This script allows you to create custom haar cascades with ease.
This script is heavily influenced from the Youtube video https://www.youtube.com/watch?v=jG3bu0tjFbk by "sentdex". I learned creating a haar cascade from that video series itself. I highly recommend you to watch it before using this tool. This tool is just used to semi automate the processes sentdex did on the video series.
input_images - This folder will contain the image(s) of the object you want to detect. Try keeping them small in size
input_resized_images - This folder will contain the resized image(s) of the object you want to detect.
negatives - This folder will contain the negative images that will be downloaded from image-net.org.
output_cascade - This folder will contain the cascade that is created using opencv_traincascade command.
positives - This folder stores the created positive samples. The positive samples are created using the images in input_resized_images and negatives using opencv_createsamples command.
uglies - Put one of each ugly images here to remove them from negatives folder.
python createCustomHaarCascade.py