11. Nuttapoom Amornpashara, Yutaka Arakawa, Morihiko Tamai, and Keiichi Yasumoto, ``Phorec: Context-Aware photography
Support System Based on Analysis of Big Data of Good Photo with Location, Time, and Weather Condition,'' ACM HotMobile 2014
From where?
Which season?
What time?
How about a weather?
Camera setting?
(ISO, exposure, etc)
Flickr分析による写真家支援システム
• 季節・時間・場所・カメラ設定を推薦
17. 脈波を用いたリアルタム睡眠状態推定
• 普及しつつある脈拍計を使い睡眠状態推定
You fell
asleep at
14:00
14:20
Wake up!
Final deadline
Desired sleep length
Ring
Alarm application
Machine Learning
Commercial
sleep monitor
Silmee, Toshiba
Palse sensor
Myo LINK
Realtime
heart rate
Sleep state and
Heart rate model
Ground Truth
- Wake
- Rem-Sleep
- Light-Sleep
- Deep-Sleep
- Pulse (bpm)
7 features
- Sleep states
- Avarage and dispersion of pulse
- Current epoch
- Previous epoch
- Next epoch
Input (devices)
Other commercial wearable heart rate monitors
(If they start providing a raw data in realtime)
Smartphone-based
Heartrate monitor
ne-based
Camera
Flash
Holding case
by 3D printer
myoelectric signal based
heart rate sensor
Collecting a sleep data Standarizing the collected data Constructing an identification model Estimating sleep states
Sleep state identifier
Deciding sleep-onset time
Set a alarm
at a proper time10 subjects
3 sleeps
/ subject
4th example is the latest my research. We are developing social data based urban sensing.
Currently, flickr-based sightseeing spots retrieving and context-aware photography support system are running as a project.
Compared with the previous example, social-data based urban sensing doesn’t require the expensive infrastructures.
It just analyze a stored past data. It means that it is not suitable for realtime sensing.