I was making the EDA using the .csv files and all of that but now i want to train the model, as my limited knowledge i was thinking about using the dataframe of the .csv files and concatenate the images in a tensor to be introduced in my model as an input, but i dont know how to make the relation between the cancer label to the image with cancer, maybe is a misunderstanding of the theory about deep learning models, if there is someone who knows how to do that or can clarify it i would be so grateful (im learning about DL models).
", "messages": [ { "id": 2117990, "postDate": "2023-01-27T17:33:52.007Z", "content": "Sure, the provided train.csv gives you a cancer column which contains a bunch of 0's and 1's for negative and positive cases. It also gives you other identifying information such as image_id and patient_id.
The image file locations are as such: train_images/<patient_id>/<image_id>.dcm
What you need to do is map the cancer scores from the train.csv to the correct images.
I highly recommend using a generator rather than pre-loading the entire image dataset onto your ram. Look into image_dataset_from_directory for Keras. Depending on what packages you are using this may be beneficial.
", "rawMarkdown": "Sure, the provided `train.csv` gives you a `cancer` column which contains a bunch of 0's and 1's for negative and positive cases. It also gives you other identifying information such as `image_id` and `patient_id`.\n\nThe image file locations are as such: `train_images/Thanks for the info, i will try that.
", "rawMarkdown": "Thanks for the info, i will try that.", "votes": 1 } ] }, { "id": 2117928, "postDate": "2023-01-27T16:45:40.883Z", "content": "I was making the EDA using the .csv files and all of that but now i want to train the model, as my limited knowledge i was thinking about using the dataframe of the .csv files and concatenate the images in a tensor to be introduced in my model as an input, but i dont know how to make the relation between the cancer label to the image with cancer, maybe is a misunderstanding of the theory about deep learning models, if there is someone who knows how to do that or can clarify it i would be so grateful (im learning about DL models).
", "rawMarkdown": "I was making the EDA using the .csv files and all of that but now i want to train the model, as my limited knowledge i was thinking about using the dataframe of the .csv files and concatenate the images in a tensor to be introduced in my model as an input, but i dont know how to make the relation between the cancer label to the image with cancer, maybe is a misunderstanding of the theory about deep learning models, if there is someone who knows how to do that or can clarify it i would be so grateful (im learning about DL models).", "votes": 2 } ], "comments": [ { "id": 2117990, "author_name": "AbdouE", "author_url": "", "post_date": "2023-01-27T17:33:52.007000", "content": "Sure, the provided train.csv gives you a cancer column which contains a bunch of 0's and 1's for negative and positive cases. It also gives you other identifying information such as image_id and patient_id.
The image file locations are as such: train_images/<patient_id>/<image_id>.dcm
What you need to do is map the cancer scores from the train.csv to the correct images.
I highly recommend using a generator rather than pre-loading the entire image dataset onto your ram. Look into image_dataset_from_directory for Keras. Depending on what packages you are using this may be beneficial.
", "votes": 1, "replies": [ { "id": 2119572, "author_name": "Alex Luna", "author_url": "", "post_date": "2023-01-28T22:38:51.630000", "content": "Thanks for the info, i will try that.
", "votes": 1, "replies": [] } ] } ], "raw_markdown_by_id": { "2117990": "Sure, the provided `train.csv` gives you a `cancer` column which contains a bunch of 0's and 1's for negative and positive cases. It also gives you other identifying information such as `image_id` and `patient_id`.\n\nThe image file locations are as such: `train_images/