How we are able to detect activity and objects
MIDL is a search platform for video surveillance. This means that our software is constantly searching the video streams from your existing cameras for any activity that you are interested in. If something you want alerted for is detected, you and/or the police will receive a notification. All other activity is cataloged for you to instantly surface whenever you would like. All you have to do is tell MIDL what you are looking for, either by typing or speaking to it.
Our power lies in being able to detect a wide range of objects and activity from your video streams (let’s define this as data) and then presenting that data to you in an extremely consumable way. This includes alerts, analytic tiles, and/or surfacing clips of activity you are interested in. This enables you and your staff to benefit from your video data in a way you have never been able to before.
So how do we enable this wide range of detections in the first place. The traditional method of enabling detections is to gather a large amount of images and train a model for a particular object or activity, such as weapons. While we still engage in training models in this traditional way, we have figured out a more efficient method of training that enables us to have a much wider range of detections than others.
Past this point of initial training, we train at your property and continually improve on your unique environment. This means that the product you get on day 1 will not be the same level of product you have on day 365. MIDL will constantly improve. If you have questions regarding our detections and how they can work for your use cases, feel free to reach out to us using the button below.