You can turn it into an Artificial Intelligence monitoring and surveillance system using our software kit in a few minutes with several easy steps
Setup the camera wherever you want monitoring to take place
Setup camera on your PC or Mac
Download and start our desktop application
Create an account for mobile notifications and event reporting
Download our app for Android or iOS and now you have a smart monitoring system which will be able to track and recognize people and common objects in the vicinity of the camera
There are two main types of the surveillance systems that are widely available today: the ones that are actively monitored by a team of people, and those that are just mostly cameras recording hours upon hours of footage, with a hope of finding the culprit long after the fact, upon reviewing this footage
The first approach requires expensive monitoring fees, usually monthly payments, as well as often hefty setup fees. The second approach is cheap, but is not really useful for anything real-time
What if there was another solution? Shouldn’t there be a possibility, in our current age of exploding Artificial Intelligence and Machine Learning advancement, to solve this problem using these new cutting age technologies?
The Surveyor is intended to be that solution. It will not require people to perform live observation of incoming camera feed, because it will be able to recognize people and common objects all by itself, and report if it sees anything suspicious, based on its configuration parameters. It will not need to have any expensive equipment to be purchased or renting a high-performance (and expensive) cloud infrastructure, because it will be able to run on the customer’s own computer, a PC or Mac
This solution will not be spamming the customer with useless notifications because it detected some motion, because it will be able to differentiate an unexpected human visitor from a squirrel. It will also have ability to be trained to recognize the members of your household, people who work at your business, or possibly neighbors, from those that it does not know, and report you that valuable information, all in real time
The Surveyor system will also be able to provide you with an event log of who was near your property and stopped by the door throughout the day, so that you could tell when your kids arrived home, and whether they have their friends with them. Any cars parked in your driveway while you are at work? You will be able to know about that as well, at a glance
We already have a functional prototype of the Surveyor solution available, and you can download it either from SourceForge, or directly from our website:
https://sourceforge.net/projects/cts-surveyor/
or
http://caerustech-solutions.com/demo/surveyor.zip
The current version will recognize and track a human being, and will be able to classify the person’s face at a sufficiently-close range. It relies on the deep neural net models to do that, which will be expended to add functionality and precision for the production version
The most critical piece of this solution is the desktop application, where the crucial AI-based computations are performed. Running the deep neural network technologies, especially those designed for visual processing space, is computationally expensive, and there are challenges at times to make those run on the regular customer’s hardware, such as home PC We have already been able to shrink the performance overhead to the level that the solftware can easily run on a regular PC, as you can see for yourself trying our prototype at https://sourceforge.net/projects/cts-surveyor/ . As we are adding additional machine learning features, we may need to further improve the performance of that part of the system, however with the continuous improvements in the ML technologies and the hardware performance, our team is confident that this is very achievable even with today’s technologies As for the delivery schedule, while the month of September of 2020 is a bit tight, and as with any challenging software projects, unanticipated complications possible, we strongly believe that this timeline is achievable based on our current progress developing the solution