Open-Source Threat Detection
We have developed an open-source threat detection application and released the code for free download. This is available for schools, communities, or anyone wishing to extend the code into their own use.
Open-Source Threat Detection
We have developed an open-source threat detection application and released the code for free download. This is available for schools, communities, or anyone wishing to extend the code into their own use.
Goal & Purpose
Iterate.ai has open-sourced a powerful Weapons Detection AI software. The AI was trained on about 100 live guns, plus 20,000 videos of robberies and threats involving weapons. Our engineers taught the AI to detect guns, knives, kevlar vests, and robbery masks.
As a company of AI and digital professionals, we want to help our communities, however we can. For this situation, network admins and developers from non-profits -- schools and religious institutions -- can install this Weapons Detection AI. It can operate on cameras that were previously installed for security. No software license fees from us will be incurred. We've removed a few important financial barriers.
Our hope is that software developers, UX designers, and AI engineers from around the world pitch in to make the detection even better. Together, we can add another layer of protection for students and teachers in our schools, members of religious organizations, and other non-profits.
Get the software at GitHub. If you're a non-profit, we invite you to use it for free. If you're a technologist, jump in and improve the project as a contributor to the open-source community.
GitHub Code is Here
As a company of AI and digital professionals, we want to help our communities, however we can. For this situation, network admins and developers from non-profits -- schools and religious institutions -- can install this Weapons Detection AI. It can operate on cameras that were previously installed for security. No software license fees from us will be incurred. We've removed a few important financial barriers.
Our hope is that software developers, UX designers, and AI engineers from around the world pitch in to make the detection even better. Together, we can add another layer of protection for students and teachers in our schools, members of religious organizations, and other non-profits.
Get the software at GitHub. If you're a non-profit, we invite you to use it for free. If you're a technologist, jump in and improve the project as a contributor to the open-source community.
GitHub Code is Here
Key Features
Guns, rifles, automatic weapons
Robust object detection
Robber masks
Mask discernment
Tactical and kevlar vests
Spotting the good guys
People counting
Enhancing overall security
Guns, rifles, automatic weapons
Robust object detection
Robber masks
Mask discernment
Tactical and kevlar vests
Spotting the good guys
People counting
Enhancing overall security
Benefits
The Challenge
Existing video security infrastructure relies on numerous cameras placed around and throughout the building. These cameras connect back to a monitoring room, where the entire system relies on 1 or 2 humans to stare at all screens for hours. Errors can occur. People get tired.
Current systems aren’t scalable. Accuracy is poor, reaction times are slow, and 24x7 monitoring is costly. This problem may especially be acute at schools where resources are limited and physical campuses are widely varied.
This open-sourced system can be configured to operate on a smartphone. School security officers can monitor any threat from their phones.
This open-sourced system can be configured to operate on a smartphone. School security officers can monitor any threat from their phones.
Configurable Application
This app is configured with basic interfaces to track cameras that are added via the admin screen. Fully configurable, the app can empower teachers and administrators with a quick view of any potential crisis that may arise on campus.
We have provided graphics and, in the spirit of open-source, will give our Figma designs for free to anyone who asks. You can see them in this animated GIF. Right now, the graphics are not tied to the back end software. As with all open-source projects, we invite the community to take this app and extend it into new functionalities and configurations.
We have provided graphics and, in the spirit of open-source, will give our Figma designs for free to anyone who asks. You can see them in this animated GIF. Right now, the graphics are not tied to the back end software. As with all open-source projects, we invite the community to take this app and extend it into new functionalities and configurations.
Deploy on Existing Video Systems
This application can be deployed and integrated wherever security cameras provide a sufficient video feed of 720p and 30 fps. Basic network documentation is included in the GitHub repository.
The current software runs on GPUs and CPUs. and we provide set up guidance in GitHub. The GPU configuration provides more frames per second. Depending on the processing unit, the current CPU configuration will capture about 20 frames per second, which may or may not be satisfactory to some security personnel. Please stay tuned for progress updates.
By introducing AI image recognition to existing systems, 100s of video streams can be monitored second-by-millisecond with new edge computers that are omni-vigilant. Reaction times are instantaneous (<30ms), with real-time alerts and high accuracy.
The current software runs on GPUs and CPUs. and we provide set up guidance in GitHub. The GPU configuration provides more frames per second. Depending on the processing unit, the current CPU configuration will capture about 20 frames per second, which may or may not be satisfactory to some security personnel. Please stay tuned for progress updates.
By introducing AI image recognition to existing systems, 100s of video streams can be monitored second-by-millisecond with new edge computers that are omni-vigilant. Reaction times are instantaneous (<30ms), with real-time alerts and high accuracy.
The Challenge
Existing video security infrastructure relies on numerous cameras placed around and throughout the building. These cameras connect back to a monitoring room, where the entire system relies on 1 or 2 humans to stare at all screens for hours. Errors can occur. People get tired.
Current systems aren’t scalable. Accuracy is poor, reaction times are slow, and 24x7 monitoring is costly. This problem may especially be acute at schools where resources are limited and physical campuses are widely varied.
This open-sourced system can be configured to operate on a smartphone. School security officers can monitor any threat from their phones.
This open-sourced system can be configured to operate on a smartphone. School security officers can monitor any threat from their phones.
Configurable Application
This app is configured with basic interfaces to track cameras that are added via the admin screen. Fully configurable, the app can empower teachers and administrators with a quick view of any potential crisis that may arise on campus.
We have provided graphics and, in the spirit of open-source, will give our Figma designs for free to anyone who asks. You can see them in this animated GIF. Right now, the graphics are not tied to the back end software. As with all open-source projects, we invite the community to take this app and extend it into new functionalities and configurations.
We have provided graphics and, in the spirit of open-source, will give our Figma designs for free to anyone who asks. You can see them in this animated GIF. Right now, the graphics are not tied to the back end software. As with all open-source projects, we invite the community to take this app and extend it into new functionalities and configurations.
Deploy on Existing Video Systems
This application can be deployed and integrated wherever security cameras provide a sufficient video feed of 720p and 30 fps. Basic network documentation is included in the GitHub repository.
The current software runs on GPUs and CPUs. and we provide set up guidance in GitHub. The GPU configuration provides more frames per second. Depending on the processing unit, the current CPU configuration will capture about 20 frames per second, which may or may not be satisfactory to some security personnel. Please stay tuned for progress updates.
By introducing AI image recognition to existing systems, 100s of video streams can be monitored second-by-millisecond with new edge computers that are omni-vigilant. Reaction times are instantaneous (<30ms), with real-time alerts and high accuracy.
The current software runs on GPUs and CPUs. and we provide set up guidance in GitHub. The GPU configuration provides more frames per second. Depending on the processing unit, the current CPU configuration will capture about 20 frames per second, which may or may not be satisfactory to some security personnel. Please stay tuned for progress updates.
By introducing AI image recognition to existing systems, 100s of video streams can be monitored second-by-millisecond with new edge computers that are omni-vigilant. Reaction times are instantaneous (<30ms), with real-time alerts and high accuracy.