Umbo Computer Vision

About Umbo Computer Vision

Umbo Computer Vision, founded in 2014 by Shawn Guan, Ping-Lin Chang, and Tingfan Wu, operates as an artificial intelligence company specializing in autonomous video security. The firm is headquartered in San Francisco, with additional offices in London and Taipei. The founding team's expertise laid the groundwork for the company's direction; CEO Shawn Guan had prior experience in business development within the surveillance industry, while CTO Ping-Lin Chang brought a strong academic background with a PhD in Robotic Vision from Imperial College London. This combination of commercial and technical knowledge guided Umbo's focus on creating AI-first security products.

The company's core business revolves around developing and providing a suite of AI-powered security products for enterprise clients, including schools, government facilities, and commercial sites. Umbo's business model is primarily subscription-based, generating revenue through annual payments per camera, which allows customers to access its cloud platform and AI features. The company has secured significant funding across multiple rounds, accumulating over $20 million from investors such as CDIB Venture Capital, TransLink Capital, and AppWorks, to fuel product development and global expansion.

Umbo offers an integrated hardware and software solution designed to reduce false alarms and minimize the need for constant human monitoring. Its product line includes AiCameras like the SmartBullet and SmartDome, the TruePlatform cloud management system, and the AI engine known as Light. The Light AI is engineered to go beyond simple motion detection to analyze video streams for specific human behaviors such as intrusion, loitering, and wall-scaling. This system processes video feeds to identify events, which can be managed through the TruePlatform. This platform can be used with Umbo's proprietary cameras or integrated with existing third-party IP cameras, offering flexibility to its clients. By leveraging deep learning algorithms, the system can distinguish between genuine threats and benign movements like falling leaves, a common issue with traditional security systems.