![\"\"](\"https://www.googleapis.com/download/storage/v1/b/kaggle-user-content/o/inbox%2F3982638%2F096a19d50a5193beba70344b2ae23307%2Fpathology.jpeg?generation=1594277800870240&alt=media\")
Most deep learning solutions for gigapixel whole slide images break them into patches, make predictions on each, and aggregate the result. But then the context of the whole image is lost.
\n\nWenqi Lu et al. instead represent whole slide images as a graph of cellular architecture and apply a graph convolutional network to predict HER2 and PR status of breast cancer.
\n\nThey recently presented their work at CVPR.
\n\nThey recently presented their work at CVPR.
\n\nPresentation: https://lnkd.in/dbqYXJv
\n\nPaper: https://lnkd.in/dnY5eSC
\n\n
See Linkedin post here
", "messages": [ { "id": 921255, "postDate": "2020-07-09T06:57:38.893Z", "content": "Most deep learning solutions for gigapixel whole slide images break them into patches, make predictions on each, and aggregate the result. But then the context of the whole image is lost.
\n\nWenqi Lu et al. instead represent whole slide images as a graph of cellular architecture and apply a graph convolutional network to predict HER2 and PR status of breast cancer.
\n\nThey recently presented their work at CVPR.
\n\nThey recently presented their work at CVPR.
\n\nPresentation: https://lnkd.in/dbqYXJv
\n\nPaper: https://lnkd.in/dnY5eSC
\n\n
See Linkedin post here
", "rawMarkdown": "Most deep learning solutions for gigapixel whole slide images break them into patches, make predictions on each, and aggregate the result. But then the context of the whole image is lost.\n\nWenqi Lu et al. instead represent whole slide images as a graph of cellular architecture and apply a graph convolutional network to predict HER2 and PR status of breast cancer.\n\nThey recently presented their work at CVPR.\n\nThey recently presented their work at CVPR.\n\nPresentation: https://lnkd.in/dbqYXJv\n\nPaper: https://lnkd.in/dnY5eSC\n\n\n\n\nSee Linkedin post [here](https://www.linkedin.com/posts/hdcouture_pathologyai-deeplearning-activity-6686602281795764224-TLg-/) ", "votes": 6 }, { "id": 921710, "postDate": "2020-07-09T13:49:45.277Z", "content": "did you use it?
", "rawMarkdown": "did you use it?", "replies": [ { "id": 921725, "postDate": "2020-07-09T13:59:53.963Z", "content": "@hiramcho nope, but i am planning to.
", "rawMarkdown": "@hiramcho nope, but i am planning to." }, { "id": 921755, "postDate": "2020-07-09T14:17:29.660Z", "content": "Two more questions that you are free to give no answer, are you going to use it alongside your principal CNN model? and what is your actual CNN model?
", "rawMarkdown": "Two more questions that you are free to give no answer, are you going to use it alongside your principal CNN model? and what is your actual CNN model?" } ] } ], "comments": [ { "id": 921710, "author_name": "Hiram Coria 🧬", "author_url": "", "post_date": "2020-07-09T13:49:45.277000", "content": "did you use it?
", "votes": 0, "replies": [ { "id": 921725, "author_name": "Dracarys", "author_url": "", "post_date": "2020-07-09T13:59:53.963000", "content": "@hiramcho nope, but i am planning to.
", "votes": 0, "replies": [] }, { "id": 921755, "author_name": "Hiram Coria 🧬", "author_url": "", "post_date": "2020-07-09T14:17:29.660000", "content": "Two more questions that you are free to give no answer, are you going to use it alongside your principal CNN model? and what is your actual CNN model?
", "votes": 0, "replies": [] } ] } ], "raw_markdown_by_id": { "921255": "Most deep learning solutions for gigapixel whole slide images break them into patches, make predictions on each, and aggregate the result. But then the context of the whole image is lost.\n\nWenqi Lu et al. instead represent whole slide images as a graph of cellular architecture and apply a graph convolutional network to predict HER2 and PR status of breast cancer.\n\nThey recently presented their work at CVPR.\n\nThey recently presented their work at CVPR.\n\nPresentation: https://lnkd.in/dbqYXJv\n\nPaper: https://lnkd.in/dnY5eSC\n\n\n\n\nSee Linkedin post [here](https://www.linkedin.com/posts/hdcouture_pathologyai-deeplearning-activity-6686602281795764224-TLg-/) ", "921710": "did you use it?" } }