Cisco Research
Revolutionizing Computer Vision with Hybrid AI and Intent-Based Framework
Cisco Deep Vision is an innovative open-source project dedicated to advancing computer vision technology by integrating deep learning, serverless computing, hybrid AI, and intent-based frameworks.
Our mission is to process large-scale streaming multimodal data—including visual, RF, and LiDAR information—and create self-supervised deep knowledge representations of the world that closely resemble human understanding.
Accurate computer vision systems are crucial for enhancing health and safety across various environments and industries, including healthcare, retail, manufacturing, and critical infrastructure. They play a vital role in the safe operation of self-driving cars, worker safety in regulated industries like mining, oil, and gas, and the implementation of high-value niche applications such as suicide prevention and anomaly detection. Cisco Deep Vision empowers intent-based analytics for computer vision systems, enabling them to see and understand the world like never before while optimizing camera performance and PTZ functionality.
By applying first principles thinking to computer vision, Cisco Deep Vision is working to create systems that are far more accurate than today's box AP metric. These systems have the potential to revolutionize a wide range of industries, from healthcare and manufacturing to transportation and beyond. Ultimately, the goal is to build computer vision systems that can see and understand the world in the same way that humans do, opening up new possibilities for innovation and progress.
Expanding the scope of our project, we also explore ambient sensing techniques like WiFi sensing (RF), thermal imaging, LiDAR, and vibration sensing to create a comprehensive understanding of the world, even in vision-less scenarios. Vibration sensing is particularly crucial for privacy-preserving analytics within the home, ensuring effective monitoring without compromising privacy. This enables Cisco Deep Vision to provide unparalleled insights into various environments.
With a focus on addressing core challenges like improving object detection and tracking, Cisco Deep Vision strives to create reliable and secure video analytics solutions that can easily integrate with existing systems. Our commitment to explainable AI and sense-making ensures complex scene understanding and opens opportunities for both deep learning and symbolic AI.
As an open-source project, Cisco Deep Vision invites researchers, developers, and industry professionals to collaborate, contribute, and drive the future of computer vision technology. Join our community today and help us shape a safer and more connected world.