Vasudev Sharma

Results 43 comments of Vasudev Sharma

> Following up on @Stoyan-I-A 's instructions for running ADS with FSLeyes on a Windows computer (see here), I'm wondering if it may be worth it to create and share...

Shall we first deal with PR #399 (implement ellipse minor axis) or first towards PR #179 which add tests for generating synthethic images of axons for the testing purpose?

My PR #441 handles the naming convention for FSLeyes with an agreed-upon naming convention as discussed in issue #439. But it doesn't cater to the scenarios when a user wishes...

Yes @Stoyan-I-A, I was thinking about saving for every visible mask on the FSLeyes canvas. But you have a good point regarding keeping all the masks visible and saving them...

Correct me if I'm wrong? The `download_data ` function in [ads_utils.py](https://github.com/neuropoly/axondeepseg/blob/master/AxonDeepSeg/ads_utils.py) script first downloads the models and saves them in a temporary folder in the working directory(`axondeepseg`), and then it...

Yes, it is. The main reason for this behaviour is that we are using harcoded paths in the [download_model.py ](https://github.com/neuropoly/axondeepseg/blob/master/AxonDeepSeg/models/download_model.py) https://github.com/neuropoly/axondeepseg/blob/e2cd6fcc7502db6241cbea4b7b47e1201599315d/AxonDeepSeg/models/download_model.py#L8-L9 When cli command runs in non root directory, it...

I can open a new PR to resolve this issue or can resolve it in this [PR](https://github.com/neuropoly/axondeepseg/pull/387). Let me know what do you think.

@mathieuboudreau or @mariehbourget could either of you review this PR?

One question, while using the CLI command to download the models, should we create destination folder if the destination folder is non existent?

@mathieuboudreau , I've added some unit tests for the changes introduced in [download_model.py](https://github.com/neuropoly/axondeepseg/blob/master/AxonDeepSeg/models/download_model.py) script. Feel free to review them. Also, I haven't included the unit test for the issue #394,...