Development of a personal identification program. Modern biometric identification methods. Keyboard handwriting recognition

Development of a software tool for identification of a person by voice PERFORMED BY: STARIKOV K.A.

Aims and Objectives To study the theoretical foundations of biometric voice identification; Analysis of existing methods for the implementation of software tools for personal identification by voice; Development of a software tool for personal identification by voice; Calculation of the economic cost of the software tool; Calculation of labor intensity in the development of software.

Biometric identification methods Face geometry Fingerprint Hand geometry Iris Drawing of veins Handwriting Voice

Personality recognition system by voice Recognition systems work in two stages: registration of a new user; identification of the registered user (recognition process). Main program Database User registration Successful / unsuccessful registration User identification Successful / unsuccessful authentication Request processing and operation execution

Information scheme Not Registered Registered User Database User Microphone Recorded Voice Processing Recording Conversion Comparison Result Display

Extraction of signs from a speech signal Signs can be divided into two types: low-level (anatomical structure of the speech apparatus); high-level (manner of pronunciation). Methods of feature extraction under consideration Mel-frequency cepstral coefficients Cepstral coefficients based on linear prediction

Chalk-frequency cepstral coefficients ALGORITHM METHOD 1. Submission by parts of the input signal (human speech) The duration of one segment (ms) was calculated by the formula: 2. Application of the weighting function (Hamming window) to reduce distortions 3. Discrete Fourier transform

Chalk-frequency cepstral coefficients 4. Division into bands using triangular filters (filter boundaries are calculated in chalk scale) Triangular filters on chalk scale 5. Triangular filters on frequency scale Discrete cosine transform (calculation of chalk-frequency cepstral coefficients)

Voice signal recording Sampling rate 44100 Hz; Encoding - 16 bit; The key phrase is "sound recording."

Example of calculating coefficients for a segment of a spoken phrase 1. A signal segment with a duration of 23 ms; 2. Filtered segment 3. Spectrum of segment

Graphs of comparison of chalk-frequency cepstral coefficients Chalk-frequency coefficients of speech signals of two different users (one segment) Chalk-frequency coefficients of speech signals of the same user (one segment) 100 80 80 60 60 40 40 20 20 0 0 -20 -40 First user Second user First entry Second entry

Choice of decision rule Distance calculation (Euclidean) Support vector machine Gaussian mixture model Nearest neighbor method Experimental method found the threshold of the Euclidean distance, which equals 0, 4. Calculation of the Euclidean distance Yes No> 0, 4 Not identification Identification

Probability of a type I error Formula Probability of a type II error N / N *, M / M *, where N * is the total number of intruders (20 people), N is the number of intruders who received unauthorized access (1 person) Calculation where M * is the number of attempts identification of the same user (20 attempts), M is the number of unsuccessful identification attempts (0 attempts) 5% 0% Type I errors - an intruder gains access to the system; Error of the second kind - the registered user cannot access the system.

Program interface 1 2 3 4 1 - toolbar (exit from the program, help about the program); 2 - login input field; 3 - "identification" button, followed by the recording of the user's speech; 4 - registration of a new user, with login and speech recording.

The result of the program The result of the program is the display of two spectrograms (recorded voice from the database) and information about the coincidence or non-coincidence of speech signs, respectively, identification or non-identification:

Conclusions The theoretical foundations of biometric voice identification have been studied; The existing methods of implementation of the software means of personal identification by voice are analyzed; A software tool for personal identification by voice has been developed; The cost of the software is calculated - 32,639 rubles; Assigned to the second class of labor intensity (permissible) when developing software.

Most people in developed countries take the ability to prove their identity for granted. This simple ritual of modern society allows you to perform many transactions that would otherwise be impossible - for example, open bank accounts or obtain a driver's license. But millions of people from developing countries are deprived of such an opportunity and remain hostages of the technological "Middle Ages".

For example, many of the poorest people in India do not have a passport, driver's license or proof of residence. They often live in villages where the names and surnames of many people are the same. As a result, for them such simple tasks as obtaining a check for payment of social needs in electronic form or a loan from a mobile operator are associated with numerous difficulties or even become an insoluble problem. And rogue entrepreneurs take advantage of the lack of a reliable identification system by creating "ghost workers" to receive funds from government public works programs.

However, all of this is a thing of the past thanks to the National Biometric Identity Program - an ambitious project that, when fully implemented, will require 10 times more storage capacity than Facebook's. Known as UIDAI (Unique Identification Administration of India), it is unquestionably the largest identification program in the world, already reaching over half of India's 1.2 billion people.

Despite political and ideological opponents, the project finds support among the poorest strata of the population, who see it as an opportunity to strengthen ties with the state, which can improve their standard of living. In rural India, people came to check-in in large groups, waiting in lines for long periods of time while the “registrars” with their laptops took pictures of everyone, scanned their fingerprints and irises. After entering the system, the person is assigned a 12-digit individual identification number.

To date, about 700 million people in India have been registered, which is almost equal to the population of Europe. This is all the more remarkable when you consider that the program is voluntary. By July 2015, it is planned to register the entire population of India.

“India is a socially oriented state, where a large number of benefits are provided to the population within the framework of subsidy programs,” said Ashok Dalwai, deputy general director of the UIDAI Technology Center in Bangalore. - Receipt of benefits is based on the identification of the person. However, due to inadequate identification systems, many of the poor are generally unable to receive benefits, while others receive it multiple times. Our project is designed to solve this problem. "

Compared to the UIDAI system, the world's largest biometric databases, including those of the FBI and US visitor programs (which have 120 million entries), appear to be dwarfs. “UIDAI performs hundreds of millions of transactions, working on billions of records and performing hundreds of trillions of biometric comparisons every day,” says Dalvai.

UIDAI is not just about size. This is a truly revolutionary technological solution. Based on an open source internet class network, the system is capable of expanding as the population grows. Its main requirements are security and confidentiality. Biometric data is transmitted in encrypted packets and stored in a data center with a triple layer of protection. At the same time, the procedure for verifying and establishing a person's identity is simple and low-cost: you only need a cell phone, smartphone, tablet or other connected device. No printed documents or smart cards are required.

With its open architecture, standard APIs, and developer independence, UIDAI is essentially an identity platform suitable for use by third-party applications that require identity verification. According to Dalvai, over time, it is planned to create conditions for the development of an ecosystem of third-party applications.

Some applications have already been developed and implemented - for example, the electronic know-your-customer (eKYC) platform used to authenticate passports and open bank accounts, a connected payment system (which directs benefits such as pensions and school scholarships directly to beneficiaries ) and government systems for distributing food, kerosene and other necessities to the poor.

Deploying such complex applications and incorporating them into the UIDAI system is just one of the many challenges facing the program, Dalvai said. It is also worth noting the provision of cost-effective standardization of a huge number of authentication devices and solving the problems of connection in a huge country, where only 1.1% of residents have fixed broadband Internet access. (India is ranked 122nd in the world in terms of fixed broadband connectivity, according to a 2013 report by the Broadband Commission.)

The other concerns are not technology-related, but persuading people to volunteer and allay their privacy and data protection concerns. Despite this, Dalvai is confident that people will appreciate the many benefits of the UIDAI system: "The most optimistic scenario is to deploy a large number of applications for social programs and other areas where identification is required."

Tags: India, biometric identification.

About Cisco

Cisco, the global leader in information technology, is helping companies seize the opportunities of the future and are proving by example that connecting the unconnected can achieve amazing results.

The company's net sales in fiscal 2014 were $ 47.1 billion. Information about solutions, technologies and current activities of the company is published on the websites www.cisco.ru and www.cisco.com.

Cisco, the Cisco logo, Cisco Systems, and the Cisco Systems logo are registered trademarks of Cisco Systems, Inc. in the United States and some other countries. All other trademarks mentioned in this document are the property of their respective owners.

Identification is the establishment of the identity of various phenomena, objects, things, persons according to their characteristic individual, inherent only features, as well as the ability of these objects to display (reflect) their properties in other objects when interacting with each other.

The investigation of any criminal or civil case begins with the identification of the victim, the suspect and other persons involved in the case. In the overwhelming majority of cases, this task is resolved by the investigating authorities by examining the certification documents. However, in a number of cases, the suspect submits forged documents, for example, in order to evade responsibility for a serious crime committed, from paying alimony, etc.

Unable to establish the identity of such a person through investigative actions, investigators resort to the help of experts: criminologists and forensic doctors. The need for forensic medical knowledge arises both when examining living people, and, if necessary, to establish the identity of a person when examining the corpses of unknown persons, dismembered and skeletonized remains. Scientific data on the medical and medico-biological properties of a person, their representations, methods of detection, research methods and assessment criteria constitute the content of this section of forensic medicine.

Forensic identification of a person is a complex of techniques and methods aimed at establishing the belonging of the remains to a particular person using a verbal portrait, dental status, special signs (congenital anomalies, traces of healing of injuries, medical interventions, tattoos, birthmarks, etc.), genetic analysis of tissues in comparison with intravital documented data of a medical and anthropological nature.

Forensic medical examination is of leading importance in establishing the identity of unidentified corpses. The identification process consists in comparing biomedical data on wanted persons and found corpses.

So, it is necessary to distinguish between forensic and forensic identification.

Forensic identification is the process of establishing the identity of a specific object by the totality of its general and particular features, carried out by comparative research in order to obtain operational or forensic evidence.

Identity, identity of material objects is determined by a number of signs in which the properties of these objects are manifested. An identification complex is only that set of features that, being identified and thoroughly studied, can serve as a sufficient basis for a specific identification of an object. The essence of identification signs (see Identification signs) lies in their relative stability, originality, expressed in deviations of signs from the average, typical values or norms, rare frequency, their specificity for a given object, and finally, availability for identification, study and comparative research.

The identification process is based on a comparison of the set of identification features, a qualitative and quantitative assessment of the coincidences and differences of the compared features and their displays on the compared objects.

The basic principles of forensic identification can be expressed as follows:

- The identification process is defined (for which it is necessary to resolve the issue of identity) and defined through the use of objects;
- Objects of identification are divided into variables and relative immutable, which are stable at the time that is determined by their identity;
- The identification process includes analysis - deep study of objects and their properties, using methods and techniques that complement each other and provide objective information about the object, and synthesis - comparison and evaluation of objects in their synthetic unity;
- Comparison of each characteristic should be considered in dynamics, how the identified objects can act in different forms;

In addition, one should bear in mind the variability of symptoms depending on time and the possibility of deliberate distortion of signs.

The basic principles of forensic identification are fully related to the identification of the object of examination. Define a person - identification of a specific person on the set of all properties and features that distinguish him from others.

One of the main tasks of the preliminary investigation is to establish the identity of the victim or suspect in the commission of a crime.

In investigative practice, it becomes necessary to establish the face of a living person (for example, a criminal who is hiding from the investigating authorities; a detainee refuses to give his name and surname or deliberately distorts them; a convict evading punishment) or a corpse - an unknown and unidentified subject who died from violent or sudden death.

The possibilities of identifying a person, both a living person and a corpse, are based on the individual uniqueness of the characteristics of each person. These include: gender, age, race, anatomical features, anthropometric indicators, antigenic properties, the presence of certain diseases, traces of various injuries, changes due to the profession, tattoos, etc.

When the bodies of unknown or unidentified persons are found, police officers, with the participation of a forensic expert, carry out their mandatory fingerprinting. If the deceased was previously subjected to fingerprinting, then his identity can be established in this way. The location of the papillary patterns on the soles of the feet is also strictly individual for each person.

The procedure for examining the place where the corpse of an unknown person was found, the organization and tactics of subsequent identification actions, and the registration of unidentified corpses throughout Russia are regulated in detail by the Instruction “on the organization and tactics of identifying citizens from unidentified corpses of patients and children who, due to health or age, cannot report information about myself ", approved by the Ministry of Internal Affairs of Russia and agreed with the General Prosecutor's Office of the Russian Federation and the Ministry of Health of the Russian Federation (1986).

It is important to immediately record the features of the verbal portrait as fully as possible at the place of discovery of the corpse, since facial features change very quickly due to posthumous processes.

As techniques and means of summing external signs, they use, in particular, the compilation of "synthetic" and drawn portraits and so-called composite images, which are used to search for identified persons. Composite (synthetic) portraits and "sketches" are produced by a forensic expert from many fragments of photographs of different faces using the composition method. Painted portraits are made by artists from the words of persons who are well aware of the signs of the wanted person.

By clicking on the "Download archive" button, you will download the file you need for free.
Before downloading this file, remember about those good abstracts, tests, term papers, theses, articles and other documents that are unclaimed on your computer. This is your work, it must participate in the development of society and benefit people. Find these works and submit to the knowledge base.
We and all students, graduate students, young scientists who use the knowledge base in their studies and work will be very grateful to you.

To download an archive with a document, in the field below, enter a five-digit number and click the "Download archive" button

Benefits of biometric identification

Biometric security is more effective than, for example, passwords, smart cards, PINs, tokens, or public key infrastructure technology. This is due to the ability of biometrics to identify not a device, but a person.

Conventional security methods are fraught with loss or theft of information, which is exposed to illegal users. An exclusive biometric identifier, such as a fingerprint, is a key that cannot be lost.

Classification of biometric methods

By the type of information used, biometric identification is divided into:

Static methods based on unique properties given to a person from birth and inalienable from him. Physiological indicators (geometry of the palm or papillary pattern of the fingers) are unchanged for a person '
Dynamic methods based on the behavioral (that is, dynamic) personality characteristics. These features are typical for subconscious movements when reproducing any actions (speech, signature, keyboard dynamics). Such behavioral characteristics are influenced by controlled and not very controllable mental factors. Because of their volatility, biometric samples must be updated as they are used.

Personal identification methods based on biometric parameters

This identification method is the most common. He uses the uniqueness of the papillary patterns of the fingers for each person. A special scanner receives an image of a fingerprint. It is transformed into a digital code and matched against the pattern entered earlier.

The identification process takes no more than a few seconds. A certain drawback holding back the development of this method is the prejudice of some people who do not want to leave their fingerprints. The counter-argument of the hardware developers is that information about the papillary pattern is not stored, but only a short identification code is stored, built on the fingerprint and does not allow the pattern to be recreated for comparison. The advantage of the method is ease of use, reliability and convenience.

Identification by the shape of the hand

This static method is based on measuring the shape of the hand. It is also a unique biometric parameter of a person. A special device allows you to get a three-dimensional view of the brush. As a result, measurements are obtained to create a unique digital code that identifies a person.

This method in its technology and accuracy is comparable to the fingerprint identification method, although the device itself for implementing the method takes up a lot of space. It is extremely unlikely that there are two identical hands with the same geometry, although hands change with age.

Today, hand geometry identification is used in legislatures, hospitals, international airports, and more.

Iris Authentication

The basis of this method is the exclusivity of the pattern on the iris of the eye. To perform it, you need a camera to get an image of the eye with sufficient resolution, and special software to extract the pattern on the iris from the resulting image. On it, a digital code is created that serves to identify a person.

The advantage of scanners is that the person is not required to focus on the target, since the iris pattern is concentrated on the surface of the eye. Scanning is possible at a distance of less than 1 m. It is convenient for use, for example, in ATMs.

Retinal identification

The retina is scanned using low-intensity infrared light, which is directed to the blood vessels in the back of the eye through the pupil. Retinal scanners are widespread in secret object access systems, since they almost never have incorrect access permissions. Errors can be attributed to head deviation from the reference position and incorrect focus of the gaze on the light source.

Even in twins, the capillary pattern of the retina is different. That is why this method can be successfully used for personal identification.

The disadvantage of such systems can be attributed to the psychological factor: not every person can look into a dark hole in which something shines into the eye. In addition, these systems are sensitive to incorrect orientation of the retina, so one must carefully monitor the position of the eye in relation to the opening.

The form faces as an object for identification

This static identification method consists in creating a two- or three-dimensional image of a person's face. The camera and specialized software emphasize the contours of the eyes, lips, eyebrows, nose, etc. on the face image. Then the distances between these elements and other parameters are calculated. Based on this information, an image is created, which is digitized for comparison.

This method is one of the fastest growing areas in the biometrics industry. Its appeal is based on the fact that no special expensive equipment is required. A personal computer and a video camera are enough. In addition, there is no physical contact with the devices. You do not need to touch anything, or stop, specifically waiting for the system to work.

Handwriting recognition

The basis of handwriting identification is the uniqueness and stability of this factor for each person. Characteristics are measured, digitized, and computer processed. That is, it is not the letter as a product that is selected for comparison, but the process itself.

Two methods of data processing are common: regular pattern matching and dynamic verification. The first one is unreliable because the signature is not always the same. This method leads to a large percentage of errors. Dynamic verification consists of more complex calculations. This method registers in real time the parameters of the signature process itself: the speed of hand movement in different areas, the force of pressure and the duration of different stages of the signature. This excludes forgery, since it is impossible to exactly copy the movements of the hand of the author of the signature.

Keyboard handwriting recognition

This method is, in general, similar to the one described above, however, the signature is replaced by a certain code word, and from the equipment only a regular keyboard is needed. The main identification characteristic is the dynamics of the keyboard set of the code word.

According to modern research, keyboard handwriting has a certain stability, so that a person can be uniquely identified. The initial data is the time between pressing the keys and holding them down. Moreover, the time between pressing shows the pace of work, and the holding time shows the style of work, that is, a smooth press or a sharp blow.

First, at the stage of filtering, the data about the "service" keys - function keys, cursor control, etc., is deleted.

Then the following user characteristics are highlighted:

the number of errors in the typing process;
time between keystrokes;
dialing speed.
time to hold keys;
arrhythmia when typing .

Voice recognition

The biometric voice identification method is easy to use. The reasons for its implementation are the widespread use of telephone networks and the embedding of microphones in computers. Factors influencing recognition can be considered disadvantages: interference in microphones, ambient noises, errors in the pronunciation process, different emotional states of a person during identification, etc.

The key to building voice authentication devices is to choose the parameters that best describe the voice personality. These signal parameters are called personality traits. Such signs, in addition to data on the characteristics of the voice, should have other properties. For example, they should be easy to measure, and little dependent on noise and interference. In addition, they must be stable over time and resist imitation.

Systems have been developed using the method of combined analysis of voice and facial expressions. It turns out that the mimicry of the speaker distinguishes only him and will be different for the other person pronouncing the same words.

Thermographic observation of facial arteries and veins

Identification of a person by face is greatly simplified if we switch to the infrared range of light waves. Thermography of the person to be identified reveals the uniqueness of the location of the arteries on the face that supply the skin with blood. The issue of backlighting for these biometric devices does not exist, since they perceive only temperature changes in the face and they do not need light. The recognition efficiency does not depend on overheating or hypothermia of the face, natural aging of the person, plastic surgeries, since they do not change the internal position of the vessels.

Facial thermography can be used to distinguish twins whose facial blood vessels are very different.

This identification method uses a specialized far infrared video camera.

Hand vein identification

There are devices on the biometric market that are based on the analysis of the individual location of the veins in the arms. The pattern of veins located on the back of the hand, clenched into a fist, is taken into account. The pattern of veins is monitored by a television camera with infrared illumination. When the image is entered, its binarization is performed, highlighting the veins. Such equipment is produced by the only British company Vinchek.

Prospects for biometrics

Fingerprint recognition remains the dominant method of personal identification. There are two main reasons for this:

the transition to biometric passports has begun in many countries;
development of updated models of fingerprint scanners for use in small devices (cell phones, pocket PCs, laptops).

Significant expansion can be expected in the signature identification sector due to the widespread adoption of digital electronic signatures. Voice recognition can also gain traction with major smart building projects.

The main forecasts boil down to the fact that the introduction of biometric security devices in the near future will acquire an avalanche nature. The fight against global terrorism will require the practical use of any achievements in this area. Thanks to the intensive development of multimedia and digital technologies and their further reduction in cost, they will allow the development and implementation of fundamentally new identification systems.

Certain biometric technologies are currently under development and some of them are recognized as promising:

infrared facial thermogram;
DNA characteristics;
finger skin spectroscopy;
handprints;
the shape of the auricle;
human gait parameters;
individual human smells;
the level of salinity of the skin.

These methods of biometric identification today can be considered mature. Perhaps soon they will move from scientific research to commercial technologies.

Similar documents