To achieve the integration of Internet of things and AI, Ruff launched face recognition
Время публикации: 2019-04-10
If scientific and technological innovation is a kind of zeitgeist, then the landing of science and technology is the destination of this zeitgeist.- Ruff
With the gradual expansion of iot business scenarios, more and more scenarios begin to upgrade based on technological innovation.AIoT, a new product officially launched by Ruff, is a Face recognition module and terminal product named "Ruff Face ID". Based on the Internet of things + machine learning technology, the original complex AI algorithm is embedded into a 35mm*42mm chip module, in which the AI chip provides neural network computing of machine vision and speech recognition ability.
Through the neural network accelerator IP, the chip can carry out high-speed convolutional neural network calculation under ultra-low power consumption, such as target detection and image classification task based on convolutional neural network, face detection and face recognition, multi-classification object detection and recognition, etc..
It is reported that Ruff Face ID has machine vision and hearing ability. In actual business scenarios, it can store and calculate data at the edge of the device to achieve low-power visual processing speed and accuracy.At the same time, it has the hardware accelerator KPU of convolutional artificial neural network, which can carry out the operation of convolutional artificial neural network efficiently.
Ruff Face ID has the following competitive advantages over other Face recognition modules in the market:
Realize the edge computing fusion of Internet of things and AI
In the case of high real-time requirements, the demand from the cloud to the edge is gradually amplified.In mission-critical applications for face recognition tracking, local data processing or edge computing will begin to become necessary.
The AI chip of Ruff Face ID can realize the offline processing of edge computing, and the advanced machine learning model of deep neural network is optimized based on the scene to run the unstructured data including video frame, speech synthesis, time series data and other devices generated by cameras, microphones and other sensors on the edge.
Edge computing fusion based on Internet of things and AI can achieve more efficient and rapid response in business scenarios, and reduce the storage and operation cost of cloud data, especially some data with low repeatability and value, which can be directly calculated at the local edge.
Out of the box, through the end - to - end data networking
For corporate clients or vendors, intelligent technology in the process of the ground to the product layer will experience all sorts of complex equipment docking with the data, including the underlying technology to the module, the module to the equipment, equipment to the data, and data to the cloud, for technical strength lack enterprise vendors, to overcome the gap not only need to spend a lot of time cost, also need to the corresponding technical personnel involved.
Ruff Face ID provides a complete set of Face recognition modules and cloud services.Cloud service is provided to users in the form of SaaS. Users can directly complete remote device management, including firmware upgrade, algorithm update, face recognition result message push, history record view, remote face feature data management and so on, by calling corresponding SDK or API, so that users can realize the convenience of out-of-the-office.
Low cost within 100 yuan, making IoT+AI more popular
In the past, the processing method of AI chip was to transfer data such as images to the cloud for processing. The processing speed was fast, but the traffic cost was high, and the real-time response was not timely enough, especially in mission-critical applications.The other is local offline processing, but running it on a general-purpose GPU or CPU can be expensive.Ruff Face ID has an AI dedicated computing unit KPU built-in, so it realizes low-cost offline AI capability.
With the overall technology maturity of AIoT and the increase of upstream and downstream manufacturers' shipments, Ruff Face ID can achieve the shipment cost within 100 yuan in terms of the price most concerned by end customers, which greatly reduces the operation and maintenance cost of end customers.
Ruff Face ID was reported by many media before it became popular
Before the official launch, Ruff Face ID appeared in the 2019 "entrepreneurship in Shanghai" international innovation and entrepreneurship competition, which was reported by many media, including Oriental finance and economics, Shanghai news, hello zhang jiang, wen wei Po, youth daily, CCTV news mobile network, etc.
- 1.ROHM introduces AC/DC converter IC with built-in 1700V SiC MOSFET
- 2.AMD officially announced the Ryzen R1000 embedded processor
- 3.Intel 14nm process continues to improve in performance and power consumption
- 4.Intel's XE graphics card may not be its own 10nm, but samsung's 5nm EUV
- 5.TDK implementation of new enterprise strategy to continue to force the Chinese market
- 6.Design a switch power supply you must understand the problem
- 7.What is the importance of the FPGA field
- 8.AMD releases arron embedded R1000 SoC: Zen architecture, dual core quad thread, 3 times performance improvement
- 9.Introduction of Molex's new 1.25mm thread pitch SlimStack floating board-to-board connector
- 10.MPS introduces the latest integrated motor drive module - for 'intelligence'
- 11.Ten bourns potentiometer recommended
- 12.Details of Maxim's new MAX77860 3A switch mode charger
- 13.77Ghz single chip millimeter wave sensor can realize automatic parking
- 14.Vishay's new automotive class IHLP inductor details overview
- 15.Qualcomm is aiming to take on standard nvidia and Intel with its Cloud AI 100 reasoning chip