Generating signatures within a network that includes a plurality of computing devices of varying processing capabilities
Generating signatures within a network that includes a plurality of computing devices of varying processing capabilities is disclosed. Chips may be transmitted, from a network video recorder and over the network, to an analytics appliance having a GPU processing power that is higher than a GPU processing power possessed by the network video recorder. The GPU processing power possessed by the analytics appliance may be employed to process the chips therein and generate respective signatures.
1. A surveillance system comprising:
a camera that captures video frames;
a Video Management Software (VMS) server stored on a computer readable medium in a first computing device housed in a first enclosure, the first computing device communicatively coupled to the camera; and
a second computing device housed in a second enclosure different than the first enclosure, the second computing device being communicatively coupled to the first computing device and having a GPU processing power that is higher than a GPU processing power possessed by the first computing device, and
wherein:
the second computing device is an analytics appliance configured to generate multi-dimensional feature vectors corresponding to objects of interest in the video frames and return the generated multi-dimensional feature vectors to the first computing device for storage and use therein, and
the generated multi-dimensional feature vectors are generated at least in part by a learning machine that spans at least two GPU cards of the plurality of GPU cards.
2. The surveillance system as claimed in claim 1 wherein the learning machine comprises a convolution neural network.
3. A method comprising:
generating a plurality of chips from video frames captured by a camera that is communicatively coupled to a first computing device within a surveillance system;
transmitting the chips, from the first computing device and over a Local Area Network (LAN), to a second computing device having a Graphics Processing Unit (GPU) processing power that is higher than a GPU processing power possessed by the first computing device, and wherein the second computing device is an analytics appliance;
employing the GPU processing power possessed by the second computing device to process the chips therein and generate respective multi-dimensional feature vectors; and
transmitting the generated signatures multi-dimensional feature vectors, from the second computing device and over the LAN, to the first computing device for storage and use therein, and
wherein the generated multi-dimensional feature vectors are generated at least in part by a learning machine that spans a plurality of GPUs of the second computing device.
4. The method as claimed in claim 3 wherein the learning machine comprises a convolution neural network.