Hey guys, I am a newbie. I wasn't able to figure out a way to convert these files into data columns to train my model, though I was able to read dicoms files. I want some insights on this.
", "messages": [ { "id": 936717, "postDate": "2020-07-20T13:20:46.117Z", "content": "You can extract important features from the scans, for example lung volume, chest size, lung surface etc. there was a great discussion on it by one of the experts of the domain. Discussion 1 and Discussion 2
", "rawMarkdown": "You can extract important features from the scans, for example lung volume, chest size, lung surface etc. there was a great discussion on it by one of the experts of the domain. [Discussion 1](https://www.kaggle.com/c/osic-pulmonary-fibrosis-progression/discussion/165727) and [Discussion 2](https://www.kaggle.com/c/osic-pulmonary-fibrosis-progression/discussion/166123)", "votes": 1 }, { "id": 936194, "postDate": "2020-07-20T04:09:03.560Z", "content": "Hey guys, I am a newbie. I wasn't able to figure out a way to convert these files into data columns to train my model, though I was able to read dicoms files. I want some insights on this.
", "rawMarkdown": "Hey guys, I am a newbie. I wasn't able to figure out a way to convert these files into data columns to train my model, though I was able to read dicoms files. I want some insights on this.", "votes": 1 }, { "id": 936245, "postDate": "2020-07-20T05:07:18.213Z", "content": "you will need a 3D convolutional neural network to learn from the DCM files also you will first need to convert these files into NumPy array of a suitable to size so that the model can handle it
", "rawMarkdown": "you will need a 3D convolutional neural network to learn from the DCM files also you will first need to convert these files into NumPy array of a suitable to size so that the model can handle it ", "votes": 2, "replies": [ { "id": 939677, "postDate": "2020-07-22T11:29:05.727Z", "content": "thanks @zainahmad
", "rawMarkdown": "thanks @zainahmad " } ] }, { "id": 936578, "postDate": "2020-07-20T10:27:01.920Z", "content": "Why do you need a 3D neural networks?
", "rawMarkdown": "Why do you need a 3D neural networks?", "replies": [ { "id": 936700, "postDate": "2020-07-20T13:04:53.060Z", "content": "You don't need the whole network to be 3D just the convolutional part needs to be 3D.. because the dicom data is in 3D each folder contains multiple DCM files and each file represents a single scan.. the whole scan is formed by stacking each of these together.. thus forming a 3d volume
", "rawMarkdown": "You don't need the whole network to be 3D just the convolutional part needs to be 3D.. because the dicom data is in 3D each folder contains multiple DCM files and each file represents a single scan.. the whole scan is formed by stacking each of these together.. thus forming a 3d volume" } ] }, { "id": 936246, "postDate": "2020-07-20T05:07:18.213Z", "rawMarkdown": "", "isDeleted": true } ], "comments": [ { "id": 936717, "author_name": "Pranav Kasela", "author_url": "", "post_date": "2020-07-20T13:20:46.117000", "content": "You can extract important features from the scans, for example lung volume, chest size, lung surface etc. there was a great discussion on it by one of the experts of the domain. Discussion 1 and Discussion 2
", "votes": 1, "replies": [] }, { "id": 936245, "author_name": "Zain Ahmad", "author_url": "", "post_date": "2020-07-20T05:07:18.213000", "content": "you will need a 3D convolutional neural network to learn from the DCM files also you will first need to convert these files into NumPy array of a suitable to size so that the model can handle it
", "votes": 2, "replies": [ { "id": 939677, "author_name": "Kartik Parsoya", "author_url": "", "post_date": "2020-07-22T11:29:05.727000", "content": "thanks @zainahmad
", "votes": 0, "replies": [] } ] }, { "id": 936578, "author_name": "Kurian Benoy", "author_url": "", "post_date": "2020-07-20T10:27:01.920000", "content": "Why do you need a 3D neural networks?
", "votes": 0, "replies": [ { "id": 936700, "author_name": "Zain Ahmad", "author_url": "", "post_date": "2020-07-20T13:04:53.060000", "content": "You don't need the whole network to be 3D just the convolutional part needs to be 3D.. because the dicom data is in 3D each folder contains multiple DCM files and each file represents a single scan.. the whole scan is formed by stacking each of these together.. thus forming a 3d volume
", "votes": 0, "replies": [] } ] }, { "id": 936246, "author_name": "", "author_url": "", "post_date": "2020-07-20T05:07:18.213000", "content": "", "votes": 0, "replies": [] } ], "raw_markdown_by_id": { "936717": "You can extract important features from the scans, for example lung volume, chest size, lung surface etc. there was a great discussion on it by one of the experts of the domain. [Discussion 1](https://www.kaggle.com/c/osic-pulmonary-fibrosis-progression/discussion/165727) and [Discussion 2](https://www.kaggle.com/c/osic-pulmonary-fibrosis-progression/discussion/166123)", "936194": "Hey guys, I am a newbie. I wasn't able to figure out a way to convert these files into data columns to train my model, though I was able to read dicoms files. I want some insights on this.", "936245": "you will need a 3D convolutional neural network to learn from the DCM files also you will first need to convert these files into NumPy array of a suitable to size so that the model can handle it ", "936578": "Why do you need a 3D neural networks?", "936246": "" } }