Skip to content
This repository has been archived by the owner on Jun 2, 2022. It is now read-only.

Support Huge Image Conversion #158

Merged
merged 71 commits into from May 23, 2019
Merged

Support Huge Image Conversion #158

merged 71 commits into from May 23, 2019

Conversation

YukihoAA
Copy link
Collaborator

@YukihoAA YukihoAA commented May 11, 2019

This PR will fix #122, closes #157

w2x works like this.

  1. denoise
  2. Get minimum 2^n scale that is bigger then input scale factor.
  3. resize to 2^n scale.
  4. down scale to input scale factor.

I added 2-1. step to prevent memory fault and wrong resizing.

  1. denoise
  2. Get minimum 2^n scale that is bigger then input scale factor.
  3. resize to 2^n scale.

2-1. slice in to 4^n pieces if input image is too big to convert.

  1. down scale to input scale factor.

@YukihoAA YukihoAA changed the title Slicing when limit Support Huge Image Conversion May 11, 2019
@YukihoAA YukihoAA requested a review from DeadSix27 May 11, 2019 14:13
It actually not happens in my test case but just for safety
@YukihoAA YukihoAA mentioned this pull request May 20, 2019
Copy link
Owner

@DeadSix27 DeadSix27 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Make sure to merge #157, #158, #163 and #164 right after this one!! (confusingly many PRs)

#164 will require separate review by you @YukihoAA.

Works fine up until 17k * 17k, everything above crashes silently for me.
Due to it exceeding my memory limit of 16G.

(Still requires a proper error (somehow)))

@YukihoAA YukihoAA merged commit 4aedbd7 into master May 23, 2019
@YukihoAA YukihoAA deleted the slicing_when_limit branch May 23, 2019 12:10
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

Upscaling very large images got crashed result
2 participants