So what does Artificial Intelligence software actually do? 

Specifically Computer Vision?

Computer Vision is an area of computer science that enables a computer to extract information from an image. Simplest terms computer vision is showing a computer how to see or in vision like a human being. Athena Gun Detection artificial intelligence has “taught” computers what a gun looks like from lots of angles and distances of images, and how it is used by a human. We talk about our system’s capability as understanding both “objects and behavior” to examine frames coming through surveillance video feeds. To improve results, the system needs multiple frames to increase accuracy. The machine processes multiple frames in under a few seconds, and produces an alarm when the model detects the threat. 

What to know more about computer vision read on:

Getting your computer to look at objects like a human some say is magic.   How exactly do our brains work? Since we have not concluded on how the brain and eyes process images in the world of humans, it’s difficult to say how well the algorithms used in production approximate our own internal mental processes.  How can we get computers to calculate using algorithms to see? The reality is that there are very few working and comprehensive theories of brain computation; so despite the fact that Neural Nets are supposed to “mimic the way the brain works,” nobody is quite sure if that’s actually true.

Computer vision is all about pattern recognition. So one way to train a computer how to understand visual data is to feed it lots images, some say more images are better and some say less images. This all depends on what computer vision algorithm you are using.  At Athena Security we thought how does the human mind work when looking for a gun and trying to detect one,  well simple. take an image and when we find what looks like a gun we look closer at the subject area.  Well that is actually what we got the computer doing to detect a gun.    We make the computer look at the whole image and then when we find the part of the image that is of interest we have the computer look at it more closely using a pixel algorithm to look at the what we think is a gun.  Yes it this process to detect a gun is slower but you are only looking at a small object 100 pixels instead of millions of pixels etc.