According to ABI Research’s report, it is projected that over 600 million mobile devices will be produced with vision-based behavior recognition features in 2017. However, there are still several technical problems need to be solved with these features first:
1. Under Different Lighting Conditions
2. Interferences From Moving Backgrounds
3. High Power Consumption
4. Large Size of the Mobile Device Itself
5. Double BOM Cost for the 2nd Camera (For color video calling and color vision sensing)
This technology solves these problems above for tablet PCs and mobile phones.
Two Implemented Applications:
The team solves the mentioned problems for tablet PCs and smart phones and has implemented 2 applications as below:
1. It successfully replaces Slide-to-Unlock patent from Apple.
2-1. No interruptions from the power-saving setting (It will turn off the backlight frequently) for long-time using circumstances. Such as watching a movie, reading a book, checking the stock market or reading a map on the mobile device.
2-2. A great deal of power can be saved from our module, as the screen from the device will be turned off immediately once the user is away. It also saves the power from using the video applications that are with Auto-play settings.
(2-1. & 2-2. also like Samsung S4′s Smart Pause and Smart Stay combined)
Click here to view the embedded video.
█ Power Consumption: 0.1444mW (Mass product edition, not prototype in video; By the way, Kinect is 13200mW):
According to the 2 major applications: To read and to unlock, the module only will be turned on immediately once it detects the user (the operation rate is really low) and it will be always off for the rest of the time when the device is not being used.
Take reading for example. The backlight timeout preset time of the most popular 10-inch iPad is 5 minutes and the backlight timeout preset time of a 7-inch Samsung tablet is 30 seconds. Our module only takes 0.1 seconds to operate, the operation rate is only 0.0222%, and the power consumption is only 0.1444 mWh. Imagine how much power our module can save from those settings.
█ Power Consumption of the Display Backlight:
Take 7 to 10 inch tablets that are released in the market recently for example, there are about 20 to 36 backlight LEDs in those tablets. When the tablet is not being used but the backlight is still on, the power consumption from all those backlight LEDs can be assumed as how much power we can save by using our module. Furthermore, our module only uses 1 to 2 LEDs, it only detects in every 15 seconds or 2.5 minutes and it takes only 0.1 seconds to operate.
█ Power Consumption from the Power Button:
According to Sony’s research, people press the power button 76 times a day on their phones. If people use their phones 15 hours a day, the button would be pressed 5 times an hour. According to the regular using habits, if people forget to press the power button to turn it off after using their phone twice an hour, our module can save more than 923 times power that is consumed from a 10 inch LCD screen.
Our previous work: http://www.kinecthacks.net/using-1-camera-to-integrate-live-foreground-fake-background-in-real-time/
Sky Chen’s LinkedIn: tw.linkedin.com/pub/sky-chen/67/489/80/