TWI844770B - System for recognizing specified actions in images to determine target is alive and method thereof - Google Patents

Abstract

A system for recognizing specified actions in images to determine a target is alive and a method thereof are provided. By deciding plurality of specified actions containing facial movements and hand movements, recognizing whether a target in an image completes specified actions in sequence for generating a result, and determining whether the target is alive based on the result, the system and the method can making sure online interviewer is alive, and can achieve the effect of increasing security of online meetings.

Description

System and method for identifying motion in an image to determine whether a target is alive

A living body recognition system and method, in particular, a system and method for recognizing movements in an image to determine whether the target is a living body.

With the advancement of the Internet and related technologies, many services that previously required identification to be handled in person can now be completed online. However, some services, such as account opening and insurance, are still restricted by domestic regulations because the identity confirmation method through the Internet is not strict enough, and the identification target cannot complete account opening and insurance online.

As for account opening, although it is currently possible to open a digital account online through online banking or mobile banking, the relevant regulations stipulate that the level of the digital account opened is different depending on the method of identity confirmation. That is, digital accounts with lower levels have many restrictions compared to digital accounts opened after more complete identity confirmation, such as some transactions cannot be executed or the amount that can be used during transactions is lower. If you want to open a digital account with more complete functions, it is still the same as opening a general account. The bank's service staff needs to confirm the identity of the account opener at the counter, which usually means confirming whether the certificate provided by the account opener is actually held by the account opener.

In addition, there are similar situations in the insurance process, that is, relevant regulations stipulate that the insured needs to meet with the salesperson face to face, and the salesperson will explain the insurance content to the insured and confirm the insured's identity based on the documents provided by the insured, and then wait for the insurance company to notify before completing the insurance.

In fact, in some cases, for example, when infectious diseases are serious, it is not suitable for account holders or policyholders to go to bank counters or meet with insurance agents. Instead, it is more suitable to open an account or purchase insurance online. For this reason, some banks/insurance companies provide online interviews for account opening/purchase. Although it can meet both the rules of distance and face-to-face interviews, interested parties can still falsify the images during the interviews with bank service personnel/insurance agents, such as by playing pre-recorded images, etc., resulting in the inability to confirm whether the account holder/insured person is a real person even through online interviews.

In summary, it can be seen that the previous technology has long had the problem of not being able to confirm whether the other party is a real person during online conversations, so it is necessary to propose improved technical means to solve this problem.

In view of the problem that the prior art cannot determine whether the identification target is the ID holder from the photo on the ID, the present invention discloses a system and method for identifying the motion in the image to determine whether the target is alive, wherein: the system for identifying the motion in the image to determine whether the target is alive disclosed by the present invention at least includes: a motion selection module for determining a predetermined number of The system comprises an image capture module for capturing multiple target images including the face of a recognition target; a liveness recognition module for judging from the target images whether the recognition target has completed the specified actions in sequence and generated a corresponding liveness recognition result; and a result judgment module for judging whether the recognition target has passed the liveness recognition based on the liveness recognition result.

The method disclosed in the present invention for identifying the actions in the image to determine whether the target is alive includes at least the following steps: determining a predetermined number of designated actions; capturing multiple target images including the face of a recognition target; judging from the target images whether the recognition target has completed the designated actions in sequence and generated a corresponding live recognition result; judging whether the recognition target has passed the live recognition according to the live recognition result.

The system and method disclosed in the present invention are as described above. The difference between the present invention and the prior art is that the present invention solves the problems existing in the prior art by determining multiple designated actions including facial actions and hand actions and identifying whether the target in the target image has completed the designated actions in sequence, and judging whether the target is alive based on the liveness recognition result, and can achieve the technical effect of increasing the security of online consultations.

100: Client

101: Processing module

110: Image capture module

120: Action selection module

130: Liveness recognition module

140: Display module

150: Data acquisition module

160: Feature acquisition module

180: Result judgment module

190: Customer communication module

200: Server

202: Trusted host

210: Servo communication module

220: Image matching module

230: Image detection module

240: Result generation module

410: Display screen

411: Indication area

412: Prompt area

Step 310: The client obtains identification information

Step 320: The client determines the designated action for the reserved quantity

Step 321: The client obtains human characteristics based on the identification data

Step 323: The client determines the designated action that matches the human body characteristics

Step 325: The client selects a predetermined number of facial and hand movements

Step 327: The client combines facial movements and hand movements to generate designated movements

Step 330: The client captures the target image including the face of the identified target

Step 331: The client determines whether the face of the recognition target is detected

Step 333: The client determines whether the face reaches a certain proportion

Step 335: The client captures the target image

Step 337: The client determines that the identification target has not passed the liveness identification.

Step 340: The client sends the target image to the server

Step 350: The client determines from the target image whether the target has completed the specified actions in sequence.

Step 351: The client performs facial recognition and motion detection on the target image

Step 355: The client determines whether the face and hand gestures of the target to be identified in the target image are consistent with the facial movements and hand movements respectively based on the face recognition results and the motion detection results to generate the living body recognition results.

Step 360: The client generates the entity identification result

Step 371: The server obtains a trusted image containing the face of the target to be identified based on the identification data

Step 373: The server compares the face in the trusted image with the target image and generates an image comparison result

Step 375: The server determines the image depth and/or image information of the target image and generates an image determination result

Step 379: The server generates and transmits image detection results to the client

Step 391: The client determines whether the identification target has passed the liveness identification based on the liveness identification result.

Step 395: The client determines whether the identification target has passed the liveness identification based on the liveness identification results and image detection results.

Figure 1 is a diagram of the system architecture of the present invention for identifying the motion in an image to determine whether the target is a living body.

Figure 2 is a schematic diagram of the components of the server mentioned in the present invention.

Figure 3A is a flow chart of the method proposed by the present invention for identifying the movements in an image to determine whether the target is a living body.

Figure 3B is a flow chart of the method for generating a specified action according to an embodiment of the present invention.

Figure 3C is a flow chart of the method for generating identification results proposed in the embodiment of the present invention.

Figure 3D is a flow chart of the method for capturing the target image proposed in the embodiment of the present invention.

Figure 3E is a flow chart of the additional method for determining a specified action proposed in the present invention.

Figure 3F is a flow chart of the method proposed by the present invention for determining whether a target is a living body based on the target image through a server.

Figure 4 is a schematic diagram of the user interface for capturing the target image according to the embodiment of the present invention.

The following will be used in conjunction with diagrams and embodiments to explain in detail the features and implementation methods of the present invention. The content is sufficient to enable anyone familiar with the relevant technology to easily and fully understand the technical means used by the present invention to solve technical problems and implement them accordingly, thereby achieving the effects that can be achieved by the present invention.

The present invention can request the identification target to perform a specified action to determine whether the identification target is alive.

The specified actions mentioned in the present invention may include facial actions and hand actions, wherein facial actions include but are not limited to turning in a specific direction (such as left and right), raising the head, nodding, blinking, smiling, etc.; hand actions include extended fingers and the positions of the fingers, etc., but the present invention is not limited to this. Generally speaking, the position of the fingers is usually on the face or around the face, but the present invention is not limited to this.

The following first uses the system architecture diagram of "Figure 1" of the present invention to illustrate the system operation of the present invention, which is a system for identifying the action in the image to determine whether the target is a living body. As shown in "Figure 1", the system of the present invention is applied to the client 100, which includes a processing module 101, an image capture module 110, and an attachable display module 140 and a client communication module 190. Among them, the client 100 can be a computing device.

The computing device mentioned in the present invention includes but is not limited to one or more processing modules, one or more memory modules, and hardware components such as a bus connecting different hardware components (including memory modules and processing modules). Through the multiple hardware components included, the computing device can load and execute the operating system, so that the operating system runs on the computing device, and can also execute software or programs. In addition, the computing device also includes a housing, and the above-mentioned hardware components are arranged in the housing.

The bus of the computing device mentioned in the present invention may include one or more types, such as data bus, address bus, control bus, expansion bus, and/or local bus. The bus of the computing device includes but is not limited to Industry Standard Architecture (ISA) bus, Peripheral Component Interconnect (PCI) bus, Video Electronics Standards Association (VESA) local bus, and serial Universal Serial Bus (USB), Peripheral Component Interconnect Express (PCI Express, PCI-E/PCIe) bus, etc.

The processing module of the computing device of the present invention is coupled to a bus. The processing module includes a register group or a register space, which can be completely set on the processing chip of the processing module, or completely or partially set outside the processing chip and coupled to the processing chip via a dedicated electrical connection and/or via a bus. The processing module can be a central processing unit, a microprocessor or any suitable processing element. If the computing device is a multi-processor device, that is, the computing device includes multiple processing modules, the processing modules included in the computing device are the same or similar, and are coupled and communicated through a bus. The processing module can interpret a computer instruction or a series of multiple computer instructions to perform specific calculations or operations, such as mathematical operations, logical operations, data comparison, copying/moving data, etc., to drive other hardware components in the computing device or run the operating system or execute various programs and/or modules.

Computing devices usually also include one or more chipsets. The processing module of the computing device can be coupled to the chipset or electrically connected to the chipset through a bus. The chipset is composed of one or more integrated circuits (ICs), including a memory controller and a peripheral input/output (I/O) controller, etc. That is, the memory controller and the peripheral input/output (I/O) controller can be included in one IC, or can be implemented using two or more ICs. The chipset usually provides input/output and memory management functions, as well as multiple general and/or dedicated registers, timers, etc., wherein the above-mentioned general and/or dedicated registers and timers can be accessed or used by one or more processing modules coupled or electrically connected to the chipset.

The processing module of the computing device can also access the data in the memory module and the mass storage area installed on the computing device through the memory controller. The above-mentioned memory module includes any type of volatile memory and/or non-volatile memory (NVRAM) memory, such as static random access memory (SRAM), dynamic random access memory (DRAM), read-only memory (ROM), flash memory, etc. The above-mentioned mass storage area can include any type of storage device or storage media, such as a hard drive, optical disc, flash drive, memory card, solid state disk (SSD), or any other storage device. In other words, the memory controller can access data in static random access memory, dynamic random access memory, flash memory, hard drive, and solid state disk.

The processing module of the computing device can also be connected and communicated with peripheral devices or interfaces such as peripheral output devices, peripheral input devices, communication interfaces, and GPS receivers through the peripheral input/output controller via the peripheral input/output bus. The peripheral input device can be any type of input device, such as a keyboard, a mouse, a trackball, a touchpad, a joystick, etc. The peripheral output device can be any type of output device, such as a display, a printer, etc. The peripheral input device and the peripheral output device can also be the same device, such as a touch screen, etc. The communication interface may include a wireless communication interface and/or a wired communication interface. The wireless communication interface may include an interface supporting a wireless local area network (such as Wi-Fi, Zigbee, etc.), Bluetooth, infrared, near-field communication (NFC), 3G/4G/5G and other mobile communication networks (cellular networks) or other wireless data transmission protocols. The wired communication interface may be an Ethernet device, a DSL modem, a cable modem, an asynchronous transfer mode (ATM) device, or an optical fiber communication interface and/or component. The processing module can periodically poll various peripheral devices and interfaces, so that the computing device can input and output data through various peripheral devices and interfaces, and can also communicate with another computing device with the hardware components described above.

In the present invention, the processing module 101 in the client 100 can load and execute one or a group of computer instructions stored in the memory module (not shown) through the bus, and can generate the action selection module 120, the living body recognition module 130, and the result judgment module 180 after executing the computer instructions. In some embodiments, the processing module 101 can also generate the data acquisition module 150 and the feature acquisition module 160 after executing the computer instructions.

The image capture module 110 may include components such as a lens and a photosensitive element (not shown in the figure). The lens allows light to pass through and illuminate the photosensitive element, and the photosensitive element can convert the light irradiated on the photosensitive element into an image signal.

The image capture module 110 is responsible for capturing the target image. The target image captured by the image capture module 110 includes a face and a gesture. When the image capture module 110 is driven by the liveness recognition module 130 or receives a cooperative message sent by the liveness recognition module 130, it can continuously use the image signal output by the photosensitive element as the target image.

The image capture module 110 can also capture data images. Generally speaking, the data images captured by the image capture module 110 include a complete string of characters or a barcode sufficient to identify the content represented, but the present invention is not limited thereto. The image capture module 110 can continue to use the image signal output by the photosensitive element as a data image when it is driven by the data acquisition module 150 or receives a coordinated message from the data acquisition module 150.

The image capture module 110 can also determine whether the face of the recognition target is detected within the image capture range, and can capture the target image when the detected face reaches a certain proportion of the image capture range.

The action selection module 120 is responsible for determining a predetermined number of designated actions. The number of designated actions determined by the action selection module 120 can be a predetermined value or set by the data acquisition module 150, and the present invention has no special limitations.

The action selection module 120 can select a facial action and a hand action from a pre-established action database (not shown in the figure), and can combine the selected facial action and hand action to generate a designated action. The action selection module 120 can repeat the above process of generating designated actions until a predetermined number of designated actions are generated.

The liveness recognition module 130 is responsible for performing liveness recognition according to the designated action generated by the action selection module 120. The process of liveness recognition performed by the liveness recognition module 130 is, for example: prompting the recognition target to complete the first designated action generated in front of the image capture module 110 through the display module 140, while continuing to perform face recognition and/or motion detection on the indication image captured by the image capture module 110 and including the completion of the first designated action to obtain the facial features and gesture features in the indication image, and judging whether the recognition target in the indication image has completed the designated action based on the facial features and gesture features in the indication image. If so, prompting the recognition target to complete the second designated action generated through the display module 140 and repeating the above comparison process until the recognition target correctly completes all designated actions, thereby performing liveness recognition. However, the method of liveness recognition proposed in the present invention is not limited to the above. Among them, the liveness recognition module 130 can usually compare the facial features/hand gesture features in the indication image with the facial features/hand gesture features corresponding to the designated action read from the action database to determine whether the recognition target has completed the designated action, but the present invention is not limited to this.

The liveness recognition module 130 is also responsible for generating corresponding liveness recognition results after liveness recognition is performed. The activity recognition module 130 can judge that the liveness recognition is passed and generate a liveness recognition result indicating that the liveness recognition is passed when all the specified actions generated are completed correctly or the number of consecutively completed specified actions reaches a parameter value; the activity recognition module 130 can also judge that the liveness recognition is not passed and generate a liveness recognition result indicating that the liveness recognition is not passed when any of the specified actions generated are not completed correctly or the number of times the specified actions are not completed correctly reaches a predetermined value.

The liveness recognition module 130 may also accumulate the number of failed attempts and re-perform liveness recognition when determining that the liveness recognition has not been passed. The liveness recognition module 130 may generate a liveness recognition result indicating that the liveness recognition has not been passed when the accumulated number of failed attempts reaches a predetermined value.

The display module 140 can display the data image, target image, and indication image captured by the image capture module 110 on the display screen of the client 100. The display module 140 can also display the indication box and text and other prompt messages generated by the data acquisition module 150 on the display screen of the client 100. The display module 140 can also display the judgment result generated by the result judgment module 180 on the display screen of the client 100.

The data acquisition module 150 can obtain identification data. More specifically, the data acquisition module 150 can obtain the mobile phone number, digital certificate, and bank card information of the identification target as identification data.

The data acquisition module 150 can also set the number of designated actions generated by the liveness recognition module 130 for liveness recognition, and can also set the number of times the liveness recognition module 130 needs to correctly complete the designated actions in succession to determine that the liveness recognition has passed (i.e., the parameter value used by the liveness recognition module 130 for determination).

The feature acquisition module 160 can acquire human features based on the identity data acquired by the data acquisition module 150. The human features acquired by the feature acquisition module 160 can indicate whether the face or hands of the recognition target are defective.

The result judgment module 180 is responsible for judging whether the identification target has passed the liveness recognition based on the liveness recognition result generated by the liveness recognition module 130. More specifically, when the liveness recognition result indicates that the liveness recognition has been passed, the result judgment module 180 can generate a judgment result indicating that the identification target has passed the liveness recognition; conversely, when the liveness recognition result indicates that the liveness recognition has not been passed, the result judgment module 180 can generate a judgment result indicating that the identification target has not passed the liveness recognition.

In addition to judging whether the identification target has passed the liveness identification based on the liveness identification result generated by the liveness identification module 130, the result judgment module 180 may also, in some embodiments, combine and refer to the image detection result received by the client communication module 190 to judge whether the identification target has passed the liveness identification. More specifically, the result judgment module 180 can generate a judgment result indicating that the target has passed the liveness recognition when the liveness recognition result indicates that the liveness recognition has been passed and the image detection result indicates that the target image has passed the detection; conversely, the result judgment module 180 can generate a judgment result indicating that the target has not passed the liveness recognition when the liveness recognition result indicates that the liveness recognition has not been passed or the image detection result indicates that the target image has not passed the detection.

The client communication module 190 can transmit the target image captured by the image capture module 110 to the server 200, and the client communication module 190 can also receive the image detection result transmitted by the server 200. In some embodiments, the client communication module 190 can also transmit the identity identification data obtained by the data acquisition module 150 and/or the liveness recognition result generated by the liveness recognition module 130 to the server 200.

The server 200 can further perform liveness detection on the target image captured by the client 100, and in some embodiments, can also store the liveness recognition result of the client 100. The server 200 can include a server communication module 210, an image matching module 220, an image detection module 230, and a result generation module 240 as shown in the component schematic diagram of "Figure 2". Generally speaking, the server 200 is also a computing device.

The server communication module 210 can receive the target image transmitted by the client 100, and can also transmit the image detection result generated by the result generation module 240 back to the client 100. In some embodiments, the server communication module 210 can also receive the identity identification data and/or liveness recognition result transmitted by the client 100.

The server communication module 210 can also download a trusted image corresponding to the received identity data from the trusted host 202. The trusted image downloaded by the server communication module 210 includes the face of the identification target, for example, the trusted image can include the ID photo of the identification target, but the present invention is not limited to this. The trusted host 202 is a server provided by a trusted unit or organization such as a government agency, and the present invention is not limited to this.

The image matching module 220 can obtain a trusted image of the identification target according to the identification data received by the servo communication module 210. The image matching module 220 can download the trusted image from the trusted host 202 through the servo communication module 210 or can read the pre-stored trusted image, wherein the trusted image can be uploaded by the identification target itself or previously downloaded by the trusted host 202.

The image matching module 220 can also compare the target image received by the servo communication module 210 with the face in the obtained credible image to generate an image matching result. For example, the image matching module 220 can compare the similarity between the target image and the face in the credible image, and can generate an image matching result indicating that the image matching has passed when the generated similarity is determined to be greater than or equal to a threshold value, or can generate an image matching result indicating that the image matching has not passed when the generated similarity is determined to be less than the threshold value.

The image detection module 230 can determine the image depth of the target image received by the servo communication module 210. The image depth determined by the image detection module 230 is the distance between the image capture module 110 and each object or region in the target image. This distance is usually an approximate value, but the present invention is not limited to this. For example, the image detection module 230 can use a convolutional neural network (CNN) such as AlexNet, GoogleNet, VGG, Residual net, etc. to determine the distance of each object or region in the target image, but the present invention is not limited to this.

The image detection module 230 can also determine the image information of the target image received by the servo communication module 210. The image information determined by the image detection module 230 can include information indicating whether the target image contains moiré, reflection, or facial distortion, but the present invention is not limited thereto.

The image detection module 230 may also generate an image determination result according to the determined image depth and/or image information. For example, the image detection module 230 can generate an image judgment result that fails depth judgment based on whether the image depths of the face and the area behind the face in the target image are similar. When the two are similar, the image detection module 230 can generate an image judgment result that fails optical judgment when there is information in the image information of the target image indicating that the face range in the target image contains moiré, reflection, or facial distortion. The image detection module 230 can also generate an image judgment result that passes image judgment when the difference in image depths of the face and the area behind the face in the target image reaches a certain value and there is no information in the image information of the target image indicating that the face range in the target image contains moiré, reflection, or facial distortion.

The result generation module 240 generates an image detection result including the image comparison result generated by the image comparison module 220 and the image judgment result generated by the image detection module 230.

Next, an example is used to explain the operating system and method of the present invention. Please refer to "Figure 3A" for the flowchart of the method of identifying the action in the image to determine whether the target is a living body.

When identifying the target using the present invention, the action selection module 120 of the client 100 may first determine a predetermined number of designated actions (step 320). In this embodiment, assuming that the predetermined number is three, the action selection module 120 may first select three (predetermined number) facial actions and three hand actions as shown in the process of "Figure 3B" (step 325), and combine one facial action and one hand action three times respectively to generate three designated actions (step 327), such as turning the head to the right, blinking, and placing the right finger under the left eye.

After the action selection module 120 of the client 100 determines the designated action (step 320), the image capture module 110 of the client 100 can capture the target image including the face of the recognition target (step 330). In this embodiment, assuming that as shown in the process of "Figure 3D", the image capture module 110 can generate a prompt message to the display module 140, the display module 140 of the client 100 can prompt the recognition target to align the lens of the mobile phone (client 100) with the front of the face and adjust the position of the mobile phone in the prompt area 412 of the display screen 410 as shown in "Figure 4", so that the recognition target can press the button when the front of the face is placed in the indication area 411 of the display screen 410 and the facial proportions are similar to the indication area 411. The shutter is pressed to trigger the image capture module 110 to capture a target image including the front face of the identified target face (step 333). For another example, the image capture module 110 can continuously determine whether the captured target image includes a face with a sufficient proportion and facing the front (step 333). When the target image includes a face facing the front and the proportion of the face is similar to that of the indication area 411 (e.g., the difference in height or width between the face and the indication area 411 is within 10%), the image capture module 110 can capture the target image (step 335). In the process of capturing the target image three times, the image capture module 110 can determine whether the face of the target continues to be within the image capture range (step 331). If not, a liveness recognition result indicating that the target has not passed liveness recognition can be generated (step 337).

Returning to "Figure 3A", after the image capture module 110 of the client 100 captures the target image (step 330), the liveness recognition module 130 of the client 100 can determine whether the recognition target has completed the generated designated actions in sequence (step 350) to perform liveness recognition, and can generate a corresponding liveness recognition result after liveness recognition (step 360). In this embodiment, it is assumed that the liveness recognition module 130 can prompt the recognition target in the prompt area 312 through the display module 140 to align the client 100 with the face so that the face is placed in the indication area 311 of the display screen 310 and the proportion of the face is close to the indication area 311, so that the recognition target completes three designated actions in sequence according to the prompt. In this way, the liveness recognition module 130 can recognize the target image as shown in the process of "Figure 3C". Perform face recognition and motion detection (step 351), and judge whether the face and hand gesture of the recognition target match the face and hand gesture of the designated action according to the face recognition result and the motion detection result (step 355), and then generate a liveness recognition result of passing liveness recognition when the recognition target successfully completes the designated action three times in a row, otherwise, the liveness recognition module 130 can generate a liveness recognition result of failing liveness recognition.

In more detail, the liveness recognition module 130 may first prompt the recognition target to complete the first designated action. When the recognition target turns its head to the right, causing the image capture module 110 to capture an indication image including the recognition target turning its head to the right, the liveness recognition module 130 may perform facial feature analysis on the indication image to determine the rotation direction of the face in the indication image. When the rotation direction is different from the designated direction, the liveness recognition module 130 can determine that the liveness recognition has not been passed; and when the rotation direction is the same as the designated direction, the liveness recognition module 130 can prompt the recognition target to complete the second designated action again. When the recognition target blinks according to the prompt, so that the image capture module 110 captures the indication image containing the recognition target blinking, the liveness recognition module 130 can perform facial feature analysis on the indication image to determine whether the face in the indication image blinks. If not, the liveness recognition module 130 can determine that the liveness recognition has not passed; if so, the liveness recognition module 130 can prompt the recognition target again to complete the third designated action. When the recognition target places the right finger under the left eye according to the prompt, so that the image capture module 110 captures the indication image containing the recognition target, the liveness recognition module 130 can perform face recognition and action detection on the indication image to determine the position of the face in the indication image and whether the right finger is correctly placed under the left eye. If not, the liveness recognition module 130 can determine that the liveness recognition has not passed; if so, the liveness recognition module 130 can determine that the liveness recognition has passed and can generate a liveness recognition result indicating that the liveness recognition has passed.

In the above-mentioned liveness recognition process, if the liveness recognition module 130 determines that the liveness recognition has not passed, the liveness recognition module 130 can accumulate the number of failures, and can provide the recognition target again for liveness recognition before the number of failures has not reached a predetermined value, that is, regenerate three designated actions, and sequentially recognize the target to complete three designated actions according to the prompts, until the accumulated number of failures reaches a predetermined value, the liveness recognition module 130 can generate a liveness recognition result indicating that the liveness recognition has not passed.

After the liveness recognition module 130 of the client 100 generates the liveness recognition result, the result judgment module 180 of the client 100 can judge whether the recognition target has passed the liveness recognition according to the liveness recognition result generated by the liveness recognition module 130 (step 391). In this embodiment, assuming that the result judgment module 180 can judge whether the liveness recognition result indicates that the recognition target has passed the liveness recognition, the result judgment module 180 can judge that the recognition target has passed the liveness recognition.

Thus, through the present invention, the client 100 can determine whether the identification target in the image is a living body based on whether the identification target performs a specified action.

In the above-mentioned embodiment, the result judgment module 180 of the client 100 can also judge whether the accumulated number of failures is higher than a preset value. If so, it means that the identification target has not passed the liveness recognition. If not, the result judgment module 180 generates a prompt message of comparison failure and the display module 140 of the client 100 displays the prompt message to prompt the identification target to align the lens of the client 100 with the front face of the identification target again so that the image capture module 140 can be used to detect the liveness of the identification target. The group 110 can capture the target image of the identification target again (step 320), and the liveness identification module 130 can also determine whether the identification target has completed the generated designated actions in sequence to perform liveness identification, and can generate a corresponding liveness identification result after liveness identification (step 350). The result determination module 180 can determine whether the identification target has passed the liveness identification according to the liveness identification result generated by the liveness identification module 130 (step 360).

In addition, the above embodiment can also be shown in the process of "Figure 3E". When the action selection module 120 of the client 100 determines the designated action (step 320), the data acquisition module 150 of the client 100 can first obtain the identification data of the identification target (step 310), and the feature acquisition module 160 of the client 100 can obtain the corresponding human body features according to the identification data obtained by the data acquisition module 150 (step 321). The action selection module 120 of the client 100 can determine the designated action that meets the human body features retrieved by the feature acquisition module 160 (step 323).

In addition, in the above embodiment, as shown in the process of "Figure 3F", after the image capture module 110 of the client 100 captures the target image containing the recognition target (step 330), the client communication module 190 of the client 100 can transmit the target image to the server 200 (step 340). In this embodiment, it is assumed that the client communication module 190 can transmit the target image together with the identity recognition data obtained by the data acquisition module 150 of the client 100 to the server 200 after the living body recognition module 130 generates the living body recognition result (step 360).

After the server communication module 210 of the server 200 receives the target image and identification data transmitted by the client 100, the image matching module 220 of the server 200 can obtain a trusted image containing the identification target according to the identification data received by the server communication module 210 (step 371). In this embodiment, it is assumed that the image matching module 220 can transmit the identification data to the trusted host 202 through the server communication module 210, and receive the certificate image returned by the trusted host 202 through the server communication module 210.

After the image comparison module 220 of the server 200 obtains the credible image, the face in the obtained credible image can be compared with the face in the target image received by the server communication module 210, and a corresponding image comparison result can be generated after the comparison (step 373). In this embodiment, it is assumed that the image comparison module 220 can extract the facial image from the credible image according to the certificate type of the credible image and calculate the image features of the facial image, and can calculate the image features of the target image, and then determine the similarity between the target image and the credible image according to the calculated image features of the facial image and the image features of the target image, and select to generate an image comparison result indicating whether the comparison is passed or not according to whether the similarity exceeds a predetermined threshold value.

After the image comparison module 220 of the server 200 generates the image comparison result, the image detection module 230 of the server 200 can determine whether the image comparison result indicates that the image comparison has been passed. If not, the image detection module 230 can terminate the execution, and the result generation module 240 of the server 200 can generate an image detection result including the image comparison result. If the image detection module 230 determines that the image comparison result indicates that the image comparison has been passed, the image detection module 230 can further determine the image depth and/or image information of the target image received by the servo communication module 210 of the server 200, and can generate an image determination result according to the determined image depth and/or image information of the target image (step 375). In this embodiment, it is assumed that the image detection module 230 can use a convolutional neural network to determine the depth of the face of the target to be identified and the image behind it in the target image, and can also determine whether there are abnormal areas such as moiré, reflection, and/or deformity or distortion on the face of the target to be identified in the target image, and can detect the difference in depth between the face of the target to be identified and the image behind it exceeds a certain value and the face of the target to be identified is not included in the image. When there is no moire, reflection, or abnormal area on the face, the image detection module 230 can generate an image judgment result indicating that the judgment has been passed. On the contrary, when the difference in depth between the face of the recognition target and the image behind it is lower than a certain value or there is moire, reflection, or abnormal area on the face of the recognition target, the image detection module 230 can generate an image judgment result indicating that the judgment has been passed.

After the image detection module 230 of the server 200 generates an image judgment result, the result generation module 240 of the server 200 can generate an image detection result based on the image judgment result, and can transmit the generated image detection result to the client 100 (step 379). In this embodiment, it is assumed that the result generation module 240 can generate an image detection result indicating that the detection has passed when the image judgment result indicates that the judgment has passed, and can generate an image detection result indicating that the detection has not passed when the image judgment result indicates that the judgment has not passed.

After the client communication module 190 of the client 100 receives the image detection result transmitted by the server 200, the result judgment module 180 of the client 100 can judge whether the identification target passes the liveness recognition according to the image detection result received by the client communication module 190 and the liveness recognition result generated by the liveness recognition module 130 of the client 100 (step 395). In this embodiment, it is assumed that the result judgment module 180 can judge that the identification target passes the liveness recognition when the liveness recognition result indicates that the identification target passes the liveness recognition and the image detection result indicates that the target image passes the image detection, otherwise it can be judged that the identification target does not pass the liveness recognition.

In summary, the difference between the present invention and the prior art is that the present invention has the technical means of determining multiple designated actions including facial actions and hand actions and identifying whether the target in the target image has completed the designated actions in sequence, and judging whether the target is alive based on the liveness recognition result. This technical means can solve the problem of the prior art that it is impossible to confirm whether the other party is a real person during an online meeting, thereby achieving the technical effect of increasing the security of online meetings.

Furthermore, the method of the present invention for identifying the motion in the image to determine whether the target is alive can be implemented in hardware, software or a combination of hardware and software, and can also be implemented in a centralized manner in a computer system or in a distributed manner with different components distributed in several interconnected computer systems.

Although the implementation methods disclosed in the present invention are as above, the contents described are not used to directly limit the scope of patent protection of the present invention. Any person with common knowledge in the technical field to which the present invention belongs, without departing from the spirit and scope disclosed by the present invention, makes slight changes and modifications to the implementation of the present invention in form and details, which are all within the scope of patent protection of the present invention. The scope of patent protection of the present invention shall still be based on the scope defined in the attached patent application.

Step 320: The client determines the designated action for the reserved quantity

Step 330: The client captures the target image including the face of the identified target

Step 350: The client determines from the target image whether the target has completed the specified actions in sequence.

Step 360: The client generates the entity identification result

Step 391: The client determines whether the identification target has passed the liveness identification based on the liveness identification result.

Claims (9)

A system for identifying movements in an image to determine whether a target is alive or not, the system at least comprising: a data acquisition module for acquiring identity recognition data; a feature acquisition module for acquiring a human feature based on the identity recognition data; a movement selection module for determining a predetermined number of facial movements and hand movements, and combining the facial movements and the hand movements one by one to generate the predetermined number of designated movements, wherein the facial movements and the hand movements conform to the human feature; an image capture module; and a live body recognition module for performing a live body recognition operation and generating a corresponding live body recognition result. In the method, the liveness recognition operation includes capturing multiple target images including the face and hand gesture of a recognition target through the image capture module, and sequentially performing face recognition and motion detection on each target image, and obtaining a designated action in a generated order, and judging whether the face and hand gesture of the recognition target in each target image are respectively consistent with one of the facial actions and one of the hand actions in the obtained designated actions according to the face recognition results and the motion detection results to judge whether the recognition target has completed each of the designated actions in sequence; and a result judgment module is used to judge whether the recognition target has passed the liveness recognition according to the liveness recognition results. A system for identifying motion in an image to determine whether a target is alive as described in claim 1, wherein the system further comprises a client communication module for transmitting the target images to a server, the server further comprises a servo communication module and an image detection module, the servo communication module is used to receive the target images, the image detection module is used to determine an image depth and/or an image information of the target image and generate an image determination result based on the image depth and/or the image information, and the servo communication module transmits an image detection result including or representing the image determination result to the client communication module, the result determination module is further used to determine whether the identified target passes the liveness identification based on the liveness identification result and the image detection result received by the client communication module. A system for identifying motion in an image to determine whether a target is a living body as described in claim 1, wherein the system further comprises a data acquisition module and a client communication module, the data acquisition module is used to obtain identity identification data, the client communication module is used to transmit the target images and the identity identification data to a server, the server further comprises a server communication module and an image matching module, the server communication module is used to receive the target images and the identity identification data and to match the target images with the identity identification data according to the identity identification data. The trusted server downloads a trusted image containing the face of the recognition target, and the image comparison module compares the trusted image with the face in the target image to generate an image comparison result. The server communication module transmits the image detection result containing or representing the image comparison result to the client communication module. The result judgment module is further used to judge whether the recognition target has passed the liveness recognition based on the liveness recognition result and the image detection result received by the client communication module. A system for determining whether a target is alive by identifying the movement in an image as described in claim 1, wherein the liveness recognition module is further used to determine whether the face of the recognition target is detected during the liveness recognition operation, and the result determination module is further used to determine that the recognition target has not passed the liveness recognition when the liveness recognition module determines that the face of the recognition target is not detected during the liveness recognition operation. A system for identifying actions in an image to determine whether a target is alive as described in claim 1, wherein the system further comprises a data acquisition module and a feature acquisition module, wherein the data acquisition module is used to obtain an identity recognition data, the feature acquisition module is used to obtain a human body feature based on the identity recognition data, and the action selection module is used to determine the specified actions that meet the human body feature. A method for identifying movements in an image to determine whether a target is alive or not, the method comprising at least the following steps: a client obtains identity recognition data, and obtains a human body feature based on the identity recognition data; the client determines a predetermined number of facial movements and hand movements, and combines the facial movements and the hand movements one by one to generate the predetermined number of designated movements, wherein the facial movements and the hand movements conform to the human body feature; the client performs a live body recognition operation and generates a corresponding live body recognition result, wherein the live body recognition operation Including: capturing multiple target images including the face and hand gesture of a recognition target; and performing face recognition and motion detection on each target image in sequence, and obtaining a designated action in the order of generation, and judging whether the face and hand gesture of the recognition target in each target image respectively match one of the facial actions and one of the hand actions in the obtained designated actions according to the face recognition results and the motion detection results to judge whether the recognition target has completed each designated action in sequence; and the client judges whether the recognition target has passed the liveness recognition according to the liveness recognition result. A method for identifying the movement in an image to determine whether the target is alive as described in claim 6, wherein the method further comprises the step of the client transmitting the target images to a server after the client captures a plurality of target images including the face of the target to be identified, the server determining an image depth and/or an image information of the target image and generating an image detection result based on the image depth and/or the image information and then transmitting the image detection result to the client, and the client determining whether the target to be identified passes the liveness identification based on the liveness identification result and the image detection result. A method for identifying the movement in an image to determine whether the target is alive as described in claim 6, wherein the method further comprises the step of the client transmitting the target image and an identification data to a server after the client captures a plurality of target images including the face of the identification target, the server obtaining a credible image including the face of the identification target based on the identification data and comparing the credible image with the face in the target image to generate an image detection result and then transmitting the image detection result to the client, and the client determining whether the identification target passes the liveness recognition based on the liveness recognition result and the image detection result. A method for identifying the movement in an image to determine whether the target is a living body as described in claim 6, wherein the client performs the live body recognition operation and generates the corresponding live body recognition result, and further includes the step of the client determining whether the face of the recognition target is not detected within the image capture range during the execution of the live body recognition operation, and generating the live body recognition result indicating that the recognition target has not passed the live body recognition when the face of the recognition target is not detected.

Images