NOTE: Please download the "resnet50_coco_best_v2.1.0.h5" model from https://github.com/OlafenwaMoses/ImageAI/releases/download/1.0/resnet50_coco_best_v2.0.1.h5 and place it in the root folder (this could not be done currently due to the size of the model). This will be integrated into the application to automatically download in the future.

Download the repository and place the images of people you want to be recognized inside "knwon_people_images" folder. Some sample images are already present. Preferably the images should be front facing. The images should contain only 1 person.

There are 3 utilities inside this zip file: memcator.insert

This utility builds the database and takes file names as arguments and can be best searched using the following command:

find ./images/. -exec ./memcator/memcator/memcator.insert {} +;

memcator.search This utility searchs inside the database for the relevant objects. The '+' is an AND operator and the '-' is a NOT operator for the prediactes. Always perform subsequent AND '+' operations first before performing the NOT '-' operations. Following are some example commands:

memcator/memcator.search ameer memcator/memcator.search bottle memcator/memcator.search ameer+bottle

memcator.delete This utility removes all indexes done in the database. memcator/memcator.delete

===================================================== Followinglibraries were used in Python 3.8 deepface==0.0.75 imageai==2.1.6 opencv_python==4.5.5.64 pandas==1.5.1 retina_face==0.0.12 retinaface==1.1.1