2002 Conference Proceedings

Go to previous article 
Go to next article 
Return to 2002 Table of Contents


CHECKING DAILY HEALTH AND SAFETY ON THE INTERNET

Nobuyuki Ishibashi
Yoshitugu Manabe
Satoru Oshiro
Motonori Doi*
Tomohiro Kuroda
Ichiro Kanaya
Yoshihiro Yasumuro
Kunihiro Chihara

Graduate School of Information Science
Nara Institute of Science and Technology
8916-5 Takayama-cho Ikoma City, Japan §630-0101
E-Mail: nobuyu-i@is.aist-nara.ac.jp 
*Osaka Electro-Communication University
18-8 Hajime-cho Neyagawa City, Osaka, Japan §572-8530
E-Mail: doi@isc.osaka.ac.jp

1.Abstruct

Numbers of elderly and single living persons are increasing recently depending on the extension span of our human life. We would like to recommend a system and simplified electronic information terminal facility for "Checking Daily Health and Safety on the Internet" by measuring the face color.

2."The Reference Notice" for Health and Safety by "Measuring the Face Color"

"The Face Color" should be measured continuously and be analyzed automatically for checking different values of "Red Color Component" employing 256 digitalized steps. Perfect color measurement of the face would be difficult and including many complicated functions, and then we would like to recommend a simple method which should compare the difference among every morning value of the face color through this facility where would be able to almost be kept in same measuring condition.

The difference value of the face color would be the reference information for checking the health condition and safety of elderly who would be living alone. This idea means that not absolute measurement value would be important checking every morning and evening condition continuously, if these would be changed so that difference value among data would be useful.

Fig.1 shows a typical using condition of this facility. Measuring points for checking the face color are indicated in Fig.2. Fig.3 explains the measuring points and the value of red color component from our students. We imagined and found out that an I-Point (under the eye) should be fit for measuring continuously.

3. Typical Opposites and Results of the Face Color Measurement

(1)Selection of the Measuring Point

Ishibashi measured first about his face such as measuring points as Fig.2, simultaneously we checked and analyzed the face color of another students as Fig.3. We decided to pick-up "Point-I"(under the eye) and "Point-C" (between eyes), because of the stability about changing values compared among the other points where the cheek and the lip were changed depending on the emotion. Fig.3 shows these typical data about above phenomena.

Typical using condition of simplified electronic mirror and terminal Measuring points

Fig.1 Typical using condition of simplified electronic
mirror and terminal

Fig.2 Measuring points

Face color data of students on measuring points

Fig.3 Face color data of students on measuring points (Red component)

Fig.4 Fluorescence natural color lighting

Fig.5 Incandescence lighting

(2) The Circumstance for Measuring

a) Back Scenery Influences

Typical colored curtain and blanket were compared to the back scenery for measuring the face color effectively. We were surprised influences of these back scenery and lighting condition. Fig.4 and Fig.5 shows these results, and Fig.6 indicates the difference of face color values depend on the back scenery.

Influences of the back scenery

Fig.6 Influences of the back scenery

b) Lighting Condition Influences

Fig.4 and Fig.5 shows differences about two kinds of lighting condition that both characteristics are almost similar influence pattern. Fig.7 shows more comparison which shows influences of two kind lighting conditions and the camera operating conditions with AF(automatic focusing) ON and WB(white balance) ON/OFF differences. The difference between I and C such as Fig.6 and Fig.7 means that a reference reliability will be useful for referring health condition such as an old proverb.

Compared data under the different lighting and camera operation

Fig.7 Compared data under the different lighting and camera operation

(3) Normalizing a Reference Red Mark

Fig.8 shows each mean value of 5 points on the face Point-I and Mark-M values. M-I difference values are 6.7%, M values are only 1.6% difference under two kinds of the lighting condition. The Marking Point-M value will be effective for normalizing and M-I value measuring data stabilization.

Mean values of C,I and Mark 5 each point

Fig.8 Mean values of C,I and Mark 5 each point

4. Analyzed Reference Data between Point-I and Mark-M

Fig.9 indicates Points-I,C,M and M-I reference values. Actual reference data for checking health condition will be inform as Fig.10 continuously measuring. From this data we can find out unusual condition such as #535,#552,#681, #780 and #727 which Ishibashi has worked hard till midnight and been tired in these day.

We are now trying to collect data from outside elderly cases.

Typical measuring data for reference check of the health condition

Fig.9 Typical measuring data for reference check of the health condition

Actual measuring data for reference of the health continuously

Fig.10 Actual measuring data for reference of the health continuously

5. The Typical Pictures of the Facility

Fig.11 shows simplified picture on the information terminal facility for easy operation by the elderly who can communicate through this on the Internet between the assistance home helper center or the hospital every day.

The picture of the Home Page The picture of the Life Mirror

Fig.11 The picture of the Home Page

Fig.12 The picture of the "Life Mirror"

6. Summary and Subjects

Now we are experimenting fundamentally, and more testing many case for realizing this recommendation continuously. We would like to recommend above data which mean the face color on the Point-I normalized by the Mark-M as difference value M-I will be useful for referring and checking our health condition comfortably under usual living style. But we are afraid that this system is not so reliable for medical purpose, and then the data will be only a reference for the safety checking of single living elderly.

7. Acknowledgement

We are grateful to Gunze Mr.Kaseyama and Mr.Ohyoshi for advices about producing the facility, and to our Institute Professor Dr.Kidode, Dr.Imura, Mr.Sasaki, Mr.Tabata, Mr.Tateishi,Mr. Kojima, Mr.Maeda and Ms.Murakami for effective discussion and kindly advices.

8. Reference

(1) A Robust Face Identification against Lighting Fluctuation; Motonori Doi, Kosuke Sato and Kunihiro Chihara, System Control Information Society, Vol.11, No.10,pp.546-553,1998

(2) Lock Control System Based on Face Identification; Motonori Doi, Doctor Report, Nara Institute of Science and Technology, DT9561029,1998

(3) Homepage Builder User's Guide Version 6,2000


Go to previous article 
Go to next article 
Return to 2002 Table of Contents 
Return to Table of Proceedings


Reprinted with author(s) permission. Author(s) retain copyright.