Syn-ISS Participants,
The instructions on how to prepare your Docker container images for the challenge can be found on the [Submissions page](https://www.synapse.org/#!Synapse:syn50908388/wiki/623401).
We have provided all information related to submissions on a GitHub repository linked in the page above.
The Docker image submissions are due **September 4, 2023**.
Please bear with us as we enter this final leg of the challenge. This is our first time doing this!
Please do reach out to us on this discussion forum or email us at EndoVis-DS@surgicalscience.com with any questions or comments.
Things coming next:
* Report writeup instructions
* Evaluation script that we would be using for assessing models
Team registration closes on **August 31** - please make sure to register your team if you have not done that yet.
Cheers,
Syn-ISS Organizing Team.
Created by Anand Malpani anand_sus Hi Anand,
Thank you for your kind instruction.
I found that the file size of the sample images which I downloaded were about 120B.
It seems that the download process didn't work properly.
I'll try that again.
Thank you!
Taiyo Taiyo, thanks for posting your question here. That is strange. The good thing is that you are able to read in other images.
One possibility is that when getting the files from GitHub the git-lfs didn't work properly for you and the image files are corrupted. What is the size of the image file? The scene image files should be around 600 KB.
I am able to open the images using `skimage` method `skimage.io.imread`.
For example, here is the output of my Python shell reading this [sample image](https://github.com/SurgicalScience/syn-iss-2023/blob/main/docker/templates/task1-binary/sample-test-data/inputs/s-017b418851a5bb01edd7.png):
```
Python 3.10.9 | packaged by conda-forge | (main, Jan 11 2023, 15:15:40) [MSC v.1916 64 bit (AMD64)] on win32
Type "help", "copyright", "credits" or "license" for more information.
>>> import skimage.io as sio
>>> img = sio.imread("s-017b418851a5bb01edd7.png")
>>> img.shape
(540, 960, 4)
>>>
```
Let me know if you still have issues. Dear Anand,
I couldn't open the sample images by neither opencv, PIL, imageio and skimage.
However, I can open the train and test images, which are included in the datasets.
I wonder these sample images have different format compared to the images in the datasets.
Will you kindly tell me the way to open them?
Best regards,
Taiyo Ishikawa Hi anand,
Thanks for our quick response. Thanks, I got the answer.
I did not find the any shape 540x960x3 in input data.
Thanks Hi @abdulkcl,
Input images are of the shape (540,960,4) in the datasets released so far as well as the sample images in the Docker template. We will keep it consistent for the test dataset as well.
Where did you find a 540 x 960 x 3 input image? Hello,
Thanks for sharing the docker. Could you check the sample images inside the sample-test-data to test our model in docker container? I could not successfully read these input sample images. I just want to confirm the input image size. It should be ((540, 960, 4) or (540, 960, 3))
Thanks.
Drop files to upload
Docker submission instructions released! page is loading…