Welcome to the State Farm Distracted Driver Detection competition. We are excited to partner with Kaggle to help solve this issue. We will be monitoring this forum and answer any questions you may have. Thank you for participating in our competition, and we look forward to your innovative solutions to help solve this challenging problem!
\n\nDan Holman - State Farm
", "messages": [ { "id": "113860", "postDate": "04/05/2016 18:37:17", "content": "Welcome to the State Farm Distracted Driver Detection competition. We are excited to partner with Kaggle to help solve this issue. We will be monitoring this forum and answer any questions you may have. Thank you for participating in our competition, and we look forward to your innovative solutions to help solve this challenging problem!
\n\nDan Holman - State Farm
", "rawMarkdown": "Welcome to the State Farm Distracted Driver Detection competition. We are excited to partner with Kaggle to help solve this issue. We will be monitoring this forum and answer any questions you may have. Thank you for participating in our competition, and we look forward to your innovative solutions to help solve this challenging problem!\r\n\r\nDan Holman - State Farm", "votes": null }, { "id": "113870", "postDate": "04/05/2016 20:33:26", "content": "Thanks Dan!
\n\nI think Ford had a scaled down but somewhat related competition 5 years ago and I was really bummed to miss it, so I can't wait to participate.
\n\nThis competition certainly has the potential to do a lot of good for the public.
", "rawMarkdown": "Thanks Dan! \r\n\r\nI think Ford had a scaled down but somewhat related competition 5 years ago and I was really bummed to miss it, so I can't wait to participate.\r\n\r\nThis competition certainly has the potential to do a lot of good for the public.", "votes": null }, { "id": "113895", "postDate": "04/06/2016 01:16:57", "content": "Thanks Dan,
\n\nI am really excited as the problem looks really interesting to me.
", "rawMarkdown": "Thanks Dan,\r\n\r\nI am really excited as the problem looks really interesting to me.", "votes": null }, { "id": "114355", "postDate": "04/09/2016 17:56:16", "content": "Hi,\nI wanted to ask if the provided labels for the images are ground truth, or are up to human mistake. For instance if we look at 'img_139.jpg' which is placed in class 0 - save driving. This to me seems like a total mistake. This is definitely c9 - talking to passenger (she is not even looking at the road, what kind of save driving is it?). Thus as so I'm quite curious for several questions:
\n\nIf the answer to Q3 is yes, than we are for a big trouble, as the perfect model would fail to give best results.
", "rawMarkdown": "Hi,\r\nI wanted to ask if the provided labels for the images are ground truth, or are up to human mistake. For instance if we look at 'img_139.jpg' which is placed in class 0 - save driving. This to me seems like a total mistake. This is definitely c9 - talking to passenger (she is not even looking at the road, what kind of save driving is it?). Thus as so I'm quite curious for several questions:\r\n\r\n1. Is this intentional?\r\n2. If not is this just human error?\r\n3. Are we to expect your ground truth labels for the test data to have same error?\r\n\r\nIf the answer to Q3 is yes, than we are for a big trouble, as the perfect model would fail to give best results.", "votes": null }, { "id": "114358", "postDate": "04/09/2016 18:28:32", "content": "I do not agree that safe driving implies you're looking straight ahead. She might be at a turn or a crossroads and looking both ways before crossing or turning onto the road. Besides that, not all hand labeling is perfect (see HomeDepot), so your model must be able to deal with some noise in the data. If you check some of the images in c9, there are cases where the person does look at the road, yet is obviously in some kind of conversation.
\n\nThe angle of the camera is also somewhat misleading, it's not taken straight from the passenger seat but somewhat before that. So when the person stares apparently into the camera, they're actually gazing slightly ahead of the passenger beside them.
\n\nA better indication of having a conversation is the position of the arm, wild gestures with either hand, wide smiles and other emotional facial cues like surprise or puzzled looks.
", "rawMarkdown": "I do not agree that safe driving implies you're looking straight ahead. She might be at a turn or a crossroads and looking both ways before crossing or turning onto the road. Besides that, not all hand labeling is perfect (see HomeDepot), so your model must be able to deal with some noise in the data. If you check some of the images in c9, there are cases where the person does look at the road, yet is obviously in some kind of conversation.\r\n\r\nThe angle of the camera is also somewhat misleading, it's not taken straight from the passenger seat but somewhat before that. So when the person stares apparently into the camera, they're actually gazing slightly ahead of the passenger beside them.\r\n\r\nA better indication of having a conversation is the position of the arm, wild gestures with either hand, wide smiles and other emotional facial cues like surprise or puzzled looks.", "votes": null }, { "id": "117985", "postDate": "05/02/2016 09:43:33", "content": "can we use MATLAB for this assignmnet
", "rawMarkdown": "can we use MATLAB for this assignmnet", "votes": null }, { "id": "121053", "postDate": "05/23/2016 09:31:43", "content": "Hello,
\n\nThanks Dan for this competition!
\n\nI have a question regarding the content of the full dataset. Could we know if there is the same drivers into the train set and the test set ?
\n\nBest regards,
\n\nJulien
", "rawMarkdown": "Hello, \r\n\r\nThanks Dan for this competition! \r\n\r\nI have a question regarding the content of the full dataset. Could we know if there is the same drivers into the train set and the test set ? \r\n\r\nBest regards,\r\n\r\nJulien", "votes": null }, { "id": "121060", "postDate": "05/23/2016 11:13:47", "content": "@Julien - Different drivers in the Test images.
", "rawMarkdown": "Julien - Different drivers in the Test images.", "votes": null }, { "id": "121068", "postDate": "05/23/2016 13:19:16", "content": "So I can't do a specific model for each driver in the train set. Thank you for your answer.
", "rawMarkdown": "So I can't do a specific model for each driver in the train set. Thank you for your answer.", "votes": null }, { "id": "126650", "postDate": "07/11/2016 15:44:41", "content": "This maybe a naive question but I didn't find it in the data description page.
\n\nThere are 26 unique drivers in the training set. How about the test set? Are the drivers in the test appear in the training set or they are new drivers?
\n\nIs the public and private LB data split by random or by drivers?
", "rawMarkdown": "This maybe a naive question but I didn't find it in the data description page.\r\n\r\nThere are 26 unique drivers in the training set. How about the test set? Are the drivers in the test appear in the training set or they are new drivers? \r\n\r\nIs the public and private LB data split by random or by drivers?", "votes": null }, { "id": "128693", "postDate": "07/22/2016 19:24:36", "content": "Can this be answered, please? https://www.kaggle.com/c/state-farm-distracted-driver-detection/forums/t/22298/can-the-test-set-images-be-processed-in-batch-mode
", "rawMarkdown": "Can this be answered, please? https://www.kaggle.com/c/state-farm-distracted-driver-detection/forums/t/22298/can-the-test-set-images-be-processed-in-batch-mode", "votes": null }, { "id": "294545", "postDate": "03/12/2018 05:58:22", "content": "Dear Julia Elliott, Dan Holman, (to competition organizer)
\n\nI am planning to write a beginner's guide to kaggle for ML community in South Korea, and include this competition as a main competition case study.
\n\nMain content of the publication will be EDA of this competition, and winner's solutions shared on github with appropriate licenses by competition winners.
\n\nI was wondering if I can include EDA part of this competition in it. Visualizations of raw data or its derivatives (attention maps on top of the image etc) are possible content candidates.
\n\nIf using the data as-is is a sensitive matter and not allowed, I plan to include executable codes written by me with no image displays so that raw data or its derivatives are not included in the publication, but readers can follow the code and reproduce those on their computer.
\n\nPlease let me know, Best regards, Kweonwoo
\n\nAbove message is posted in a separate discussion post\nhttps://www.kaggle.com/c/state-farm-distracted-driver-detection/discussion/51687
", "rawMarkdown": "Dear Julia Elliott, Dan Holman, (to competition organizer)\n\nI am planning to write a beginner's guide to kaggle for ML community in South Korea, and include this competition as a main competition case study.\n\nMain content of the publication will be EDA of this competition, and winner's solutions shared on github with appropriate licenses by competition winners.\n\nI was wondering if I can include EDA part of this competition in it. Visualizations of raw data or its derivatives (attention maps on top of the image etc) are possible content candidates.\n\nIf using the data as-is is a sensitive matter and not allowed, I plan to include executable codes written by me with no image displays so that raw data or its derivatives are not included in the publication, but readers can follow the code and reproduce those on their computer.\n\nPlease let me know, Best regards, Kweonwoo\n\nAbove message is posted in a separate discussion post\nhttps://www.kaggle.com/c/state-farm-distracted-driver-detection/discussion/51687", "votes": null } ], "comments": [ { "id": 113870, "author_name": "millerintllc", "author_url": "", "post_date": "04/05/2016 20:33:26", "content": "Thanks Dan!
\n\nI think Ford had a scaled down but somewhat related competition 5 years ago and I was really bummed to miss it, so I can't wait to participate.
\n\nThis competition certainly has the potential to do a lot of good for the public.
", "votes": null, "replies": [] }, { "id": 113895, "author_name": "shubhamp05", "author_url": "", "post_date": "04/06/2016 01:16:57", "content": "Thanks Dan,
\n\nI am really excited as the problem looks really interesting to me.
", "votes": null, "replies": [] }, { "id": 114355, "author_name": "alexanderbotev", "author_url": "", "post_date": "04/09/2016 17:56:16", "content": "Hi,\nI wanted to ask if the provided labels for the images are ground truth, or are up to human mistake. For instance if we look at 'img_139.jpg' which is placed in class 0 - save driving. This to me seems like a total mistake. This is definitely c9 - talking to passenger (she is not even looking at the road, what kind of save driving is it?). Thus as so I'm quite curious for several questions:
\n\nIf the answer to Q3 is yes, than we are for a big trouble, as the perfect model would fail to give best results.
", "votes": null, "replies": [] }, { "id": 114358, "author_name": "remap1", "author_url": "", "post_date": "04/09/2016 18:28:32", "content": "I do not agree that safe driving implies you're looking straight ahead. She might be at a turn or a crossroads and looking both ways before crossing or turning onto the road. Besides that, not all hand labeling is perfect (see HomeDepot), so your model must be able to deal with some noise in the data. If you check some of the images in c9, there are cases where the person does look at the road, yet is obviously in some kind of conversation.
\n\nThe angle of the camera is also somewhat misleading, it's not taken straight from the passenger seat but somewhat before that. So when the person stares apparently into the camera, they're actually gazing slightly ahead of the passenger beside them.
\n\nA better indication of having a conversation is the position of the arm, wild gestures with either hand, wide smiles and other emotional facial cues like surprise or puzzled looks.
", "votes": null, "replies": [] }, { "id": 117985, "author_name": "chetanksreddy", "author_url": "", "post_date": "05/02/2016 09:43:33", "content": "can we use MATLAB for this assignmnet
", "votes": null, "replies": [] }, { "id": 121053, "author_name": "krywyk", "author_url": "", "post_date": "05/23/2016 09:31:43", "content": "Hello,
\n\nThanks Dan for this competition!
\n\nI have a question regarding the content of the full dataset. Could we know if there is the same drivers into the train set and the test set ?
\n\nBest regards,
\n\nJulien
", "votes": null, "replies": [] }, { "id": 121060, "author_name": "inversion", "author_url": "", "post_date": "05/23/2016 11:13:47", "content": "@Julien - Different drivers in the Test images.
", "votes": null, "replies": [] }, { "id": 121068, "author_name": "krywyk", "author_url": "", "post_date": "05/23/2016 13:19:16", "content": "So I can't do a specific model for each driver in the train set. Thank you for your answer.
", "votes": null, "replies": [] }, { "id": 126650, "author_name": "aikinogard", "author_url": "", "post_date": "07/11/2016 15:44:41", "content": "This maybe a naive question but I didn't find it in the data description page.
\n\nThere are 26 unique drivers in the training set. How about the test set? Are the drivers in the test appear in the training set or they are new drivers?
\n\nIs the public and private LB data split by random or by drivers?
", "votes": null, "replies": [] }, { "id": 128693, "author_name": "bobutis", "author_url": "", "post_date": "07/22/2016 19:24:36", "content": "Can this be answered, please? https://www.kaggle.com/c/state-farm-distracted-driver-detection/forums/t/22298/can-the-test-set-images-be-processed-in-batch-mode
", "votes": null, "replies": [] }, { "id": 294545, "author_name": "kweonwooj", "author_url": "", "post_date": "03/12/2018 05:58:22", "content": "Dear Julia Elliott, Dan Holman, (to competition organizer)
\n\nI am planning to write a beginner's guide to kaggle for ML community in South Korea, and include this competition as a main competition case study.
\n\nMain content of the publication will be EDA of this competition, and winner's solutions shared on github with appropriate licenses by competition winners.
\n\nI was wondering if I can include EDA part of this competition in it. Visualizations of raw data or its derivatives (attention maps on top of the image etc) are possible content candidates.
\n\nIf using the data as-is is a sensitive matter and not allowed, I plan to include executable codes written by me with no image displays so that raw data or its derivatives are not included in the publication, but readers can follow the code and reproduce those on their computer.
\n\nPlease let me know, Best regards, Kweonwoo
\n\nAbove message is posted in a separate discussion post\nhttps://www.kaggle.com/c/state-farm-distracted-driver-detection/discussion/51687
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "113860": "Welcome to the State Farm Distracted Driver Detection competition. We are excited to partner with Kaggle to help solve this issue. We will be monitoring this forum and answer any questions you may have. Thank you for participating in our competition, and we look forward to your innovative solutions to help solve this challenging problem!\r\n\r\nDan Holman - State Farm", "113870": "Thanks Dan! \r\n\r\nI think Ford had a scaled down but somewhat related competition 5 years ago and I was really bummed to miss it, so I can't wait to participate.\r\n\r\nThis competition certainly has the potential to do a lot of good for the public.", "113895": "Thanks Dan,\r\n\r\nI am really excited as the problem looks really interesting to me.", "114355": "Hi,\r\nI wanted to ask if the provided labels for the images are ground truth, or are up to human mistake. For instance if we look at 'img_139.jpg' which is placed in class 0 - save driving. This to me seems like a total mistake. This is definitely c9 - talking to passenger (she is not even looking at the road, what kind of save driving is it?). Thus as so I'm quite curious for several questions:\r\n\r\n1. Is this intentional?\r\n2. If not is this just human error?\r\n3. Are we to expect your ground truth labels for the test data to have same error?\r\n\r\nIf the answer to Q3 is yes, than we are for a big trouble, as the perfect model would fail to give best results.", "114358": "I do not agree that safe driving implies you're looking straight ahead. She might be at a turn or a crossroads and looking both ways before crossing or turning onto the road. Besides that, not all hand labeling is perfect (see HomeDepot), so your model must be able to deal with some noise in the data. If you check some of the images in c9, there are cases where the person does look at the road, yet is obviously in some kind of conversation.\r\n\r\nThe angle of the camera is also somewhat misleading, it's not taken straight from the passenger seat but somewhat before that. So when the person stares apparently into the camera, they're actually gazing slightly ahead of the passenger beside them.\r\n\r\nA better indication of having a conversation is the position of the arm, wild gestures with either hand, wide smiles and other emotional facial cues like surprise or puzzled looks.", "117985": "can we use MATLAB for this assignmnet", "121053": "Hello, \r\n\r\nThanks Dan for this competition! \r\n\r\nI have a question regarding the content of the full dataset. Could we know if there is the same drivers into the train set and the test set ? \r\n\r\nBest regards,\r\n\r\nJulien", "121060": "Julien - Different drivers in the Test images.", "121068": "So I can't do a specific model for each driver in the train set. Thank you for your answer.", "126650": "This maybe a naive question but I didn't find it in the data description page.\r\n\r\nThere are 26 unique drivers in the training set. How about the test set? Are the drivers in the test appear in the training set or they are new drivers? \r\n\r\nIs the public and private LB data split by random or by drivers?", "128693": "Can this be answered, please? https://www.kaggle.com/c/state-farm-distracted-driver-detection/forums/t/22298/can-the-test-set-images-be-processed-in-batch-mode", "294545": "Dear Julia Elliott, Dan Holman, (to competition organizer)\n\nI am planning to write a beginner's guide to kaggle for ML community in South Korea, and include this competition as a main competition case study.\n\nMain content of the publication will be EDA of this competition, and winner's solutions shared on github with appropriate licenses by competition winners.\n\nI was wondering if I can include EDA part of this competition in it. Visualizations of raw data or its derivatives (attention maps on top of the image etc) are possible content candidates.\n\nIf using the data as-is is a sensitive matter and not allowed, I plan to include executable codes written by me with no image displays so that raw data or its derivatives are not included in the publication, but readers can follow the code and reproduce those on their computer.\n\nPlease let me know, Best regards, Kweonwoo\n\nAbove message is posted in a separate discussion post\nhttps://www.kaggle.com/c/state-farm-distracted-driver-detection/discussion/51687" }, "source": "meta" }