20/01/2018
Geovision GV-VD8700 , Qualcomm & Intellivision collaborate on smart AI cameras Taipei, Taiwan
With this new camera Geovision brings what it believes to be the world-first deep learning-based, 4K HEVC, intelligent video analytics (IVA) on-device smart camera, featuring on-camera face detection, tracking, recognition of up to eight faces, human detection, and intrusion detection for various security applications.
Traditionally, video analytics processing for video surveillance has been limited to server or cloud management due to CPU or GPU constraints, which are often still time lagging to properly provide meaningful and useful analytic results in real-world applications, and on time. Qualcomm’s Snapdragon 600 connected camera platform and Intellivision’s video analytic algorithm fundamentally changes how video is captured and analysed, to truly build a new class of “Artificial Intelligence (AI) camera”, so that the camera can help to sense, gather, and analyse data. With the new platform, the GV-VD8700 can easily add deep learning capabilities with very low power requirements; reduce data streaming, and create immediate action and event alerts. Video analytics features such as facial recognition, human detection, and intrusion detection will be able to operate on multiple channels in real time.
According to IHS, in 2016 the Access Control System (ACS) and Intrusion Detection System (IDS) market is growing faster than traditional Video Surveillance System (VSS). With such intelligence on the camera itself, customers can move from expensive, server-based video analytics to timely, efficient edge computing camera devices. Applications include surveillance of casinos, hospital, retail stores, high value storage and offices. Integrating a GV-VD8700 to an existing access control system, will immediately boost the level of security by providing real-time, multiple-channel, face-indexed black/white list without additional cost for the servers. Meanwhile, those data can broadcast through human resources (HR) departments for facility access or absence report records.
