I want to create my own dataset for YOLOv3 Object Detection. It looks like there is a huge set of offline tools for marking bounded boxes, for example: Yolo_mark, Microsoft VoTT, LabelImg. Is there any zero-config online tool, which will work right in my web browser? Also It would also be convenient to be able to crop ROI of the image right in the process of marking along with bounding boxes.
I could not find a tool that fit the search criteria, so I decided to write it myself. It published by direct link in Github Pages. A killer feature of this tool is support for touch screen input. ROI cropping with automatic shift of bounding boxes is also implemented.
Source code is published here. The tool uses a minimum of third-party dependencies: only Typescript compiler is required for deployment. Bounding boxes will be written when saved onto txt files by darknet framework format.