README.md

Project HADES ( Hostility Analysis for Defence by Empirical Surveillance )

An AI-powered monitoring solution for defence surveillance.

Defence monitoring video feeds are continuously monitored by humans which is cumbersome and inefficient employment of manpower. The legacy COTS (Commercially-off-the-shelf) systems currently used to achieve battlefield transparency, need to be continuously manned and monitored. This poses a problem of inefficient management for manpower, for resource manipulation. Thus, an innovative adaptation is required to help manage human resources, without compromising on the level of security deployed

How To Run:

• Dataset/soldiers - contains our custom COCO dataset of training, testing and validation dataset images used.

• mrcnn folder - Contains the Original Mask RCNN Library

• Run requirements.txt first

• Download our trained model from the link: https://drive.google.com/uc?export=download&id=1g0hYm7wSPOm0IJSkWo8sOk4TKZPVTc1u

• Place this model file in the root directory

• Run all cells in validation.ipynb

You can see that we have got high accuracy detection using Mask RCNN with very low training images.

With More images to train and with video detection using opencv, we can use this model for continous surveillance and get much higher results than expected.