add click-drag-resizing - Githubissues

Cartucho / OpenLabeling

Label images and video for Computer Vision applications

Apache License 2.0

926 stars 266 forks source link

add click-drag-resizing #38

Closed vuthede closed 5 years ago

vuthede commented 5 years ago

In Each bbox, I create some small resizing-anchors: TL(Top Left), TR, BL, BR, TM(Top Middle), BM, LM, RM. Currently, I handle click-drag-resizing on 2 mouse events:

First: Left Mouse Down. To check what bbox is chosen to be dragged and what on what resizing-anchor
Second: Left Mouse Up . Return new size of chosen bbox after dragging.

Cartucho commented 5 years ago

Great! Let's work on this step-by-simple-step.

I will have a look at this initial commit!

vuthede commented 5 years ago

Thanks

Cartucho commented 5 years ago

Hey! Do you think I can commit to this PR? And make some modifications?

vuthede commented 5 years ago

Yeah! Go ahead bro!

Cartucho commented 5 years ago

I'm currently working on:

drawing the anchors
drawing/updating the bbox when dragged (so that one knows that the bbox is being resized)
merging the function edit_bbox and resize_bbox into a single function.

I will commit as soon as I have it ready

vuthede commented 5 years ago

Great!

Cartucho commented 5 years ago

Hey, could you please Allow edits from maintainers? https://help.github.com/articles/allowing-changes-to-a-pull-request-branch-created-from-a-fork/

vuthede commented 5 years ago

I think i allowed edits from maintainer by checking the check box as the link u sent

Cartucho commented 5 years ago

I managed to simplify/reduce the lines of code and add the new features! There is nothing like teamwork!

I will only do some final cleaning and then it will be ready to be merged.

Cartucho commented 5 years ago

Done! Please test it a bit and if you agree let's merge it.

vuthede commented 5 years ago

Yeah, I have tested, I look so cool. Thanks a lot. There is a minor "bug" causing by different verison of python and opencv i think. I used python 3.6 and opencv4:

First: In python 3, the tuple parameters of function are removed https://www.python.org/dev/peps/pep-3113/ . So i changed draw_bbox_anchors(tmp_img, (xmin, ymin), (xmax, ymax), color) into draw_bbox_anchors(tmp_img, xmin, ymin, xmax, ymax, color).
Second: cv2.retangle(img, (x1, y1), (x2,y2), -1) takes x1, y1, x2, y2 as integer type. I also cast it into integer Is there any different with your version of opencv and python?

if u agree, i will push a commit for abovementioned changes and we will merge it