Combining body movement information and foot contact forces to recognize human identity by analyzing walking method

Dissertation for M.Sc degree

Trend: relationship between human, machine, computer

Abstract

Today with rapid technological changes, the level of security Individuals and organizations have also decreased and this security has been attacked by profiteers and vandals. In the meantime, the increasing progress of biometric methods has made tremendous achievements in the fields of military, commercial and financial security. Biometrics refers to the technology of measuring and processing the characteristics of people's bodies for identification purposes. A person's gait is one of the physical characteristics of a person's body, which is used as a biometric in identity recognition systems today. All existing methods based on walking for identity recognition analyze the images taken of a person while walking. In this thesis, footprints are introduced as a new biometric method. This biometric is based on the fact that people's footprints differ in some ways when they walk. The distinguishing features of people in terms of footprints can be the size and angle of the foot while walking. 12 footprint features were selected to identify people from each other, and a three-layer perceptron neural network was trained to recognize people based on these features. It was shown that by using these 12 characteristics, people can be accurately identified and distinguished from each other.

Introduction:

Man has needed to distinguish friend from enemy for survival, and identity recognition was and still is a vital thing for him, so today attempts have been made to mechanize identification systems. These developments are proof of the need of society and the world. A need in which progress reduces violations, increases security, speeds up daily affairs, and so on. has been In the past, fingerprint and facial imaging were used to identify criminals, but now more sophisticated methods have been developed for this purpose.

Human walking, which is a new biometric with the aim of identifying people while walking, is increasingly playing an important role in video surveillance applications.

We use equipment every day. We use basic security, for example, we have a key to enter our house, we need a username and password to use a computer, and when using a bank teller, we need both a card and a password. We need to pass. It is very annoying to misplace our key, forget our password and lose our debit card. But it is worse when someone uses them instead of us, just as if we used them ourselves. This is at a time when, with rapid changes in technology, the security factor of individuals and organizations has also decreased, and this security has been attacked by profiteers and vandals. Therefore, according to the rapid development of computer networks and communications, the need for reliable methods of verifying people's identity is felt more and more. The common methods that are used today to identify people's identity are based on two things, things that we own, such as house keys and credit cards, which if they are lost, others can use them under certain conditions, or Things that we know about, such as username and password, if they are simple, they can be easily obtained by guessing or testing, and if they are very complicated, it will be difficult to hand them over and we will be forced to go somewhere. Let's write them down. In this case, there is a possibility of them being lost or stolen. Now let's imagine that the body is converted into a key or password. Physiological characteristics of the body (such as face and fingerprints) and habits and behaviors (such as handwriting and voice) and the way of walking are so complicated that they increase the security factor and we certainly do not leave them or not. We forget. In fact, biometric systems use the human to confirm the identity.

Many methods have been proposed to recognize the identity of people from walking in video images, which can be classified into two general categories: statistical approaches and model-based approaches, the presented approaches include three general phases. Pre-processing, feature extraction, recognition..

Biometrics is an important and automatic method of identifying people based on physiological or behavioral characteristics. Walking recognition in biometric technology has become an important and safe identification and personal verification solutions.

Problem statement:

In biometric science, body parts They are considered to be easier to use and less harmful. Each of the used methods have weak points and strengths that can be eliminated by combining them with other security methods. No one wants to find out that the balance is empty when checking their balance through online banking networks. guess and by understanding our number in the bank, they can easily enter our account using the bank's online network and empty our balance. Usually, the majority of people in a society experience such problems. Due to the significant growth rate of world trade and the importance of trade, it is not possible to use old manual or present systems for a long time, on the other hand, using these old methods causes waste of energy and Time has increased and less work is done in a long period of time. Therefore, in business, the need for electronic commerce is felt, and the very important issue that is being considered today is the issue of security. By using reliable methods, biometrics can be the answer to such problems to a large extent. Biometric science is not only used in electronic commerce, but also in many other cases. For example, in important and sensitive laboratories or the entrances of buildings that we are sensitive about entering and exiting, or we can use the locks on which the keypad is installed and pass to the people concerned by the name and password. Let's let them use it when they log in, but this method is also not very reliable, it won't be worth it if you lose the password. But when it is caused by the finger or the palm of the hand or the way of walking. If it is used for identification and permission to enter, these problems will no longer arise.

Many solutions have been proposed to solve these problems, such as identity recognition through fingerprints, identity recognition through eyes, identity recognition through Through the face, identifying the identity through the iris, which we will explain some of them in the next chapters. In this thesis, we will introduce a new method to identify the identity of people. Our idea is to use the method of identifying a person from walking, we want to identify his identity by using specific footprints while walking to solve the problems and issues that we discussed above to a large extent. We control the entry and exit of people by using the word "footprints" and allow people whose details are registered in our database to enter and exit.

Research hypotheses:

The way different people walk is different.

The forces that people apply to the ground while walking are different.

The angle between the legs and the size of different parts of the soles of the feet are different.

The identity of the person can be recognized from the way of walking and the size and angle of different points of the soles of the feet.

1-3-1- A brief overview of the history of biometrics:

Scientific texts about the quantitative measurement of human behavior and habits and external characteristics for identification purposes go back to the 1870s and Alphonse Bertillon's measurement system. In America, the Bertillon system was used to measure the body, including measuring the diameter of the skull and the length of the arms and legs to identify prisoners until the 1920s.

Henry Folds and William Herschler and Sir Francis Galton proposed quantitative identification through fingerprints and face measurements in the 1880s. The development of digital signal processing techniques in the 1960s quickly led to work in the field of human automatic recognition.