RU2644525C2 - Method and system of identifying a living person in the sequence of images by identifying a pulse at separate parts of a face - Google Patents

Abstract

FIELD: computer engineering.

SUBSTANCE: invention relates to computer engineering and specifically to image processing systems. Method of identifying a living person on a sequence of frames by detecting the pulse in individual areas of a person's face contains: obtaining a sequence of frames; highlighting a person's face in the image of each frame; splitting the area of the image of the selected face into segments in the form of a grid, and more than half of the image is divided into cells that differ in area by no more than twice; carrying out the temporal and spatial filtering of the face image segments obtained in the previous step; determining the presence of periodicity in changing the magnitude of the brightness of image segments in the entire sequence of frames, which indicates the presence of a pulse; determination of the correspondence between the periodicity of the variation in the luminance of segments and the pulse rate of a person; making a decision about the presence of a living person in the case of coincidence of the periodicity of the change in the magnitude of the brightness of image segments to the pulse rate of a living object.

EFFECT: technical result is to increase the accuracy of identifying a living person on the sequence of personnel.

21 cl, 11 dwg

Description

FIELD OF TECHNOLOGY

[0001] The invention relates to computer technology, in particular to methods and systems for detecting a living person by detecting a pulse in certain areas of a person's face.

BACKGROUND

[0002] Modern access control systems use biometric information about a person to identify registered users and deny access to unauthorized users.

[0003] One of the features of the development of modern society is the growing need to restrict access to various facilities, such as offices, warehouses, ATMs, military facilities. This is due to the need to ensure the security of these objects, to prevent theft of intellectual property and goods.

[0004] Previously widely used security systems based on the presence of a person, such as a security guard, at the entrance to a protected area or using video surveillance systems, the data from which are also analyzed by a person, become expensive and do not provide the necessary degree of reliability. It is in connection with these shortcomings in many public and private organizations that systems based on identification magnetic and smart cards have become widespread. The card contains information identifying the user using it. The scanner located at the entrance to the secure room reads this information and makes a decision on providing the user access to the room. A significant drawback of such systems is the ability to use the card by an attacker, for example, he can steal it.

[0005] Biometric systems are one of the most promising solutions devoid of these disadvantages. Such systems are based on the analysis of biometric information about the user: facial features, voice, gestures, fingerprints, etc. User biometric parameters automatically read by the system are compared with templates stored in the database. If one of the templates matches the received data, then the user is considered identified and he is allowed access.

[0006] Modern biometric systems are based on the analysis of the following biometric parameters of a person: face, voice, iris, gestures. Other types of parameters either do not provide sufficient accuracy for user identification or require contact with a reader (as is the case with fingerprints).

[0007] However, existing biometric systems have a number of disadvantages that allow unauthorized access to a protected area.

[0008] The closest analogue to patent No. 2316051 “Method and system for automatically checking the presence of a living person in biometric security systems”, patent holder: SAMSUNG ELECTRONICS, published: 01/27/2008.

[0009] The invention relates to safety and control systems. Its use allows to obtain a technical result in the form of increasing reliability and speed in detecting attempts of unauthorized access to an object. This is achieved due to the fact that as the main mechanism they use the detection of the face of a living person and the detection of unauthorized users present next to the registered user. The invention uses methods for tracking a three-dimensional object reduced to the first normalized face shape, using a quick method for measuring and comparing facial expressions with a template, as well as methods for detecting local features and representing the face in three different normalized forms. In addition, they use a quick method of measuring and comparing with a template such a behavioral biometric characteristic as a phoneme signature obtained as a result of user executing system commands.

[00010] The disadvantage of this technical is that it fixes only the presence on the image of the image of the face and its inherent physical aspects, such as facial expressions. This method does not allow to distinguish the generated image from the image of a real living object.

[00011] Also known from the prior art are solutions offering algorithms and control systems based on the analysis of two-dimensional images of a human face (see published US patents Nos. 6,633,655 and 6,681,032). These systems use video cameras to capture face images, realize the allocation of facial areas and calculate their characteristics, which are compared with templates stored in the database. Such systems process 2D models of a person’s face and thereby make it possible to gain unauthorized access by providing cameras with photos of a registered user.

ESSENCE OF TECHNICAL SOLUTION

[00012] This technical solution is aimed at eliminating the disadvantages inherent in existing solutions.

[00013] The technical problem of the claimed invention is the creation of a method and system for recognizing biometric information by determining the heart rate in individual parts of an image of a living person by amplifying small color changes in the image. Often, biometric data are used for identification purposes, and biometric data that changes in dynamics, such as facial expressions, is used for added security. But even this approach does not provide comprehensive protection. The solution is to determine the heart rate in specific areas of the image when changing facial expressions, which allows us to conclude that the image is a real person, not a generated image.

[00014] The technical result manifested in solving the specified technical problem is to increase the possibility of identifying a living person.

[00015] Available in the prior art solutions do not allow to distinguish the generated image from the real image of a living object. While the proposed method allows to distinguish the generated image from the real object with a probability of 90-97% depending on external conditions (for example, illumination).

[00016] The indicated technical result is achieved due to the method for detecting a living person in a sequence of frames by detecting a pulse in separate parts of a person’s face in which a sequence of frames is obtained; distinguish a person’s face in the image of each frame; break the image area of the selected face into segments in the form of a grid; conduct temporary and spatial filtering of the face image segments obtained in the previous step; determine the presence of periodicity in changing the brightness of image segments in the entire sequence of frames, which indicates the presence of a pulse; determine the correspondence of the periodicity of changes in the magnitude of the brightness of the segments to the pulse rate of a person; make a decision on the presence of a living person in case of coincidence of the periodicity of the change in the brightness value of the image segments and the pulse rate of a living object.

[00017] In some embodiments, a sequence of frames is obtained in the form of streaming video coming from a real-time video source or from storage from a local video server or central server.

[00018] In some embodiments, a frame sequence is obtained over the real-time streaming protocol RTSP and / or RTMP and / or HLS and / or DASH.

[00019] In some embodiments, a sequence of frames is obtained in the form of a list of video sources of various quality and bit rate.

[00020] In some embodiments, a compressed sequence of frames is obtained using an H.264 encoder and / or VP8 and / or MJPEG and / or JPEG and / or PEG2000.

[00021] In some embodiments, a frame sequence is obtained over wireless networks, such as GSM, or CDMA, or LTE, or Wi-Fi.

[00022] In some embodiments, when isolating a person’s face in an image, the Viola Jones method, or a neural network, or flexible graph comparison method, or hidden Markov models are used.

[00023] In some embodiments, when dividing the image area of the selected face into grid segments, the shape of the segment is round, or triangular, or square, or rectangular.

[00024] In some embodiments, when dividing the image area of the selected face into grid segments, more than half of the image should be divided into cells that differ from each other in area by no more than two times.

[00025] In some embodiments, when performing temporal and spatial filtering of face image segments, low frequencies are filtered in a sequence of frames and the resolution of their computational efficiency is reduced.

[00026] In some embodiments, when performing temporal and spatial filtering of face image segments, a full Laplace pyramid is calculated.

[00027] Also, the indicated technical result is achieved thanks to a system for detecting a living person in a sequence of frames by detecting a pulse in separate parts of a person’s face, comprising: a processor; memory for storing instructions executed by the processor; wherein the processor is configured to: obtain a sequence of frames; highlighting the person’s face in the image of each frame; dividing the area of the image of the selected face into segments in the form of a grid; conducting temporal and spatial filtering of face image segments obtained in the previous step; determining the presence of periodicity in changing the brightness value of image segments in the entire sequence of frames, which indicates the presence of a pulse; determine the correspondence of the periodicity of changes in the magnitude of the brightness of the segments to the pulse rate of a person; making decisions on the presence of a living person in case of coincidence of the periodicity of changing the magnitude of the brightness of the image segments and the pulse rate of a living object.

[00028] In some embodiments, the processor is application or graphics.

[00029] In some embodiments, the processor receives a sequence of frames in the form of streaming video coming from a real-time video source or from storage from a local video server or central server.

[00030] In some embodiments, the processor receives a frame sequence over the real-time streaming protocol RTSP and / or RTMP and / or HLS and / or DASH.

[00031] In some embodiments, the processor receives a sequence of frames in the form of a list of video sources of various quality and bit rate.

[00032] In some embodiments, the processor receives the sequence of frames in compressed form using an H.264 encoder and / or VP8 and / or MJPEG and / or JPEG and / or PEG2000.

[00033] In some embodiments, the processor receives a frame sequence over wireless networks, such as GSM, or CDMA, or LTE, or Wi-Fi.

[00034] In some embodiments, when the processor selects a person’s face on the image, use the Viola Jones method, or the neural network, or the method of flexible comparison on graphs, or hidden Markov models.

[00035] In some embodiments, when the processor splits the image area of the selected face into grid segments, the shape of the segment is round, or triangular, or square, or rectangular.

[00036] In some embodiments, when the processor splits the image area of the selected face into grid segments, more than half the image should be divided into cells that differ from each other in area by no more than two times.

[00037] In some embodiments, when a processor performs temporal and spatial filtering of facial image segments, low frequencies are filtered in a sequence of frames and the resolution of their computational efficiency decreases.

[00038] In some embodiments, when a processor performs temporal and spatial filtering of face image segments, a full Laplace pyramid is calculated.

BRIEF DESCRIPTION OF THE DRAWINGS

[00039] The features and advantages of the present invention will become apparent from the following detailed description of the invention and the accompanying drawings, in which:

[00040] in FIG. 1 shows an example of the implementation of a technical solution according to a method for detecting a living person in a sequence of frames by detecting a pulse in separate parts of a person’s face;

[00041] in FIG. 2 shows the principle of operation of a method for amplifying small changes in a video stream;

[00042] in FIG. 3 illustrates an approximation of spatial displacement using temporal filtering;

[00043] in FIG. 4 shows an illustration of amplification of motion along a 1D signal for various spatial frequencies;

[00044] in FIG. 5 shows the motion of the magnification error;

[00045] in FIG. 6 shows the gain, but, as a function of the spatial wavelength, to enhance the motion;

[00046] in FIG. 7 shows the architecture of a convolutional neural network;

[00047] in FIG. 8 shows an example of dividing a person’s face into segments;

[00048] in FIG. 9 shows an example of temporal and spatial filtering of face image segments;

[00049] in FIG. 10 shows an example of the implementation of the dynamics of changes in brightness on the studied 120 pictures obtained in 5 seconds (24 frames per second);

[00050] in FIG. 11 shows an example of the implementation of a technical solution according to a system for detecting a living person in a sequence of frames by detecting a pulse in certain areas of a person's face.

DETAILED DESCRIPTION OF THE TECHNICAL SOLUTION

[00051] This technical solution can be implemented on a computer, in the form of a system or computer-readable medium containing instructions for performing the above method.

[00052] The technical solution may be implemented as a distributed computer system.

[00053] In this solution, a system means a computer system, a computer (electronic computer), CNC (numerical control), PLC (programmable logic controller), computerized control systems, and any other devices that can perform a given, well-defined sequence of operations (actions, instructions).

[00054] An instruction processing device is understood to mean an electronic unit or an integrated circuit (microprocessor) that executes machine instructions (programs).

[00055] An instruction processing device reads and executes machine instructions (programs) from one or more data storage devices. Hard disk drives (HDD), flash memory, ROM (read only memory), solid state drives (SSD), optical drives can be used as storage devices.

[00056] A program is a sequence of instructions for execution by a computer control device or an instruction processing device.

[00057] Below will be described the concepts and terms necessary for a detailed description and implementation of a technical solution.

[00058] A frame is a single image limited to a certain size on a photo or film, on a television or movie screen, a separate photographic image.

[00059] The complete Laplace pyramid is a mathematical term for a subband signal transform.

[00060] A segment is a part of the whole, limited in size.

[00061] Segment brightness is the amount of light energy emitted / reflected by an image segment.

[00062] Spectral characteristic - part of the spectrum of white light that is emitted, transmitted or absorbed by a radiation source, substance or surface

[00063] Filtering is the conversion of signals represented in digital form.

[00064] Spatial filtering is the transformation of a signal in the time domain.

[00065] Image intensity is a scalar physical quantity that quantitatively characterizes the radiation power received from an object in the image.

[00066] Motion signal - the fact of a change in the geometric properties of an object

[00067] Gain - a quantitative characteristic showing how many times a particular parameter has been amplified

[00068] Convolution of a spatial pyramid is a mathematical operation for converting a signal for a spatial pyramid.

[00069] Modulation of spectral components - the process of changing the parameters of the spectral characteristic

[00070] Linear approximation - approximation of an arbitrary function using a linear function

[00071] According to the claimed technical solution, a method for detecting a living person in a sequence of frames by detecting a pulse in separate parts of a person’s face, which is shown in FIG. 1 is implemented as follows.

[00072] Step 101: obtaining a sequence of frames.

[00073] The sequence of frames may be transmitted in the form of streaming video coming from a video source in real time or from storage from a local video server or central server. For streaming video, standard RTSP (Real Time Streaming Protocol), RTMP (Real Time Messaging Protocol), HLS (HTTP Live Streaming) and DASH (Dynamic Adaptive Streaming over HTTP) protocols can be used. At the same time, the speed and quality of the transmitted video data can automatically adapt to the communication channel of the mobile device.

[00074] The frame sequence may be transmitted in compressed form, for example, using H.264, VP8, MJPEG, JPEG, JPEG2000 encoders.

[00075] The frame sequence may be transmitted as separate files. In this case, standard containers can be used, for example WebM, OGV, MKV, MP4, TS, JPG, etc.

[00076] The frame sequence may be transmitted as a list of video sources of various quality and bit rate, for example, in the M3U8 format.

[00077] The frame sequence can be transmitted over wireless networks such as GSM (Global System for Mobile Communications), CDMA (Code division multiple access), LTE (Long Term Evolution), Wi-Fi (Wireless Fidelity). In some embodiments, the implementation of the present invention, the receipt and / or sending of data is carried out using several technologies described above, or technologies for receiving / transmitting data that will be invented after filing the application for the present invention.

[00078] Step 102: highlighting a person’s face in the image of each frame.

[00079] A person’s face is identified in the image of each frame obtained earlier by the face extraction algorithm, which is, but not limited to, the Viola Jones method, neural networks (Fig. 7), flexible comparison method on graphs, hidden Markov models (CMM, HMM).

[00080] Step 103: divide the image area of the selected face into grid segments.

[00081] A face is divided into segments that are selected by dividing the face image with a grid whose segment area is not more than one hundredth of the image area. Segments can be of any shape, the main factor is the full coverage of the face segments. The shape of the cell is also not fundamental. Moreover, more than half of the image should be divided into cells that differ from each other in area by no more than two times. The vertices of the grid can be selected arbitrarily.

[00082] Step 104: conduct temporary and spatial filtering of the face image segments obtained in the previous step.

[00083] This step is necessary to highlight subtle temporal changes in the sequence of frames, which combines spatial and temporal processing (Fig. 2). First, the sequence of frames is laid out in different ranges of spatial frequencies. The frequency ranges can be increased, because in a different way (a) they could have different signal-to-noise ratios or (b) they could contain spatial frequencies that do not have the linear approximation used in our algorithm to increase movement. In the latter case, the gain for these bands decreases in order to suppress artifacts. When the goal of spatial processing is simply to increase the temporal signal-to-noise ratio by combining several pixels, low frequencies are filtered in a sequence of frames and the resolution of their computational efficiency is reduced. In the general case, however, the full Laplace pyramid is calculated.

[00084] The filtered spatial bands are then amplified by a predetermined number of times, which is determined by manual or machine search in the range from 0 to 100, until the oscillations in the sequence of frames become distinguishable, and again added to the original signal to generate the output video. The choice of a temporary filter can be configured to support various applications. Then, time processing of each spatial range is performed. Consider a time series corresponding to a pixel value in a frequency band, and apply a band-pass filter to extract the frequency ranges of interest. Temporal processing is the same for all spatial levels and for all pixels in each level. Then, the extracted signal is multiplied by the gain a. This value can be set by the user and can be weakened automatically. Next, an enlarged signal is added to the original, and the spatial pyramid is convolved to obtain the final result. Since real videos in space and time are constant and since the filtering described above is performed uniformly for all pixels, this method implicitly supports the principle of spatio-temporal coherence of results.

[00085] Step 105: determining the presence of periodicity in changing the magnitude of the brightness of image segments in the entire frame sequence, which indicates the presence of a pulse.

[00086] To explain the relationship between time processing and the increase in small movements, consider a simple case of a 1D signal that undergoes translational motion. This analysis directly generalizes locally translational motion in 2D. Let I (x, t), where I denote the intensity of the image at position x and time t. Since the image undergoes translational motion, we can express the observed intensities with respect to the displacement function δ (t), so that I (x, t) = F (x + δ (t)) and I (x, 0) = F ( x). The purpose of increasing motion is to synthesize a signal with a certain gain a.

[00087]

.[00088] Assuming that the image can be approximated through first-order Taylor expansions, we write the image at time T, F (x + δ (T)) in the first-order Taylor extension in x, as

.

[00089]

[00090] Let B (xt) be the result of applying the broadband temporary band-pass filter I (X, T) for any position x (selecting all but F (X) in equation 2. Assume that the motion signal δ (T) is within the passband of the time-domain bandpass filter.Then we have Df (x)

[00091]

[00092] In our process, after amplification of the bandpass signal, we add it to I (x, t), as a result of which in the processed signal

[00093]

[00094] After combining the formulas 2, 3 and 4, we have

[00095]

, можно связать усиление с временным увеличением сигнала движения. Обработать выходной сигнал легко.[00096] Assuming that the first Taylor expansion is performed to amplify large perturbations,

, you can associate the gain with a temporary increase in the motion signal. Handling the output is easy.

[00097]

[00098] This shows that processing increases the spatial displacement movement δ (T) of the local image F (X) at time t when amplified to a value of (1 + α).

[00099] This process is illustrated in one of the sinusoids in FIG. 3. For a low-frequency wave, a cosine wave is observed and a relatively small volume is characteristic.

[000100] FIG. 3: temporal filtering can approximate spatial displacement.

в момент времени t и

в момент времени

.[000101] This effect is demonstrated here on a 1D signal, but equally applies to 2D signals. The input signal is displayed at two points in time:

at time t and

at time

.

. При повышении временной составляющей сигнала по а и последующем добавлении его обратно в I(X, T) мы приближаем эту волну смешения к (1+α)δ. Для полноты рассмотрим более общий случай, когда δ(T) находится не полностью в пределах полосы пропускания временного фильтра. В этом случае пусть δk(T) индексируется по k, представляя различные временные спектральные компоненты δ(t). Каждый δk(t) будет ослабляется временной фильтрацией с коэффициентом γk. Это приводит к полосовому сигналу.[000102] Taylor's first order expansion serves as a good approximation for a biased signal over time

. With increasing the time component of the signal with respect to a and then adding it back to I (X, T), we bring this mixing wave closer to (1 + α) δ. For completeness, we consider the more general case when δ (T) is not completely within the bandwidth of the time filter. In this case, let δk (T) be indexed by k, representing various temporal spectral components δ (t). Each δk (t) will be attenuated by time filtering with coefficient γk. This results in a bandpass signal.

[000103]

[000104] Due to the multiplication in Equation 4, we get the attenuation, while the temporal frequency can be equivalently interpreted as a frequency-dependent motion with an increase factor aK = γkα, resulting in increased motion at the output.

[000105]

[000106] The result, as might be expected for linear analysis, was that the modulation of the spectral components of the signal motion becomes a modulation factor with the amplification of the motion ak for each time subzone δk of the motion signal.

[000107] Step 106: determine the correspondence of the periodicity of the change in the magnitude of the brightness of the segments to the pulse rate of a person.

[000108] A spectral characteristic is constructed in which a bandwidth of 0.66-3.33 Hz is selected by band-pass filtering, which corresponds to a human pulse frequency of 40-200 beats per minute. If there are repeated changes in spectral power in this range, with an average difference of at least 30% from the maximum values, a conclusion is made about the presence of a periodic change corresponding to the human pulse frequency.

[000109] Step 107: decide on the presence of a living person in case of coincidence of the periodicity of the change in the brightness value of the image segments and the pulse rate of the living object.

[000110] This conclusion is made when detecting a periodic change in the sequence of frames corresponding to the frequency range of the pulse of the person described in step 106.

[000111] FIG. 11 is a block diagram showing a system for detecting a living person in a sequence of frames by detecting a pulse in separate areas of a person’s face according to yet another embodiment of the invention. For example, system 1100 may be a mobile phone, tablet computer, digital broadcast terminal, messaging device, game console, tablet, medical device, exercise equipment, and personal digital assistant, etc.

[000112] Referring to Figure 11, a system 1100 may include one or more of the following components: processing component 1102, memory 1104, power component 1106, multimedia component 1108, audio component 1110, input / output (I / O) interface 1112, touch component 1114, data transmission component 1116.

[000113] The processing component 1102 mainly controls all operations of the system 1100, such as a display, a phone call, data transfer, camera operation, and recording operation. Processing component 1102 may include one or more processors 1118 that implement instructions for completing all or part of the steps of the above methods. In addition, the processing component 1102 may include one or more modules for a convenient interaction process between the processing components 1102 and other components. For example, the processing component 1102 may include a multimedia module for conveniently facilitating interaction between the multimedia component 1108 and the processing component 1102.

[000114] The processor 1118 may be application or graphics.

[000115] The memory 1104 is configured to store various types of data to support the operation of the system 1100. Examples of such data include instructions from any application or method, contact data, address book data, messages, images, videos, etc. Memory 1104 may be implemented as any type of non-volatile memory, non-volatile memory, or a combination thereof, for example Static Random Access Memory (RAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Storage Device (EEPROM), Permanent Storage Device (ROM), magnetic memory, flash memory, magnetic or an optical disc.

[000116] Power component 1106 provides electricity to various components of system 1100. Power component 1106 may include a power management system, one or more power supplies, and other nodes for generating, controlling, and distributing power to system 1100.

[000117] The multimedia component 1108 includes a screen providing an output interface between the system 1100 and the user. In some embodiments, the screen may be a liquid crystal display (LCD) or touch panel (SP). If the screen includes a touch panel, the screen may be implemented as a touch screen for receiving an input signal from a user. The touch panel includes one or more touch sensors in the sense of gesturing, touching and sliding the touch panel. The touch sensor can not only feel the boundary of the touch or the gesture of turning over, but also determine the duration of time and pressure associated with the operation mode on touch and sliding. In some embodiments, the multimedia assembly 1208 includes one front camera and / or one rear camera. When the system 1100 is in an operating mode, such as a shooting mode or a video mode, the front camera and / or the rear camera can receive multimedia data from the outside. Each front camera and rear camera may be a single fixed optical system of the lens or may have a focal length or optical zoom.

[000118] The audio component 1110 is configured to output and / or input an audio signal. For example, the audio component 1110 includes a single microphone (MIC) that is configured to receive an external audio signal when the system 1100 is in an operation mode, such as a call mode, a recording mode, and a speech recognition mode. The received audio signal may be further stored in the memory 1104 or sent through the data transmission component 1116. In some embodiments, the audio component 1110 also includes a single speaker configured to output an audio signal.

[000119] The I / O interface 1112 provides an interface between the processing component 1102 and the peripheral interface module. The above peripheral interface module may be a keyboard, steering wheel, button, etc. A button may include, but is not limited to, a homepage button, a volume button, a power button, and a lock button.

[000120] The sensor component 1114 contains one or more sensors and is configured to provide various aspects of assessing the state of the system 1100. For example, the sensor component 1114 can detect on / off states of the device 1100, the relative position of components, such as the display and keypad of the device 1100, change the position of the system 1100 or one component of the system 1100, the presence or absence of contact between the user and the system 1100, as well as the orientation or acceleration / deceleration and temperature change of the system 1100. Sensors component 1114 comprises a proximity sensor configured to detect the presence of an object in the vicinity when there is no physical contact. The sensor component 1114 comprises an optical sensor (e.g., CMOS or CCD image sensor) configured to be used in visualizing an application. In some embodiments, the sensor component 1114 comprises an acceleration sensor, a gyro sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.

[000121] The communication component 1116 is configured to facilitate wired or wireless communication between the system 1100 and other devices. System 1100 can access a wireless network based on communication standards such as Wi-Fi, 2G, or 3G, or a combination thereof. In one exemplary embodiment, the data transmission component 1116 receives a broadcast signal or broadcast, associated information from an external broadcast control system via a broadcast channel. In one embodiment, the data transmission component 1116 comprises a near field communication (NFC) module to facilitate near field communication. For example, the NFC module can be based on radio frequency identification (RFID) technology, infrared data association technology (IrDA), ultra-wideband (UWB) technology, Bluetooth (BT) technology, and other technologies.

[000122] In an exemplary embodiment, system 1100 can be implemented by one or more Specialized Integrated Circuits (ICs), Digital Signal Processors (DSPs), Digital Signal Processing Devices (DSPs), Programmable Logic Devices (PLCs), logic circuits programmed in operating conditions (PPVM), controller, microcontroller, microprocessor or other electronic components and can be configured to implement a method for identifying a living person in a sequence of frames by identifying the pulse in some parts of the human face.

[000123] In an exemplary embodiment, the non-volatile computer-readable medium also includes instructions provided, for example, memory 1104 includes instructions where instructions are executed by a processor 1118 of system 1100 to implement the album display methods described above. For example, a non-volatile computer-readable medium may be a ROM, random access memory (RAM), CD, magnetic tape, floppy disks, optical storage devices and the like.

[000124] In the specified system 1100, the processor is configured to:

[000125] obtaining a sequence of frames;

[000126] highlighting a person’s face in the image of each frame;

[000127] dividing the image area of the selected face into grid segments;

[000128] conducting temporal and spatial filtering of the face image segments obtained in the previous step;

[000129] determining the presence of periodicity in changing the brightness value of image segments in the entire sequence of frames, which indicates the presence of a pulse;

[000130] determining the correspondence of the periodicity of the change in the magnitude of the brightness of the segments to the human pulse rate;

[000131] deciding on the presence of a living person in case of coincidence of the periodicity of the change in the magnitude of the brightness of the image segments and the pulse rate of the living object.

[000132] A detailed description of each step that is performed by the processor is described above with a detailed disclosure of a method for detecting a living person in a sequence of frames by detecting a pulse in certain areas of a person's face.

[000133] A person skilled in the art can easily understand other variations of the invention from the above description of the technical solution. This invention is intended to cover any variations, uses or adaptations of the following general principles of the invention, including those deviations from the present invention that appear within the scope of known or ordinary practice in the art. It is assumed that the description and examples are considered only as exemplary, with the essence and scope of the present invention indicated by the claims.

[000134] It should be appreciated that the present invention is not limited to the precise constructions described above and illustrated in the accompanying drawings, and that various modifications and changes can be made without departing from its scope. It is intended that the scope of the invention be limited only by the appended claims.

SPECIFIC EXAMPLE OF IMPLEMENTATION

[000135] Consider an example of the implementation of the above method of detecting a living person in a sequence of frames by detecting the pulse in certain areas of the person’s face.

[000136] A sequence of 120 frames is obtained.

[000137] On each of the frames, a person’s face is distinguished by using the Viola Jones method, after which the person’s face is broken, as shown in FIG. 8.

[000138] Then, temporal and spatial filtering of the face image segments obtained in the previous step shown in FIG. 9, obtaining the dynamics of changing the brightness of the image pixels.

[000139] Check if there is a periodicity in the brightness of each pixel on the example of the dynamics of brightness in the studied 120 pictures obtained in 5 seconds (24 frames per second) (Fig. 10).

[000140] From the obtained dependence it is seen that there is a periodicity with a frequency of 0.8 Hz. A frequency of 0.8 Hz corresponds to a range of 0.66-3.33 Hz, which indicates that a person’s pulse is recorded in the image. If such a picture of changes in brightness dynamics is observed in more than 95% of the segments, it is concluded that the face of a living person is on the image.

Claims (38)

1. A method for identifying a living person in a sequence of frames by detecting a pulse in certain areas of a person’s face, comprising the following steps: - receive a sequence of frames; - highlight the person’s face in the image of each frame; - divide the image area of the selected face into segments in the form of a grid, more than half of the image is divided into cells that differ from each other in area by no more than two times; - conduct temporary and spatial filtering of the face image segments obtained in the previous step; - determine the presence of periodicity in changing the brightness of image segments in the entire sequence of frames, which indicates the presence of a pulse; - determine the correspondence of the periodicity of changes in the magnitude of the brightness of the segments to the pulse rate of a person; - decide on the presence of a living person in case of coincidence of the periodicity of the change in the magnitude of the brightness of the image segments and the pulse rate of a living object. 2. The method according to p. 1, characterized in that they receive a sequence of frames in the form of streaming video coming from a video source in real time, or from storage from a local video server, or a central server. 3. The method according to p. 1, characterized in that they receive a sequence of frames over the real-time streaming protocol RTSP, and / or RTMP, and / or HLS, and / or DASH. 4. The method according to p. 1, characterized in that they receive a sequence of frames in the form of a list of video sources of various quality and bit rate. 5. The method according to claim 1, characterized in that the sequence of frames in compressed form is obtained using an H.264 encoder, and / or VP8, and / or MJPEG, and / or JPEG, and / or JPEG2000. 6. The method according to p. 1, characterized in that they receive a sequence of frames over wireless networks, such as GSM, or CDMA, or LTE, or Wi-Fi. 7. The method according to claim 1, characterized in that when highlighting a person’s face in the image, use the Viola Jones method, or a neural network, or a method of flexible comparison on graphs, or hidden Markov models. 8. The method according to p. 1, characterized in that when dividing the image area of the selected face into segments in the form of a grid, the shape of the segment is round, or triangular, or square, or rectangular. 9. The method according to p. 1, characterized in that when conducting temporary and spatial filtering of segments of the image of the face low frequencies are filtered in a sequence of frames with a decrease in their resolution. 10. The method according to p. 1, characterized in that when conducting temporary and spatial filtering of segments of the image of the face calculate the full Laplace pyramid. 11. A system for detecting a living person in a sequence of frames by detecting a pulse in certain areas of a person’s face, containing: - processor; - memory for storing instructions executed by the processor, wherein the processor is configured to: - obtaining a sequence of frames; - highlighting the person’s face in the image of each frame; - dividing the area of the image of the selected face into segments in the form of a grid, more than half of the image being divided into cells that differ from each other in area by no more than two times; - conducting temporal and spatial filtering of face image segments obtained in the previous step; - determining the presence of periodicity in changing the brightness of image segments in the entire sequence of frames, which indicates the presence of a pulse; - determine the correspondence of the periodicity of changes in the magnitude of the brightness of the segments to the pulse rate of a person; - making a decision on the presence of a living person in case of coincidence of the periodicity of the change in the brightness of the image segments with the pulse rate of a living object. 12. The system of claim 11, wherein the processor is application or graphics. 13. The system of claim 11, wherein the processor receives a sequence of frames in the form of streaming video coming from a video source in real time or from storage from a local video server or central server. 14. The system of claim 11, wherein the processor receives a sequence of frames over the real-time streaming protocol RTSP, and / or RTMP, and / or HLS, and / or DASH. 15. The system according to claim 11, in which the processor receives a sequence of frames in the form of a list of video sources of various quality and bit rate. 16. The system of claim 11, wherein the processor receives the sequence of frames in compressed form using an H.264 encoder, and / or VP8, and / or MJPEG, and / or JPEG, and / or JPEG2000. 17. The system of claim 11, wherein the processor receives a frame sequence over wireless networks, such as GSM or CDMA or LTE or Wi-Fi. 18. The system according to claim 11, in which, when the processor distinguishes a person’s face in the image, use the Viola Jones method, or a neural network, or the method of flexible comparison on graphs, or hidden Markov models. 19. The system of claim 11, wherein when the processor breaks the image area of the selected face into segments in the form of a grid, the shape of the segment is round, or triangular, or square, or rectangular. 20. The system of claim 11, wherein when the processor performs temporal and spatial filtering of the face image segments, low frequencies are filtered in a sequence of frames with a decrease in their resolution. 21. The system according to claim 11, in which when the processor performs temporal and spatial filtering of segments of the image of the face, the full Laplace pyramid is calculated.

Images