Just wonder why no one have mentioned that part of the data set was probably labeled using EVE Online players. There is some tutorial on their wiki that can help to understand some (20 of them) labels. Hope it will give you some ideas what features we are trying to extract from images.
", "messages": [ { "id": "406478", "postDate": "10/19/2018 10:01:12", "content": "Just wonder why no one have mentioned that part of the data set was probably labeled using EVE Online players. There is some tutorial on their wiki that can help to understand some (20 of them) labels. Hope it will give you some ideas what features we are trying to extract from images.
", "rawMarkdown": "Just wonder why no one have mentioned that part of the data set was probably labeled using EVE Online players. [There][1] is some tutorial on their wiki that can help to understand some (20 of them) labels. Hope it will give you some ideas what features we are trying to extract from images.\n\n\n [1]: https://wiki.eveuniversity.org/Project_Discovery:_Human_Protein_Atlas", "votes": null }, { "id": "406570", "postDate": "10/19/2018 13:03:55", "content": "You mean ground truth labels in the data set are not reliable and may have some noise?
", "rawMarkdown": "You mean ground truth labels in the data set are not reliable and may have some noise?", "votes": null }, { "id": "406583", "postDate": "10/19/2018 13:24:14", "content": "There is always room for human error, but I just wanted to share tutorial that gives examples of how target features may look like.
", "rawMarkdown": "There is always room for human error, but I just wanted to share tutorial that gives examples of how target features may look like.", "votes": null }, { "id": "406674", "postDate": "10/19/2018 16:06:05", "content": "As a former EVE player I thought it was pretty cool.
", "rawMarkdown": "As a former EVE player I thought it was pretty cool.", "votes": null }, { "id": "406941", "postDate": "10/20/2018 03:54:23", "content": "Thank you.\nI found another tutorial from [HPA site][1]\nhttps://www.proteinatlas.org/learn/dictionary/cell \nit covers *-somes.
\n\nIn EVE Online: Discovery Project,\nBlue & Red marker are used but not Yellow?
\n\nSome Kernels neglect yellow marker, is Yellow marker less informative...?
", "rawMarkdown": "Thank you.\nI found another tutorial from [HPA site][1]\nhttps://www.proteinatlas.org/learn/dictionary/cell \nit covers *-somes.\n\nIn EVE Online: Discovery Project,\nBlue & Red marker are used but not Yellow?\n\nSome Kernels neglect yellow marker, is Yellow marker less informative...?", "votes": null }, { "id": "407254", "postDate": "10/20/2018 18:01:53", "content": "I have not found any previous examples using yellow. They all used RG or RGB. I was using RGBY initially, discarding yellow runs faster and didn't seem to change the result. I plan to experiment with discarding B as well.
", "rawMarkdown": "I have not found any previous examples using yellow. They all used RG or RGB. I was using RGBY initially, discarding yellow runs faster and didn't seem to change the result. I plan to experiment with discarding B as well.", "votes": null } ], "comments": [ { "id": 406570, "author_name": "tomomimoriyama", "author_url": "", "post_date": "10/19/2018 13:03:55", "content": "You mean ground truth labels in the data set are not reliable and may have some noise?
", "votes": null, "replies": [] }, { "id": 406583, "author_name": "mortido", "author_url": "", "post_date": "10/19/2018 13:24:14", "content": "There is always room for human error, but I just wanted to share tutorial that gives examples of how target features may look like.
", "votes": null, "replies": [ { "id": 406941, "author_name": "tomomimoriyama", "author_url": "", "post_date": "10/20/2018 03:54:23", "content": "Thank you.\nI found another tutorial from [HPA site][1]\nhttps://www.proteinatlas.org/learn/dictionary/cell \nit covers *-somes.
\n\nIn EVE Online: Discovery Project,\nBlue & Red marker are used but not Yellow?
\n\nSome Kernels neglect yellow marker, is Yellow marker less informative...?
", "votes": null, "replies": [] }, { "id": 407254, "author_name": "ldm314", "author_url": "", "post_date": "10/20/2018 18:01:53", "content": "I have not found any previous examples using yellow. They all used RG or RGB. I was using RGBY initially, discarding yellow runs faster and didn't seem to change the result. I plan to experiment with discarding B as well.
", "votes": null, "replies": [] } ] }, { "id": 406674, "author_name": "ldm314", "author_url": "", "post_date": "10/19/2018 16:06:05", "content": "As a former EVE player I thought it was pretty cool.
", "votes": null, "replies": [] } ], "raw_markdown_by_id": { "406478": "Just wonder why no one have mentioned that part of the data set was probably labeled using EVE Online players. [There][1] is some tutorial on their wiki that can help to understand some (20 of them) labels. Hope it will give you some ideas what features we are trying to extract from images.\n\n\n [1]: https://wiki.eveuniversity.org/Project_Discovery:_Human_Protein_Atlas", "406570": "You mean ground truth labels in the data set are not reliable and may have some noise?", "406583": "There is always room for human error, but I just wanted to share tutorial that gives examples of how target features may look like.", "406674": "As a former EVE player I thought it was pretty cool.", "406941": "Thank you.\nI found another tutorial from [HPA site][1]\nhttps://www.proteinatlas.org/learn/dictionary/cell \nit covers *-somes.\n\nIn EVE Online: Discovery Project,\nBlue & Red marker are used but not Yellow?\n\nSome Kernels neglect yellow marker, is Yellow marker less informative...?", "407254": "I have not found any previous examples using yellow. They all used RG or RGB. I was using RGBY initially, discarding yellow runs faster and didn't seem to change the result. I plan to experiment with discarding B as well." }, "source": "meta" }