sowmyakavali / Generate-annotations

Generate-annotations

Its just a simple opencv project to directly generate images and annotations from video

Process

Create Virtual environment

$python -m venv venv
$venv\Scripts\activate
$pip install -r requirements.txt

Run

python main.py -v {path to video} -l {class name}

Ex : python main.py -v video.mp4 -l car

> Here you can able to annotate for only one class at a time
> After video opened , Enter 
        s key to pause that frame and 
        draw rectanle box on the frame where the object is present,after then press Enter key , 
        again draw another box and press enter if there are multiple objects. 
> Press esc key after you completed marking all objects in that particular frame. 
> Repeat the step 2 and 3.    

Output

Generates image and .xml file, which having the bounding boxes.