback. Currently, we are working with Kaggle to find these problems.
", "rawMarkdown": "boliu0 @hesene Thanks for your feedback. Currently, we are working with Kaggle to find these problems.", "votes": null }, { "id": "683182", "postDate": "11/28/2019 07:12:22", "content": "@frkaka Hi, thanks for your feedback, would you provide the image id of these images please?
", "rawMarkdown": "frkaka Hi, thanks for your feedback, would you provide the image id of these images please?", "votes": null }, { "id": "683465", "postDate": "11/28/2019 13:10:27", "content": "Thank you for looking into this, here are the images IDs. Thankfully there don't seem to be errors in the test set.
\n\n\nID_1a5a10365.jpg\nID_4d238ae90.jpg\nID_408f58e9f.jpg\nID_bb1d991f6.jpg\nID_c44983aeb.jpg\n
HI @dingfuzhou ,
\n\nYou are emphasizing the fact, that there are \"industry-grade CAD car models\", but there is no reference of which CAD model belongs to which car id ('model type' like it is called in the data section), is there?
", "rawMarkdown": "HI @dingfuzhou ,\n\nYou are emphasizing the fact, that there are \"industry-grade CAD car models\", but there is no reference of which CAD model belongs to which car id ('model type' like it is called in the data section), is there?", "votes": null }, { "id": "693154", "postDate": "12/12/2019 04:57:55", "content": "Is it possible to increase the team size limit? Thanks! @dingfuzhou
", "rawMarkdown": "Is it possible to increase the team size limit? Thanks! @dingfuzhou", "votes": null }, { "id": "701890", "postDate": "12/24/2019 02:50:27", "content": "Hi, thanks for your interest in this competion. The think the team size if ruled by Kaggle. I can help you to ask whether we can change it.
\n\nBest
", "rawMarkdown": "Hi, thanks for your interest in this competion. The think the team size if ruled by Kaggle. I can help you to ask whether we can change it. \n\nBest", "votes": null } ], "comments": [ { "id": 678211, "author_name": "boliu0", "author_url": "", "post_date": "11/21/2019 05:13:04", "content": "Hi @dingfuzhou, it looks like leaderboard's scores are wrong. Please see discussions here: https://www.kaggle.com/c/pku-autonomous-driving/discussion/116661
\n\nCan you provide the evaluation code? Thanks.
", "votes": null, "replies": [ { "id": 679656, "author_name": "hesene", "author_url": "", "post_date": "11/23/2019 05:35:12", "content": "@dingfuzhou I hope you could answer Bo‘s question, which is very important for competitors, it will determine how we optimize the code
", "votes": null, "replies": [] } ] }, { "id": 678244, "author_name": "frkaka", "author_url": "", "post_date": "11/21/2019 06:28:01", "content": "Hi, @dingfuzhou \nbroken images\nFive pictures seem to be broken.Is there any approach to solve this problem
", "votes": null, "replies": [] }, { "id": 683178, "author_name": "dingfuzhou", "author_url": "", "post_date": "11/28/2019 07:08:55", "content": "@boliu0 @hesene Thanks for your feedback. Currently, we are working with Kaggle to find these problems.
", "votes": null, "replies": [] }, { "id": 683182, "author_name": "dingfuzhou", "author_url": "", "post_date": "11/28/2019 07:12:22", "content": "@frkaka Hi, thanks for your feedback, would you provide the image id of these images please?
", "votes": null, "replies": [ { "id": 683465, "author_name": "ilu000", "author_url": "", "post_date": "11/28/2019 13:10:27", "content": "Thank you for looking into this, here are the images IDs. Thankfully there don't seem to be errors in the test set.
\n\n\nID_1a5a10365.jpg\nID_4d238ae90.jpg\nID_408f58e9f.jpg\nID_bb1d991f6.jpg\nID_c44983aeb.jpg\n
HI @dingfuzhou ,
\n\nYou are emphasizing the fact, that there are \"industry-grade CAD car models\", but there is no reference of which CAD model belongs to which car id ('model type' like it is called in the data section), is there?
", "votes": null, "replies": [] }, { "id": 693154, "author_name": "tonychenxyz", "author_url": "", "post_date": "12/12/2019 04:57:55", "content": "Is it possible to increase the team size limit? Thanks! @dingfuzhou
", "votes": null, "replies": [ { "id": 701890, "author_name": "dingfuzhou", "author_url": "", "post_date": "12/24/2019 02:50:27", "content": "Hi, thanks for your interest in this competion. The think the team size if ruled by Kaggle. I can help you to ask whether we can change it.
\n\nBest
", "votes": null, "replies": [] } ] } ], "raw_markdown_by_id": { "678193": "Welcome to the 3D Car Pose Estimation Challenge on Autonomous Driving. We (Peking University/Baidu and Kaggle) are so excited for to host this competition, which provides a very challenging task to estimate the vehicle pose from a single frame. For this challenge, we provide about 60,000 labeled 3D car instances from 5,324 real-world images, based on industry-grade CAD car models.\n\nIn the training dataset, we provide the RGB image, camera intrinsic parameters, CAD car models, ignore mask and corresponding car poses in each frame. For challenge competition, you are asked to give the 3D pose for each vehicle in the image. To get the best results, you are encouraged to explore the CAD models for providing you some prior information of the vehicle.\n\nBy the way, please refer the following website for more challenges on Autonomous Driving. http://apolloscape.auto/\n\nPlease feel free to ask your questions in this thread.\n\nGood luck!", "678211": "Hi @dingfuzhou, it looks like leaderboard's scores are wrong. Please see discussions here: https://www.kaggle.com/c/pku-autonomous-driving/discussion/116661\n\nCan you provide the evaluation code? Thanks.", "678244": "Hi, @dingfuzhou \n[broken images](https://www.kaggle.com/c/pku-autonomous-driving/discussion/117621#latest-677894)\nFive pictures seem to be broken.Is there any approach to solve this problem", "679656": "dingfuzhou I hope you could answer Bo‘s question, which is very important for competitors, it will determine how we optimize the code", "683178": "boliu0 @hesene Thanks for your feedback. Currently, we are working with Kaggle to find these problems.", "683182": "frkaka Hi, thanks for your feedback, would you provide the image id of these images please?", "683465": "Thank you for looking into this, here are the images IDs. Thankfully there don't seem to be errors in the test set.\n\n```\nID_1a5a10365.jpg\nID_4d238ae90.jpg\nID_408f58e9f.jpg\nID_bb1d991f6.jpg\nID_c44983aeb.jpg\n```", "688541": "HI @dingfuzhou ,\n\nYou are emphasizing the fact, that there are \"industry-grade CAD car models\", but there is no reference of which CAD model belongs to which car id ('model type' like it is called in the data section), is there?", "693154": "Is it possible to increase the team size limit? Thanks! @dingfuzhou", "701890": "Hi, thanks for your interest in this competion. The think the team size if ruled by Kaggle. I can help you to ask whether we can change it. \n\nBest" }, "source": "meta" }