According to Syrian Network For Human Rights (SNHR) reports: Year Casualties
2015 21,179 (civilians including 3,443(Children) killed)
2016 16,913 (civilians including 1,984(Children) killed)
2017 10,204 (civilians including 2,298(Children) killed)
2018 167 (civilians including 48(Children) killed)
This project is focused on saving by identifying innocent children(harmless) from critical situations like war_zones using deep learning and computer vision techniques.A swarm of drones can help for making this possible.
The training dataset for deep learning task consists of images of two classification
1.Harmless(331)- children in the war
2.Harmful (376)-Military weapons, army soldier(on both sides of the war), Army tanks.
Keras is the module used for deep learning algorithm implementation and is trained for 50 epoches.
Result for some test images is shown below
