![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F20559325%2F2a90ee14111e89e302a5ed241269b991%2FScreenshot%202024-11-08%20085422.png?generation=1731036454960990&alt=media\")
are we required to first classify the type of particle and then predict their x,y,z positions or do we only need to predict x,y,z value for a particular type of particleCan anyone tell me what do these transformation represent or mean
", "messages": [ { "id": 3039159, "postDate": "2024-11-07T18:21:06.847Z", "content": "Can anyone tell me what do these transformation represent or mean
", "rawMarkdown": "Can anyone tell me what do these transformation represent or mean ", "votes": 3 }, { "id": 3039192, "postDate": "2024-11-07T19:14:52.007Z", "content": "Hi @avish006, when processing cryoET data we sometimes determine the orientation of particles in the tomogram. For this reason, the copick picks format has the capability to store orientations as a transformation matrix. For the purpose of this competition, we chose not to provide orientation information for the experimental data, so you can safely ignore the transformation information (it should be an identity matrix in all provided files).
\nIf you think this information could be useful to you, take a look at the simulated dataset on the CZ cryoET data portal for examples of data that includes this information.
", "rawMarkdown": "Hi @avish006, when processing cryoET data we sometimes determine the orientation of particles in the tomogram. For this reason, the copick picks format has the capability to store orientations as a transformation matrix. For the purpose of this competition, we chose not to provide orientation information for the experimental data, so you can safely ignore the transformation information (it should be an identity matrix in all provided files). \n\nIf you think this information could be useful to you, take a look at the [simulated dataset](https://cryoetdataportal.czscience.com/datasets/10441) on the CZ cryoET data portal for examples of data that includes this information.", "votes": 4, "replies": [ { "id": 3039430, "postDate": "2024-11-08T03:52:04.147Z", "content": "Thanks for the quick response ,one more question @uermel :)
\nin this are we required to first classify the type of particle and then predict their x,y,z positions or do we only need to predict x,y,z value for a particular type of particle
Hi @avish006,
\nYour submission should contain predicted x,y,z positions and particle classes for all particle types in all experiments.
\nYour question suggests 2 different strategies, and both are approaches that people use:
\nI would just point out that it isn't an option to \"first classify the type of particle\" because we do not provide positions, so you'll still need to figure out what positions in the image to try to classify.
\nIs this clear?
\nThank you!
\nKyle
Ok got it , thanks for the clear explanation
", "rawMarkdown": "Ok got it , thanks for the clear explanation" } ] } ] }, { "id": 3041529, "postDate": "2024-11-10T13:34:55.633Z", "content": "A related question: The (x, y, z) in the json file is for the particle's geometric center or \"mass\" center? Or different dataset can use different particle centers?
", "rawMarkdown": "A related question: The (x, y, z) in the json file is for the particle's geometric center or \"mass\" center? Or different dataset can use different particle centers? ", "votes": 2, "isDeleted": true, "replies": [ { "id": 3041590, "postDate": "2024-11-10T14:58:55.563Z", "content": "Although i don't remember where exactly i saw this written but i believe the location is at the center of mass of the object
", "rawMarkdown": "Although i don't remember where exactly i saw this written but i believe the location is at the center of mass of the object", "votes": 1, "replies": [ { "id": 3103313, "postDate": "2025-01-23T10:13:24.330Z", "content": "https://www.kaggle.com/competitions/czii-cryo-et-object-identification/discussion/544895#3039311
\nThis was the comment which mentioned predicted coordinated should be the center of mass
", "rawMarkdown": "https://www.kaggle.com/competitions/czii-cryo-et-object-identification/discussion/544895#3039311\n\nThis was the comment which mentioned predicted coordinated should be the center of mass" } ] } ] } ], "comments": [ { "id": 3039192, "author_name": "Utz Ermel", "author_url": "", "post_date": "2024-11-07T19:14:52.007000", "content": "Hi @avish006, when processing cryoET data we sometimes determine the orientation of particles in the tomogram. For this reason, the copick picks format has the capability to store orientations as a transformation matrix. For the purpose of this competition, we chose not to provide orientation information for the experimental data, so you can safely ignore the transformation information (it should be an identity matrix in all provided files).
\nIf you think this information could be useful to you, take a look at the simulated dataset on the CZ cryoET data portal for examples of data that includes this information.
", "votes": 4, "replies": [ { "id": 3039430, "author_name": "Avish Sinha", "author_url": "", "post_date": "2024-11-08T03:52:04.147000", "content": "Thanks for the quick response ,one more question @uermel :)
\nin this are we required to first classify the type of particle and then predict their x,y,z positions or do we only need to predict x,y,z value for a particular type of particle
Hi @avish006,
\nYour submission should contain predicted x,y,z positions and particle classes for all particle types in all experiments.
\nYour question suggests 2 different strategies, and both are approaches that people use:
\nI would just point out that it isn't an option to \"first classify the type of particle\" because we do not provide positions, so you'll still need to figure out what positions in the image to try to classify.
\nIs this clear?
\nThank you!
\nKyle
Ok got it , thanks for the clear explanation
", "votes": 0, "replies": [] } ] } ] }, { "id": 3041529, "author_name": "", "author_url": "", "post_date": "2024-11-10T13:34:55.633000", "content": "A related question: The (x, y, z) in the json file is for the particle's geometric center or \"mass\" center? Or different dataset can use different particle centers?
", "votes": 2, "replies": [ { "id": 3041590, "author_name": "Avish Sinha", "author_url": "", "post_date": "2024-11-10T14:58:55.563000", "content": "Although i don't remember where exactly i saw this written but i believe the location is at the center of mass of the object
", "votes": 1, "replies": [ { "id": 3103313, "author_name": "Keesari Vigneshwar Reddy", "author_url": "", "post_date": "2025-01-23T10:13:24.330000", "content": "https://www.kaggle.com/competitions/czii-cryo-et-object-identification/discussion/544895#3039311
\nThis was the comment which mentioned predicted coordinated should be the center of mass
", "votes": 0, "replies": [] } ] } ] } ], "raw_markdown_by_id": { "3039159": "Can anyone tell me what do these transformation represent or mean ", "3039192": "Hi @avish006, when processing cryoET data we sometimes determine the orientation of particles in the tomogram. For this reason, the copick picks format has the capability to store orientations as a transformation matrix. For the purpose of this competition, we chose not to provide orientation information for the experimental data, so you can safely ignore the transformation information (it should be an identity matrix in all provided files). \n\nIf you think this information could be useful to you, take a look at the [simulated dataset](https://cryoetdataportal.czscience.com/datasets/10441) on the CZ cryoET data portal for examples of data that includes this information.", "3041529": "A related question: The (x, y, z) in the json file is for the particle's geometric center or \"mass\" center? Or different dataset can use different particle centers? " } }