Doing everything according to guide. However, the program does not work properly.
I add image directory, then file with class names. Then names do not appear in the programm and I cannot draw bounding boxes.

Hello,
Thank you for this great tool !
Is it possible to shift or resize an existing annotation without removing it ?
Thank you

Some of the images I loaded are rotated when displaying (maybe due to their size or anything else) which causes positions of boxing boxes to be shifted.

(Cannot send images because of privacy terms)

the max resolution for my screen is 1368x768, hence theres no way for me to reach the buttons at the bottom, sad af, pls help

Hello, thanks for making a cool annotation tool.
We are trying it out but facing an issue with the numbering. when we switch from img#109 to 110 it instead goes to 1001. i cant find the relevant code in source, but would love any pointers or a fix. im on win so compiling is abit iffy aswell.

HI, I wonder what the image coordinate format is.
Also, the coordinates are marked with decimal points, so I wonder if it's normal.

I WILL COME BACK IN JULY.

Visualize Class Name on the bounding box

Is there potential to add functionality to set the image directory and object text file from the command line?

Hi!

First off many thanks for the amazing tool. Very handy and simple to use.

I feel a bit stupid, but I'm not sure how to convert the decimal coordinates produced by the program into pixel coordinates.

All of the coordinate values in my label .txt files are decimals between 0 and 1, as shown below.

I saw in issue #40 that that user tried to convert the decimals with an int() call... however for me this converts all values to 0.

What am I missing here?

Thanks in advance!

Unable to resize the exe file to fit on laptop screen

@developer0hye the Tool is awesome, almost perfect but it would be great if there was no restriction on what size you can mark a bounding box.

It doesn't allow small enough objects when tagging small things on a frame from a 4k video :). I have really small objects!

Thanks!

Hi , a little documentation on the result file would be appreciable.

PS: the software is incredibly simple and powerful 👍

:( However about HCI, some improvements could be:

• "remove" button should not be aside "save" button (deleted some of my samples stupidly and cannot retrieve them, a validation box can be appreciable)
• tips about where " remove" button sends your beloved samples (tooltips on each button can be appreciable)
• " change directory " and " open files " buttons : I don't know the difference and they are too close from "save" button although they are not used so often also clicking on them is not cancelable

saving coords with this code

_> double midX = objBox.box.x() + objBox.box.width() / 2.;

        double midY     = objBox.box.y() + objBox.box.height() / 2.;
        double width    = objBox.box.width();
        double height   = objBox.box.height();_

but in Yolo

Where x, y, width, and height are relative to the image's width and height.

not from mid

while trying to run make command in MacOs Venture 13.2.1 I got the error

Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/clang++ -c -pipe -stdlib=libc++ -O2 -std=gnu++1z  -arch x86_64 -isysroot /Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX13.0.sdk -mmacosx-version-min=10.10 -Wall -W -fPIC -DQT_DEPRECATED_WARNINGS -DQT_NO_DEBUG -DQT_WIDGETS_LIB -DQT_GUI_LIB -DQT_CORE_LIB -I. -I../../opt/anaconda3/include/qt -I../../opt/anaconda3/include/qt/QtWidgets -I../../opt/anaconda3/include/qt/QtGui -I../../opt/anaconda3/include/qt/QtCore -I. -I/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX13.0.sdk/System/Library/Frameworks/OpenGL.framework/Headers -I/Applications/Xcode.app/Contents/Developer/Platforms/MacOSX.platform/Developer/SDKs/MacOSX13.0.sdk/System/Library/Frameworks/AGL.framework/Headers -I. -I../../opt/anaconda3/mkspecs/macx-clang -o label_img.o label_img.cpp
In file included from label_img.cpp:1:
In file included from ./label_img.h:7:
In file included from ../../opt/anaconda3/include/qt/QtGui/QMouseEvent:1:
In file included from ../../opt/anaconda3/include/qt/QtGui/qevent.h:50:
../../opt/anaconda3/include/qt/QtCore/qvariant.h:380:16: warning: definition of implicit copy assignment operator for 'Private' is deprecated because it has a user-provided copy constructor [-Wdeprecated-copy-with-user-provided-copy]
        inline Private(const Private &other) Q_DECL_NOTHROW
               ^
../../opt/anaconda3/include/qt/QtCore/qvariant.h:265:15: note: in implicit copy assignment operator for 'QVariant::Private' first required here
    { other.d = Private(); }
              ^
label_img.cpp:74:48: warning: unused parameter 'ev' [-Wunused-parameter]
void label_img::mouseReleaseEvent(QMouseEvent *ev)
                                               ^
label_img.cpp:312:31: error: use of undeclared identifier 'QColorConstants'
            blackPen.setColor(QColorConstants::Black);
                              ^
label_img.cpp:314:31: error: use of undeclared identifier 'QColorConstants'
            blackPen.setColor(QColorConstants::White);
                              ^
2 warnings and 2 errors generated.
make: *** [label_img.o] Error 1

Do you have any recommendation to solve it

I think a project file would be fine
so there is no need to adjust image file directory and label file every time

Hi,
First of all: great tool, you reinvented the wheel but made it super simple and intuitive.
I am thinking about using Yolo_Label to correct annotations automatically generated by my yolo model. The idea is to have a an automated pre-annotation, then I just need to resize the bounding boxes for perfect annotations. However, this requires a resizing mode, otherwise I will need to delete the "ok" annotation in order to make a perfect one.
Are you planning on developing such an option/mode?

Also, an image zoom would be a great addition to the projet in my opinion.
Thank you for you answer

image files showned without same aspect ratio

Yolo was originally designed for linux so a linux binary would be appreciated

I'm using a W10 computer and whenever I try to choose the directory where the images are, the window freezes and the program shuts down.

Hi!

First, thank you @developer0hye for this repo, this sounds like a really handy alternative of the famous Yolo annotation tools.

I'm trying to use Yolo-Label With some random images and the whole interface seems to freeze with my images.

I have no problem with the same images on Yolo-Mark.

All files are .jpg but some are .jfif rename to .jpg, could this be the pb?

In the zip file below is two images on which the pb occurs.

Is there a quick fix I can set up?

By the way, is there a method to create the "train.txt" file with all the images' names (one per line)?

img2.zip

tool randomly rotates images

look how picture looks like in labeling tool

and in my folder

What a good tool YOLOLABEL is! Hope for a Mac verison.

sometimes .jpg (not .png or .jpeg) images are not displayed on the UI, if there are a lot of images, neither using the slider or the prev next buttons will display the image, the path will be updated tho.

I tried isolating the image what was not being shown, but it doesn't work either:

I am working with relatively small sets (less than 100 images per labeling session) so I don't know if this becomes more often with larger amounts.

although it's kind of random, is there a way to fix this?

thank you.

1. When I run it in debug mode, the dll looks different. Like Qt5cored.dll, d is added. Does this issue affect mouse hangup?

2. Mouse crosses are very slow to follow. And it got worse when the window was maximized.
   The readme's window and the current window look different and updated, but where is the object name and color part cut off the color part? In other words, is it because the slider bar that pushes it to the side is below it?
   (한국인이시면 한국어로 해주시면 감사하겠습니다.)

What type of unit is used to define the bounding box in the .txt file?

I have hundreds of images correctly being saved but there's these 2 jpg images where no matter how I save, the coords are not being added to the txt file:

for some reason, only the, percentages? are being saved. These two images are the only ones where this is happening. I tried with another 600 images(jpg and png) and both fields are being saved. What could it be?

Thank you

This is a great labeling tool with a great UI.

But I can't zoon in/out.

It's so uncomfortable in a picture with a lot of small objects. Could you please add the zoon in/out function?

When the mouse is out of the image boundary , the second click doesn't work, I have to move the mouse back and forth to see the "+" mark and click to complete the label.
Maybe I'm asking too much. But getting the mouse right at the edge is really annoying. The purpose of this tool is to save some time, right?
I could just drag the mouse out of the image while using labelImg.

Hi,
I completed my annotation and had a look at the .txt files. All the values of my bounding boxes start with 0.

For eg :
0 0.485048 0.334155 0.253589 0.059186
1 0.449761 0.529593 0.180622 0.060419
2 0.437799 0.739211 0.153110 0.053021
3 0.550239 0.215783 0.250000 0.054254
This is for a single image with 4 classes.

I used PIL and OpenCV to crop my image using these values considering they are [xmin ymin xmax ymax] and I got the results wrong. When I reopen Yolo_Label, the boxes are perfectly there. Is there anything I need to do? any solution?

The code I used to crop the boxes :

from PIL import Image
import matplotlib.pyplot as plt
bbox=(int(0.485048), int(0.334155), int(0.253589), int(0.059186)) #for one bounding box of an image
im=Image.open('sample.jpg')
im=im.crop(bbox)
plt.subplot(3,3,i+1)
plt.axis("off")
plt.imshow(im)

Hi @developer0hye,
Thank you for this nice and simple labeling tool
Can you add a license to your Yolo_Label project (preferably an open license)?
Thank you.

Bug: label list read white space

Excellent Yolov3 annotation tool, can this tool generate xml for each image annotation object? Do you need Xml to train your image? What is the use of an XML file?

Thanks for the great tool.
It might be convenient to have an option to filter out an image that has at least one label.
In many cases I am adding new images and don't want to skip the ones that are already marked.
It is especially convenient when there is a big amount of images.

TrainYourOwnYOLO requires CSV file as data. Is there any way we can do that with Yolo_Label?