![\"2\"](\"http://diegojohnson.github.io/car1.gif\")
It's gif, maybe need to waite several seconds.\n
It's gif, maybe need to waite several seconds.\n
thanks, I learn a lot from this kernel. @ebouteillon\nhttps://www.kaggle.com/ebouteillon/augmented-reality
", "rawMarkdown": "thanks, I learn a lot from this kernel. @ebouteillon\nhttps://www.kaggle.com/ebouteillon/augmented-reality", "votes": null }, { "id": "707341", "postDate": "12/31/2019 17:06:33", "content": "Hi @diegojohnson , you're welcomed! 😄 \nIt seems your post contains an image, unfortunately, I cant see it.
", "rawMarkdown": "Hi @diegojohnson , you're welcomed! 😄 \nIt seems your post contains an image, unfortunately, I cant see it.", "votes": null }, { "id": "707346", "postDate": "12/31/2019 17:18:34", "content": "😂 okay, I'll fix it.
", "rawMarkdown": "😂 okay, I'll fix it.", "votes": null }, { "id": "707586", "postDate": "01/01/2020 05:30:26", "content": "hi,这个不能用来训练吧？我尝试了以后效果不太好...还是说透明度调高就可以用来训练了？
", "rawMarkdown": "hi,这个不能用来训练吧？我尝试了以后效果不太好...还是说透明度调高就可以用来训练了？", "votes": null }, { "id": "707592", "postDate": "01/01/2020 05:51:34", "content": "只是为了好看，基本上用不到。
", "rawMarkdown": "只是为了好看，基本上用不到。", "votes": null }, { "id": "707760", "postDate": "01/01/2020 12:37:23", "content": "Nice animation
", "rawMarkdown": "Nice animation", "votes": null }, { "id": "708098", "postDate": "01/02/2020 01:31:11", "content": "I think we can use them to get masks and bbox of cars (and this may lead to two-stage det models). \nTry \n> cv2.fillConvexPoly(image, coord, (0,0,255))
\n\ninstead of
\n\n> cv2.polylines(image, np.int32([coord]), 1, (0,0,255))
\n\nin code block no.5.
", "rawMarkdown": "I think we can use them to get masks and bbox of cars (and this may lead to two-stage det models). \nTry \n> cv2.fillConvexPoly(image, coord, (0,0,255))\n\ninstead of \n\n> cv2.polylines(image, np.int32([coord]), 1, (0,0,255))\n\nin code block no.5.", "votes": null }, { "id": "708110", "postDate": "01/02/2020 02:15:29", "content": "Thanks, good idea .It is better than my way to get bbox. I using the mean (Width&Height&Length) and center point to calculate the bbox , the error is a little bit bigger
", "rawMarkdown": "Thanks, good idea .It is better than my way to get bbox. I using the mean (Width&Height&Length) and center point to calculate the bbox , the error is a little bit bigger", "votes": null }, { "id": "708168", "postDate": "01/02/2020 03:25:47", "content": "Thanks for the excellent post
", "rawMarkdown": "Thanks for the excellent post", "votes": null }, { "id": "708339", "postDate": "01/02/2020 07:54:57", "content": "Hi,thanks for you excellent post, I am trying 3d target detection for the first time. I have a question, when the picture is flipped vertically, how to deal with the Angle?
", "rawMarkdown": "Hi,thanks for you excellent post, I am trying 3d target detection for the first time. I have a question, when the picture is flipped vertically, how to deal with the Angle?", "votes": null }, { "id": "708394", "postDate": "01/02/2020 08:59:10", "content": "![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F3876174%2F18e61736928a13d0b7353dca11adc166%2F2020-01-02_165813.png?generation=1577955547593169&alt=media\")
thanks, I learn a lot from this kernel. @ebouteillon\nhttps://www.kaggle.com/ebouteillon/augmented-reality
", "votes": null, "replies": [ { "id": 707586, "author_name": "guobaozi", "author_url": "", "post_date": "01/01/2020 05:30:26", "content": "hi,这个不能用来训练吧？我尝试了以后效果不太好...还是说透明度调高就可以用来训练了？
", "votes": null, "replies": [] }, { "id": 707592, "author_name": "diegojohnson", "author_url": "", "post_date": "01/01/2020 05:51:34", "content": "只是为了好看，基本上用不到。
", "votes": null, "replies": [] }, { "id": 708098, "author_name": "richardwong1994", "author_url": "", "post_date": "01/02/2020 01:31:11", "content": "I think we can use them to get masks and bbox of cars (and this may lead to two-stage det models). \nTry \n> cv2.fillConvexPoly(image, coord, (0,0,255))
\n\ninstead of
\n\n> cv2.polylines(image, np.int32([coord]), 1, (0,0,255))
\n\nin code block no.5.
", "votes": null, "replies": [] }, { "id": 708110, "author_name": "guobaozi", "author_url": "", "post_date": "01/02/2020 02:15:29", "content": "Thanks, good idea .It is better than my way to get bbox. I using the mean (Width&Height&Length) and center point to calculate the bbox , the error is a little bit bigger
", "votes": null, "replies": [] } ] }, { "id": 707341, "author_name": "ebouteillon", "author_url": "", "post_date": "12/31/2019 17:06:33", "content": "Hi @diegojohnson , you're welcomed! 😄 \nIt seems your post contains an image, unfortunately, I cant see it.
", "votes": null, "replies": [ { "id": 707346, "author_name": "diegojohnson", "author_url": "", "post_date": "12/31/2019 17:18:34", "content": "😂 okay, I'll fix it.
", "votes": null, "replies": [] } ] }, { "id": 707760, "author_name": "ebouteillon", "author_url": "", "post_date": "01/01/2020 12:37:23", "content": "Nice animation
", "votes": null, "replies": [] }, { "id": 708168, "author_name": "ahmedsmara", "author_url": "", "post_date": "01/02/2020 03:25:47", "content": "Thanks for the excellent post
", "votes": null, "replies": [] }, { "id": 708339, "author_name": "changebin", "author_url": "", "post_date": "01/02/2020 07:54:57", "content": "Hi,thanks for you excellent post, I am trying 3d target detection for the first time. I have a question, when the picture is flipped vertically, how to deal with the Angle?
", "votes": null, "replies": [ { "id": 708394, "author_name": "diegojohnson", "author_url": "", "post_date": "01/02/2020 08:59:10", "content": "