Some infectious diseases can spread rapidly via a community of human or animals or both, either through airborne particles or viruses. Such rapid spread diseases may become a local, national or international widespread and contagious threat. As a symptom of infection, the body temperature of a disease carrier is higher than normal people. In this chapter, flu detection system using thermal imaging tool and computer vision techniques are discussed. An automatic flu detection method adopting human object extraction algorithm and fuzzy logic based Viola Jones algorithm are also discussed. The proposed system able to capture a thermogram of the human subject, detecting the eye region of the human subject, calculating the pixels values around the detected eye region, converted to temperature readings and further classified the subject's body temperature whether the subject satisfies a flu condition or not. Experimental results also shown that the proposed fuzzy logic based Viola Jones algorithm can trace out flu infectious personal from the input thermal images up to 80% of accuracy.
TopIntroduction
Flu is a type of infectious disease induced by the outspread of influenza virus. The dominant symptoms on the flu infected personal comprise of: high fever, sore throat, running nose, headache, muscle pains, coughing, fatigue, congestion and muscle aches. From all these symptoms, a rise or drop in body temperature may be an early symptom of such disease, and as such it may be desirable to measure the body temperature of each person for flu detection. This may help to detect and quarantine infected individuals at certain locations such as airports or border, to prevent a disease entering a region, state or country. Conventional flu detection system using contact measurement system (clinical thermometer), whereby modern flu detection system applying contactless measurement system (infrared thermometer and infrared camera), both are to trace out whether there is an existence of infectious disease active carrier or not for release pass or quarantine purposes. The main merit of modern flu detection system as compare to conventional flu detection system is that in the contactless flu detection system, the security personal no need to have physical contact with the examinee. In a better version of contactless flu detection system, the image processing based flu detection system can help free up the long que faster, saving times, at the meantime offering the examinee a more relaxing body temperature measurement way.
In general, the contactless flu detection system can be divided into two main categories, one is by the infrared thermometer measurement and another one is by the infrared imaging capturement. Infrared thermometer measurement is by the use of medical infrared thermometer or the door type forehead thermometer. The infrared thermometer first invented by (Egawa, 1997) comprising an infrared sensor for receiving heat radiation from an object and an arithmetic circuit which calculates the temperature readings. The door type forehead thermometer manufactured by TECOS (Topteches 2015) is a walkthrough forehead thermometer. The examinee is allowed to enter the passageway through a door. An infrared optical recognition system above the door checks the temperature of the examinee. One problem encounter in infrared thermometer measurement is the requirement of long queuing, since only one person at a time enters the passageway to get the measurement. Hence it is better to apply the infrared imaging capturement tool that can help free up the long queue, since the photo of the people who pass through the camera will be captured in video rate, and it will be process with image processing technique to extract the human objects in the captured image and further identify flu infected personnel (human body temperature > 37.5degree Celcius). Infrared cameras make small temperature difference visible. They are currently applied widely in many new or existing security networks.
Conventional infrared imaging system normally employs human observers to analyze the captured video. In the long run this is vulnerable to error due to deterioration in attention of the human observer (Hampapur, 2003). It is a fact that a human’s visual attention drops below acceptable levels when assigned to visual monitoring and this fact holds true even for a trained personnel (Hampapur, 2005; Green, 1999). The weakness in conventional infrared imaging system has raised the need for a smart flu detection system where it employs computer and pattern recognition techniques to analyze information from situated imaging tools and automatically detect flu infected personal. An automatic flu detection method is discussed in this chapter, this includes human object extraction algorithm and fuzzy logic based Viola Jones algorithm. With the algorithms proposed in this chapter, it can automatically detect the flu infected personal at certain location captured by the infrared imaging tool, without required long queue for one by one checking and not to disturb the moving personal proceed to their destination ahead. These monitoring and subsequent analyses of the images from the inspection can alert security personnel to take further action on either let pass the healthy personal or quarantine the flu infectious people effectively.