![\"\"](\"https://www.radiomics.io/img/pyrad1.jpg\")
I saw a lot of notebooks with an end-2-end approach, but in many papers, they are using a feature extraction approach. I found pyradiomics tool. Has anyone already tried that one or maybe other feature extraction tools?
\n
I saw a lot of notebooks with an end-2-end approach, but in many papers, they are using a feature extraction approach. I found pyradiomics tool. Has anyone already tried that one or maybe other feature extraction tools?
\n
I'm just using one of the Timm models. Get one of the high scoring models from this page : https://github.com/rwightman/pytorch-image-models/blob/master/results/results-imagenet-real.csv. Use the forward_feature() method, save the torch tensors to disc using torch.save(), and run them through an RNN.
", "rawMarkdown": "I'm just using one of the Timm models. Get one of the high scoring models from this page : https://github.com/rwightman/pytorch-image-models/blob/master/results/results-imagenet-real.csv. Use the forward_feature() method, save the torch tensors to disc using torch.save(), and run them through an RNN.", "votes": 1, "replies": [ { "id": 1481455, "postDate": "2021-08-19T14:08:31.950Z", "content": "I mean statistical features like the volume of the tumor, the roughness of shape, etc. But your proposed idea also sounds like a good one to give a try. Thanks!
", "rawMarkdown": "I mean statistical features like the volume of the tumor, the roughness of shape, etc. But your proposed idea also sounds like a good one to give a try. Thanks!" }, { "id": 1481497, "postDate": "2021-08-19T14:44:38.373Z", "content": "Hello, I am trying with pyradiomics and although while saving version and producing submission.csv file take 3 hours but when I submitted it takes more than 9 hours and I got notebook timeout error.
\nAny idea how to solve notebook timeout error.
", "rawMarkdown": "Hello, I am trying with pyradiomics and although while saving version and producing submission.csv file take 3 hours but when I submitted it takes more than 9 hours and I got notebook timeout error.\n\nAny idea how to solve notebook timeout error." }, { "id": 1481532, "postDate": "2021-08-19T15:06:24.563Z", "rawMarkdown": "", "isDeleted": true }, { "id": 1481566, "postDate": "2021-08-19T15:25:47.307Z", "content": "Thank you for your reply, I was installing pyradiomics without internet and saved my version without internet. So it works and generate submission file.
\nBut when I submit for public scoring, it takes more than 9 hours.
", "rawMarkdown": "Thank you for your reply, I was installing pyradiomics without internet and saved my version without internet. So it works and generate submission file.\n\nBut when I submit for public scoring, it takes more than 9 hours." }, { "id": 1481692, "postDate": "2021-08-19T16:22:56.147Z", "rawMarkdown": "", "isDeleted": true }, { "id": 1481712, "postDate": "2021-08-19T16:31:45Z", "content": "Yes, I agree with you but previously when I worked with DL models, it took me 4/5 hours to save a version and get the submission.csv file and after submitting for LB scoring, it takes the same amount of time to generate the public score.
\nAlthough, I made modifications to get result with 1 hour 2 min with ML model to generate submission and I submitted again, let's see what happen this time.
", "rawMarkdown": "Yes, I agree with you but previously when I worked with DL models, it took me 4/5 hours to save a version and get the submission.csv file and after submitting for LB scoring, it takes the same amount of time to generate the public score.\n\nAlthough, I made modifications to get result with 1 hour 2 min with ML model to generate submission and I submitted again, let's see what happen this time." } ] }, { "id": 1524978, "postDate": "2021-09-27T03:48:36.187Z", "content": "HI @wfarzana can you let me know how you installed pyradiomics without internet. I tried with some wheels but havent been able to install it
", "rawMarkdown": "HI @wfarzana can you let me know how you installed pyradiomics without internet. I tried with some wheels but havent been able to install it", "replies": [ { "id": 1525098, "postDate": "2021-09-27T06:09:44.660Z", "content": "Hi @pankajattri I am also experimenting with pyradiomics. I was able to install it by installing the following dependencies first:
\n# install dependencies for pyradiomics\n!pip install ../input/your-wheel-dataset/docopt-0.6.2/docopt-0.6.2\n!pip install ../input/your-wheel-dataset/ruamel.yaml.clib-0.2.6-cp37-cp37m-manylinux1_x86_64.whl\n!pip install ../input/your-wheel-dataset/ruamel.yaml-0.17.16-py3-none-any.whl\n!pip install ../input/your-wheel-dataset/pykwalify-1.8.0-py2.py3-none-any.whl\n\n# install pyradiomics\n!pip install ../input/your-wheel-dataset/pyradiomics-3.0.1-cp37-cp37m-manylinux1_x86_64.whl\nYou can download the wheels on your own or use my dataset here.
\nHope it helps you.
I'm just using one of the Timm models. Get one of the high scoring models from this page : https://github.com/rwightman/pytorch-image-models/blob/master/results/results-imagenet-real.csv. Use the forward_feature() method, save the torch tensors to disc using torch.save(), and run them through an RNN.
", "votes": 1, "replies": [ { "id": 1481455, "author_name": "Narek Maloyan", "author_url": "", "post_date": "2021-08-19T14:08:31.950000", "content": "I mean statistical features like the volume of the tumor, the roughness of shape, etc. But your proposed idea also sounds like a good one to give a try. Thanks!
", "votes": 0, "replies": [] }, { "id": 1481497, "author_name": "Wfarzana", "author_url": "", "post_date": "2021-08-19T14:44:38.373000", "content": "Hello, I am trying with pyradiomics and although while saving version and producing submission.csv file take 3 hours but when I submitted it takes more than 9 hours and I got notebook timeout error.
\nAny idea how to solve notebook timeout error.
", "votes": 0, "replies": [] }, { "id": 1481532, "author_name": "", "author_url": "", "post_date": "2021-08-19T15:06:24.563000", "content": "", "votes": 0, "replies": [] }, { "id": 1481566, "author_name": "Wfarzana", "author_url": "", "post_date": "2021-08-19T15:25:47.307000", "content": "Thank you for your reply, I was installing pyradiomics without internet and saved my version without internet. So it works and generate submission file.
\nBut when I submit for public scoring, it takes more than 9 hours.
", "votes": 0, "replies": [] }, { "id": 1481692, "author_name": "", "author_url": "", "post_date": "2021-08-19T16:22:56.147000", "content": "", "votes": 0, "replies": [] }, { "id": 1481712, "author_name": "Wfarzana", "author_url": "", "post_date": "2021-08-19T16:31:45", "content": "Yes, I agree with you but previously when I worked with DL models, it took me 4/5 hours to save a version and get the submission.csv file and after submitting for LB scoring, it takes the same amount of time to generate the public score.
\nAlthough, I made modifications to get result with 1 hour 2 min with ML model to generate submission and I submitted again, let's see what happen this time.
", "votes": 0, "replies": [] } ] }, { "id": 1524978, "author_name": "ImBczVr", "author_url": "", "post_date": "2021-09-27T03:48:36.187000", "content": "HI @wfarzana can you let me know how you installed pyradiomics without internet. I tried with some wheels but havent been able to install it
", "votes": 0, "replies": [ { "id": 1525098, "author_name": "NaN", "author_url": "", "post_date": "2021-09-27T06:09:44.660000", "content": "Hi @pankajattri I am also experimenting with pyradiomics. I was able to install it by installing the following dependencies first:
\n# install dependencies for pyradiomics\n!pip install ../input/your-wheel-dataset/docopt-0.6.2/docopt-0.6.2\n!pip install ../input/your-wheel-dataset/ruamel.yaml.clib-0.2.6-cp37-cp37m-manylinux1_x86_64.whl\n!pip install ../input/your-wheel-dataset/ruamel.yaml-0.17.16-py3-none-any.whl\n!pip install ../input/your-wheel-dataset/pykwalify-1.8.0-py2.py3-none-any.whl\n\n# install pyradiomics\n!pip install ../input/your-wheel-dataset/pyradiomics-3.0.1-cp37-cp37m-manylinux1_x86_64.whl\nYou can download the wheels on your own or use my dataset here.
\nHope it helps you.