Windows & Linux GUI for marking bounded boxes of objects in images for training Yolo v3 and v2

• To compile on Windows open yolo_mark.sln in MSVS2013/2015, compile it x64 & Release and run the file: x64/Release/yolo_mark.cmd. Change paths in yolo_mark.sln to the OpenCV 2.x/3.x installed on your computer:

  • (right click on project) -> properties -> C/C++ -> General -> Additional Include Directories: C:\opencv_3.0\opencv\build\include;

  • (right click on project) -> properties -> Linker -> General -> Additional Library Directories: C:\opencv_3.0\opencv\build\x64\vc14\lib;

• To compile on Linux type in console 3 commands:

  cmake .
  make
  ./linux_mark.sh
  

Supported both: OpenCV 2.x and OpenCV 3.x

1. To test, simply run

• on Windows: x64/Release/yolo_mark.cmd
• on Linux: ./linux_mark.sh

2. To use for labeling your custom images:

• delete all files from directory x64/Release/data/img
• put your .jpg-images to this directory x64/Release/data/img
• change numer of classes (objects for detection) in file x64/Release/data/obj.data: https://github.com/AlexeyAB/Yolo_mark/blob/master/x64/Release/data/obj.data#L1
• put names of objects, one for each line in file x64/Release/data/obj.names: https://github.com/AlexeyAB/Yolo_mark/blob/master/x64/Release/data/obj.names
• run file: x64\Release\yolo_mark.cmd

3. To training for your custom objects, you should change 2 lines in file x64/Release/yolo-obj.cfg:

• set number of classes (objects): https://github.com/AlexeyAB/Yolo_mark/blob/master/x64/Release/yolo-obj.cfg#L230
• set filter-value
  • For Yolov2 (classes + 5)*5: https://github.com/AlexeyAB/Yolo_mark/blob/master/x64/Release/yolo-obj.cfg#L224
  • For Yolov3 (classes + 5)*3

3.1 Download pre-trained weights for the convolutional layers (76 MB): http://pjreddie.com/media/files/darknet19_448.conv.23

3.2 Put files: yolo-obj.cfg, data/train.txt, data/obj.names, data/obj.data, darknet19_448.conv.23 and directory data/img near with executable darknet-file, and start training: darknet detector train data/obj.data yolo-obj.cfg darknet19_448.conv.23

For a detailed description, see: https://github.com/AlexeyAB/darknet#how-to-train-to-detect-your-custom-objects

To get frames from videofile (save each N frame, in example N=10), you can use this command:

• on Windows: yolo_mark.exe data/img cap_video test.mp4 10
• on Linux: ./yolo_mark x64/Release/data/img cap_video test.mp4 10

Directory data/img should be created before this. Also on Windows, the file opencv_ffmpeg340_64.dll from opencv\build\bin should be placed near with yolo_mark.exe.

As a result, many frames will be collected in the directory data/img. Then you can label them manually using such command:

• on Windows: yolo_mark.exe data/img data/train.txt data/obj.names
• on Linux: ./yolo_mark x64/Release/data/img x64/Release/data/train.txt x64/Release/data/obj.names

• /x64/Release/

  • yolo_mark.cmd - example hot to use yolo mark: yolo_mark.exe data/img data/train.txt data/obj.names
  • train_obj.cmd - example how to train yolo for your custom objects (put this file near with darknet.exe): darknet.exe detector train data/obj.data yolo-obj.cfg darknet19_448.conv.23
  • yolo-obj.cfg - example of yoloV3-neural-network for 2 object

• /x64/Release/data/

  • obj.names - example of list with object names
  • obj.data - example with configuration for training Yolo v3
  • train.txt - example with list of image filenames for training Yolo v3

• /x64/Release/data/img/air4.txt - example with coordinates of objects on image air4.jpg with aircrafts (class=0)