A simple Python code that recognize the faces of humans, and rarely, faces of ghosts. More features coming soon as I continue to study the OpenCV documentation.

This is a project by Lin as a way to challenge himself to create a facial detection software within 2 days. As he is typing this right now, he is currently under immense headache from the back all the way to the front. If you wish to add more features into this project, you are free to contribute.

To use the code for your personal project, follow these easy steps (read everything first before doing so):

1. Crop and copy the images of the faces of people you want the computer to identify to the resx folder. Using different folders, separate the images to their corresponding folder names (use their real name, or not; depends if you want to make a fool of somebody by labelling them something else). The name of the folders will be used as labels. The name of the images doesn't matter.
2. Run os-walk.py. This will create the trainer.yml and labels.pickle files. These files contain the data necessary for person identification for main.py.
3. Run main.py and enjoy!
4. coco-detect.py is for object detection. It utilizes the files frozen_inference_graph.pb and ssd_mobilenet_v3_large_coco_2020_01_14.pbtxt, so it doesn't require steps #1-#3. You can simply run this as is!