Generic Security System with Thermal Imaging Technology
American International University-Bangladesh (AIUB), Department of Computer Science and Engineering.

Abstract. In this paper, I proposed a model to build thermal imaging based security system which can be used to control the access on private places and computer systems. The security system will maintain a database in which the thermogram of individuals will be kept. To get access on those places and computers, recent thermal image of the person will be matched with database in a systematic way. Here, I derived some equations for considering the dispersion of body temperature to match with the database in different physical and environmental condition. Key words: Thermogram, Thermal imager, Infrared radiator, Biometric system, Image processing, Image patterns recognition.

1

Introduction

Thermal imaging system is the one of the most prominent and advanced technology to detect the radiated temperature (as image) from any objects or bodies by capturing it’s emitted infrared waves. Since almost every object radiate several volume of thermal energy, special type of device called thermal imager can be used to build the detailed temperature pattern of any object which is known as thermogram. Thermal image is varied from person to person and it’s unique [1]. So for ensuring true identification and correct authentication on different security system, thermal imaging technology can be kept in our general interest list. Recently different biometric parameters such as Fingerprint, Retinal, iris, palm print are being used in different security systems [2]. These parameters are also varied from person to person [3] but this can be stolen or copied or distorted. In some research, fingerprinting system is mentioned for securing the system such as credit card authentication by matching the finger print of a specific person with the saved data of server’s database [4]. But problematic situation can arise; if the person loses his/her finger in any accident then he/she will not be able to verify himself/herself and with the misuse of modern technologies the fingerprint can be stolen. Facial recognition can be considered as a parameter of security system [5] but face pattern may change with the age of a person [12]. Iris has its own patterns which varies person to person [12] but different eye disease may change the structure of iris. Beside iris detection requires more advanced and complex technology to use in security systems. Among all of them, thermal imaging can ensure more advanced security system for some of its special properties.

Thermal image of a specific person can’t be stolen or copied for other person or bodies because it’s impossible to apply ones particular thermal pattern to another person. Thermal imager always requires solid bodies or warm-blooded animals to build the thermogram [13] so false projection image will not work against a thermal camera. Different physical and weather condition may effect on individuals body temper for which his/her thermal pattern may change on certain ratio [6]. There is not enough research with thermal imaging technology to use it on security systems but its recently being used by firefighter or military force to detect hide person, arms or finding person in smoky environment. Rest part of this paper will show, how to identify the individuals thermal image in different body temper to provide authentication of general security systems.

2

Related works

As stated before, the research work on thermal imaging based security system is petty low. But some research has been done on thermal imaging to detect anxiety from human face, object detection, human activity detection and individuals face recognition. Bhanu and Hun proposed a model to recognize human repetitive activity like walking, running etc. from his/her thermal image by extracting silhouettes and employing Efficient Spatio-Temporal Representation (ESTR) and Gait Energy Image (GEI) [7]. They showed a mathematical equation for GEI to represent human motion sequence. Miyaura and Kimura developed a model for identifying human face by using thermal imaging [8]. They just normalize the front view of human face from thermogram and measured the temperature distribution according with average environment temperature and shape factor of human face. Levine and Pavlidis proposed a model to detect a person engaged in harmful behavior or illegal activities by analyzing the thermogram of facial pattern of that person [9]. They investigate the thermal image of anxious, fearful and alert faces in broad way and analyze the probability to evolve a person in illegal activities. Martinez and Kempen developed a model to detect the presence of landmines by capturing the thermal signature of bare soil [10]. Hoffman and Riggan designed a way to map the landscape temperature, spot the fire, measurement of flame radiation by thermal Imaging for remote places [11]. After reviewing these related works with thermal imaging technology, it can be considered that thermogram based security system would be a new and sophisticated approach in security sector.

3

Proposed method

For designing thermography based security system, we can use Uncooled infrared detector. Identifying individual by thermal imaging is well-suited for accessing control on privet area, intruder detection on restricted place, accessing control to secure computer system, increasing surveillance by analyzing imagery [1]. So, for ensuring security measure of those places and systems via thermography, first we need to capture the thermogram of individuals. Then the captured thermogram

and other identical information will be kept in a database of the system. To get access on those private places or secured computer systems, a person must have to authenticate himself on the system. For this, a thermal device will capture an image of the person and the thermal pattern of the captured image will be matched with the database. 3.1 Thermogram process and matching

The thermogram of person’s face portion at any instant is influenced by the vascular system of the body and ambient temperature of the environment [1]. So, when to match the captured thermal image with the database, the system should consider enough intelligence to cope with different physical and environmental condition. To do this, we have to divide the facial area into several parts. According to pavlidis and Levine, human facial area can be divided into five sub-areas as Periorbital area, nasal area, cheeks, chin area and Neck area [9]. They introduced that, for any person these five sub-areas of face can change the

Fig. 1. Five sub-areas of human facial area.

thermal pattern with the change of their physical condition such as his sickness or anxiousness. In order to consider this disparity of thermal image of same person, following steps may provide the solution. 3.1.1 Color introducing according to temperature

Uncooled thermal imager detect the radiated infrared energy by using false color index in which white hue means highest and blue hue means lowest temperature. Between white and blue, other orders of temperature spectrum are use red, yellow and green hue. By using inverse linear rainbow transformation, each pixel value of RGB triplet can be mapped to a single value in the range of 0 to 255 where 0 represents the lowest and 255 represents the highest temperature [9].

3.1.2

Measurement of dispersion of thermal pattern

Human body temperature can fluctuate between the range of 28 0 C (82.4 0 F) to 40 0 C (104 0 F) [14]. Although different weather and environment condition (hot or cold) and physical disease (such as fever) keep effect on body temperature, but body temperature doesnt cross that certain limit. So, to maintain this dispersion of body temperature, the system database should accept certain range of pixel value for human facial area. Such as if a person suffers from fever then the temperature of his cheek and neck will increase which enhance the pixel value of those sub-areas of face. Again for some reason such as cold atmosphere in winter, the body temperature may fall-down which decrease the pixel value. To accept this several fluctuation of pixel index of thermal image, database would maintain following equation: Pixel Pixel Pixel Pixel Pixel value value value value value of of of of of periorbital area = Θp ± Θrp nasal area = Θna ± Θrna cheek area = Θc ± Θrc chin area = Θch ± Θrch neck area = Θne ± Θrne

Here, Θp , Θna , Θc , Θch , Θne represents the general captured pixel value of periorbital, nasal, cheek, chin, neck area of face. Beside Θrp , Θrna , Θrc , Θrch , Θrne are used to maintain the range of the dispersion of pixel value for creating certain limits. Let, the general temperature of the cheeks of a person is 30 0 C and its equivalent pixel value is Θc = 231. Now, when to keep this value in database, the administrator will add or deduct certain value (according to need) such as Θrc = 9 (3 0 C). So, the range of the pixel value of cheek will be 222 to 240. In next, if that parson suffers from fever and wants to get access from the system, thermal imager may pixelate the index value of his/her cheeks to 235. Since, this value is in the range of saved value in database, so the system will consider this value as correct for authenticating. Same procedure is applicable for all other sub-areas of a face.

4

Future work

Although, I proposed this model on the basis of theoretical knowledge but in future the implementation of this work can be done by any person or group. This paper introduce the individual facial recognition for security purpose but more advanced work can be constituted with thermal imaging such as whole body identification for security system or finding disease via thermogram or authentication for transaction in ATM booth. Since, I could not use any thermal camera in this experiment practically, so error can be occurred in assumption.In this case, if any other group or person works on this concept in practically, it may possible to develop more generalize form of the equations or more accurate result or the experiment.

5

Conclusion

Since, thermal imaging system is very expensive for the high cost of the devices but it’s a most innovative approach to use this technology on different sector of modern life. For the uniqueness, anti-theft and less distortion properties of this technology, it can ensure more security of any system. This paper introduces a concept of designing a generic security system with thermal imaging which can be considered as a frame at different security measure. Beside this, I showed a way how to set pixel values limit of different facial sub-areas to manage the database system. Pixel value management helps to consider the fluctuation of thermal pattern of the body surface in different physical and environmental condition.

References
1. Prokoski, F.J., Riedel R.B., Coffin J.S.: Identification of Individuals by Means of Facial Thermography. Proc. IEEE International Conference Carnahan Security technology, Crime Countermeasures, 120-125 (1992) 2. Le, C.: A Survey of biometric Security System. Last Modified: Nov. 28, (2011) 3. Jain, A.K., Ross, A., Prabhakar, S.: An introduction to biometric recognition. Circuits and Systems for Video Technology, IEEE Transactions on 14.1, 4-20 (2004) 4. Wang P.T., Wu S.M.: Personal fingerprint authentication method of bank card and credit card. Patent Application 09/849,279, May 7, (2001) 5. Schroeder, C.C.: Biometric security process for authenticating identity and Credit cards, Visas, Passports and facial recognition. U.S. Patent No. 5,787,186, issued 28 Jul. (1998) 6. Arlowe, H.D: Thermal detection contrast of human targets. Proc. IEEE International Carnahan Conference on Security Technology, pages 2733, 1992. 7. Han J., Bhanu B.: Human Activity Recognition in Thermal Infrared Imagery. Proc. IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops (CVPR’05), Page 17-17, (2005) 8. Yoshitomi, Y., Miyaura, T., Tomita, S., Kimura, S.: Face identication using thermal image processing. Proc. IEEE International Workshop on Robot and Human Communication (RO-MAN’97), pages 374379, (1997) 9. Pavlidis, I., Levine J.,Baukol P.: Thermal imaging for anxiety detection. Proc. IEEE Workshop on Computer Vision Beyond the Visible Spectrum: Methods and Applications, pages 104109, June (2000) 10. Martinez P.L, van Kempen L., Sahli H., Ferrer D.C: Improved thermal analysis of buried landmines. IEEE Trans. Geoscience and Remote Sensing, 42(9):19651975, (2004) 11. Riggan P.J, Hoffman. J.W.: Field applications of a multi-spectral, thermal imaging radiometer. Proc. IEEE Aerospace Conference, vol 3, pages 443449, (1999) 12. “Biometrics new portal”, UK 2011 http://www.biometricnewsportal.com/ 13. “Thermography”, modified on 8 April, 2013 http://en.wikipedia.org/wiki/ Thermography 14. “Human body temperature”, modified on 26 April, 2013 http://en.wikipedia. org/wiki/Human_body_temperature

al