- The following is content from the press release -
The "eyes" of the AI monitor sleep, detecting and notifying the user of the danger of falling asleep, leaving the bed, falling over, and so on. Provides a safer sleep environment by reducing the burden on staff at kindergartens, daycare centers, hospitals, and nursing care facilities.
(Head Office: Chuo-ku, Tokyo; Representative Director: Kenichi Ozaki) announces the launch of the DC900, an AI sleep monitoring system using the latest AI technology.
AI video analysis functionality with AI algorithms
Advanced AI face recognition technology developed for the DC77 AI face recognition camera is evolved to achieve advanced sleep state diagnosis through high-speed hardware processing of the CPU and NPU (AI accelerator) chipset.
Unique sleep recognition algorithm
Using a proprietary AI algorithm to analyze complex sleep patterns, we have succeeded in recognizing sleep states with an average accuracy of 90% or better under all circumstances.
For example, advanced AI processing can determine not only whether a person is "sleeping on his/her back" or "not sleeping on his/her back," but also whether he/she is "likely to sleep on his/her back.
Video of the DC900 in action. You can see the reaction of the DC900 in a situation where the child on the far right is gradually put to sleep on his/her back.
Up to 4 cameras can be connected
Thanks to its powerful processing power, up to four webcams can be connected to one unit to watch over separate rooms simultaneously.
Support for a variety of scenarios
In kindergartens and nursery schools, naptime monitoring can be performed, and in hospitals and nursing homes, sleep monitoring, as well as fall detection and bed release detection can be added.
The camera uses a commercially available webcam
There is no need to use a dedicated camera. The DC900 is equipped with eight PoE-compatible LAN ports to simplify wiring work.
Basic operation with commercially available tablets
Considering staff not accustomed to operating PCs, the system is designed so that basic operations can be performed with commercially available tablets. (Recommended number of simultaneously connected tablets: 2)
Centralized management of multiple DC900s in the cloud
Multiple DC900s can be centrally managed in the cloud. This makes it possible to build large systems, as well as version control of DC900 and tablet applications.
State-of-the-art hardware for quietness and stability
The CPU+NPU (AI accelerator) chipset is used for high performance yet low power operation. Minimized heat generation, sealed metal housing with no cooling holes and fanless construction ensure completely quiet operation.
This is a particular concern because it is a device that watches over your sleep.
Support for various notifications and collaborations
Flexible notification functions and expandability, such as contact output, allow various interlocking actions when an alert occurs.
Various alarms
Email Notification
Interlocking with patrol lights, etc.
Evolving systems with AI learning
Although the system has already learned more than 90% average accuracy at the time of shipment, it is equipped with a learning evolutionary algorithm that can further improve accuracy through AI learning by accumulating examples.
Connection image
DC900 Application Scenes
Main Specifications
Hardware Specifications
| processor | 8-core ARM CPU integrated NPU (AI accelerator) |
| memory | 6GB DDR4 up to 32GB x 4 |
| EMMC | Standard 24GB 32GB/64GB can be changed to 32GB/64GB |
| storage | 1TB |
| NVR | 4-8 channel video real-time processing, video export, platform management |
| decode | 4-8 channel 1080p video @ 30fps |
| compatibility | Onvif |
interface
| Video output | HDMI x 1, 1080P/4K compatible |
| network | 1000BASE-T x 1 100BASE-T (PoE support) × 8 |
| USB | USB Host x 1, OTG x 1, NVR x 2 |
| Flash Card Slot | Two expansion card slots up to 256 GB |
| audio | 1 system (output 5W) |
| Extended phoenix terminal socket | Dual RS845 with 12V power supply Input port DI port for control x 2 High/Low level control Output port for control x 2 contact output 30V@500mA |
| MIPI | 40-pin x 1, I2C with touch sensor |
Other Specifications
| power (button on TV, etc.) | AC100V / DC24V/3A |
| Operating temperature | -10°C to 55°C |
| Operating humidity | 10% to 90% RH |
| Size | 180mm(W) × 210mm(H) × 60mm(D) |
| Installation method | Desktop or wall mounted |
Price
open
DC900 Home Page
https://dc77.jp/dc900/
Company Profile
Company name: Dambonet Systems, Inc.
Location: CANALTOWER 1F - 10F, 9-3, Koamicho, Nihonbashi, Chuo-ku, Tokyo
Representative: Kenichi Ozaki, Representative Director
Established: December 1997
URL:https://www.dumbonet.com
For inquiries from customers, please contact.
Phone number: 03-5623-9666
Email receptionist: info@dumbonet.com
