Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Feature Request : Simple Image Tweaking / Filter #207

Open
rasyidf opened this issue Sep 23, 2021 · 5 comments
Open

Feature Request : Simple Image Tweaking / Filter #207

rasyidf opened this issue Sep 23, 2021 · 5 comments
Assignees
Labels
ms-beta Part of make-sense beta efforts new-feature New feature or request

Comments

@rasyidf
Copy link
Contributor

rasyidf commented Sep 23, 2021

Is your feature request related to a problem? Please describe.
Would it be better if we can change the brightness, gamma, contrast, or just make it grayscale to manipulate images. make it session / project based, that can persist when selected image changed.

Describe the solution you'd like
Something like drawer or panel that can manipulate base image tweaking. Brightness, Contrast, Shadow or Highlight.

it would be great if we can preprocess the image before annotating, that'd be another feature request

Describe alternatives you've considered

Additional context
image

PS: Sorry, i used your profile

@SkalskiP
Copy link
Owner

Hello, again @rasyidf 👋 This is such a good idea 😮 ! It is such a cool feature and something that we could be proud to have. However, we need to plan it in more detail. Questions that immediately come to my mind:

  1. Do we want to preserve the result of that operation or is it just temporal to make it easier to label? If we want to preserve results this task would be dependant on saving my continuous progress so i won't have to start over. #169 as we, first of all, need to have a way to save and load not only labels but the whole project. I myself think those operations should be temporal, especially at the beginning to keep things simple and not overcomplicate implementation.
  2. Do we want to allow users to combine transformations? Can a single image have changed brightness and gamma at the same time?
  3. Do we want to allow users to input the level of transformation via some scroll or number input?
  4. Do we want to allow users for per image transformations?
  5. Where those processing tools should be located in the UI? I think they should be easily accessable so I think that they can be placed in the left or right sidebar. Those bars are designed in a way that they can have multiple tabs inside a single sidebar. For now, the left sidebar contains only the images tab and the right only labels tab, but we can add new tabs quite easly.

@SkalskiP SkalskiP self-assigned this Sep 24, 2021
@SkalskiP SkalskiP added ms-beta Part of make-sense beta efforts new-feature New feature or request labels Sep 24, 2021
@SkalskiP SkalskiP added this to To do in make-sense beta via automation Sep 24, 2021
@SkalskiP
Copy link
Owner

@rasyidf we should also think about task prioritization. We have lots of new ideas, and that is great! But we need to focus and decide which of them are most crucial and have the biggest impact and start with those. I was thinking maybe we can use https://github.com/SkalskiP/make-sense/discussions to do that?

@rasyidf
Copy link
Contributor Author

rasyidf commented Sep 24, 2021

you're right, we should prioritize more on completing beta phases. and discussion board would be a great idea. let's discuss more there.

@arronsullivan
Copy link

This would be very helpful. Specifically just allowing basic min/max contrast settings. Just to make it easier to label, not to be retained.
Fiji/ImageJ has a nice 'auto' contrast where it sets the max such that 5% of pixels are saturated I believe, a very nice shortcut to see dim images.
Another 'advanced' setting would be to apply a user set min/max contrast to all images as they are opened/viewed for speed.

Someone has implemented this into labelImg but has not shared it anywhere that I can see.

Thank you, make-sense looks very nice

@SkalskiP
Copy link
Owner

Hello @arronsullivan 👋 ! Thank you very much for your opinion. I'm happy to see that this issue is getting more traction as I feel it would be a very nice addition to make sense. I think we will gradually add more operations. Start with something small like the contrast.

@SkalskiP SkalskiP pinned this issue Jul 29, 2022
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
ms-beta Part of make-sense beta efforts new-feature New feature or request
Projects
Development

No branches or pull requests

3 participants