$ cd my/directory/of/choice
$ git clone https://github.com/krisskad/rmLogo.git

$ cd path/to/rmLogo
$ python -m pip install -r requirements.txt

Computer vision program to watermark and logos. Utilises computer
vision and image analysis to return approximated feature dimensions.

positional arguments:
  input_dir             The path of the image folder which containts images [JPG, PNG].
  output_dir            The path of the output folder to save output images [It will store the image with same name as original image]

optional arguments:
  single_image          You can provide a single image to test the model [just provide the image path and output_dir and make other as None]
  kernel_size 2         To perform smoothing process we need odd values [the greather the value more soft will be result. It will damage quality]


def process_image(input_dir: Any = input_dir,
               output_dir: Any = output_dir,
               single_image: Any = None,
               kernel_size: Any = None)

from main import process_image
# run process_image function with given parameters
process_image("E:\project\rmLogo\INPUT_DIR", "E:\project\rmLogo\OUTPUT_DIR", None, (5,5))

* Remove noise and watermark
* Remove logo using yolo object detection and masking
* Apply some post processing to increase quality