question

[DANwelsen]

What type of data is input to this network ?

[fhaghighi]

Hi, thanks for your interest in our work. Any type of data can be passed as the input to our network. Our self-discovery process can extract anatomical visual words from medical images, such as CT or MRI. The input to the network is the self-discovered visual words. We have made the visual words extracted from chest CT scans publicly available (download link is available on our Github page, here is the link as well). Our method can be extended to other modalities as well.

[DANwelsen]

Thank you for answering my question, but i still have a question about your code. In pattern_generator_3D, I don't know how to get “prev_coordinates".

[fhaghighi]

You do not need any prev_coordinates the first time you run the code (its default value is set to None). This argument helps the user to generate coordinates at different times. For example, you can run the code to generate 50 random coordinates, which will be saved in a text file. Later, if you want to generate another 50 coordinates, you can utilize the stored text file as the prev_coordinates to save time while keeping the distance threshold between the coordinates. However, you can generate all coordinates in a single run, so in this case, you do not need to prev_coordinates at all.

[DANwelsen]

I'm sorry to bother you again.After training ,when I run "pattern_generator" this code:

this need prev_coordinates , but if it's none, it can not run the code under "if" statement.

------------------ 原始邮件 ------------------ 发件人: "fhaghighi/TransVW" @.>; 发送时间: 2021年10月14日(星期四) 下午3:20 @.>; @.@.>; 主题: Re: [fhaghighi/TransVW] question (#4)

You do not need any prev_coordinates the first time you run the code (its default value is set to None). This argument helps the user to generate coordinates at different times. For example, you can run the code to generate 50 random coordinates, which will be saved in a text file. Later, if you want to generate another 50 coordinates, you can utilize the stored text file as the prev_coordinates to save time while keeping the distance threshold between the coordinates. However, you can generate all coordinates in a single run, so in this case, you do not need to prev_coordinates at all.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe. Triage notifications on the go with GitHub Mobile for iOS or Android.

[fhaghighi]

You just need to run the pattern_generator.py with the number of the coordinates that you want. If you do not have the previ_coordinates, that is fine.