WO2021005702A1 - 顔検出処理装置および顔検出処理方法 - Google Patents
顔検出処理装置および顔検出処理方法 Download PDFInfo
- Publication number
- WO2021005702A1 WO2021005702A1 PCT/JP2019/027103 JP2019027103W WO2021005702A1 WO 2021005702 A1 WO2021005702 A1 WO 2021005702A1 JP 2019027103 W JP2019027103 W JP 2019027103W WO 2021005702 A1 WO2021005702 A1 WO 2021005702A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- face
- image
- person
- face detection
- detected
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/59—Context or environment of the image inside of a vehicle, e.g. relating to seat occupancy, driver state or inner lighting conditions
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/161—Detection; Localisation; Normalisation
Definitions
- the present invention relates to a face detection processing device and a face detection processing method.
- the driver monitoring system which has been actively developed in recent years, is a technology necessary for determining the transfer of driving initiative from a vehicle to a human in a partially autonomous vehicle.
- the process of detecting the featured part of the passenger's face or facial parts is an elemental technology in the driver monitoring system.
- the feature site detection process is required not only to determine the accuracy but also to reduce the amount of calculation and perform the process at high speed.
- the vehicle control device described in Patent Document 1 discloses a face detection processing technique that reduces the amount of calculation and increases the processing speed.
- the position of the face may shift from the center of the screen due to the movement of the driver.
- the position of the face is usually the position of the headrest.
- the vehicle control device of Patent Document 1 searches for a face from a portion where the driver's face is likely to exist by grasping the position of the headrest on the shooting screen in advance. As a result, the search time is shortened.
- the face detection device may detect the face of the person to be non-detected. In that case, the face of the person to be detected is excluded from the subsequent search area, so that the detection accuracy is lowered.
- the present invention has been made to solve the above problems, and an object of the present invention is to provide a face detection processing device that enables a face detection device to accurately detect the face of a person to be detected. ..
- the face detection processing device outputs an image to the face detection device.
- the face detection device detects the face of the person in the search area set in at least a part of the sequentially input images, and is input next based on the success or failure of the detection of the face of the person. Change the search area in the image.
- the face detection processing device includes a validation unit and an image insertion unit.
- the face of the person detected by the face detection device is the face of the person to be detected in advance based on the predetermined conditions regarding the face of the person detected by the face detection device. Judge whether or not.
- the image insertion unit uses a dummy image as an image for the face detection device to cause the face detection device to fail to detect the face of a person in the search area and change the search area. Output.
- a face detection processing device that enables a face detection device to accurately detect the face of a person to be detected.
- FIG. It is a block diagram which shows the structure of the face detection processing apparatus in Embodiment 1.
- FIG. It is a figure which shows an example of the structure of the processing circuit which a face detection processing apparatus has. It is a figure which shows another example of the structure of the processing circuit which a face detection processing apparatus has.
- FIG. It is a flowchart which shows the face detection processing method in Embodiment 2.
- FIG. It is a figure which shows an example of the image image
- FIG. It is a flowchart which shows the face detection processing method in Embodiment 3. It is a figure which shows an example of the simulated image in Embodiment 3.
- FIG. It is a figure which shows an example of the image image
- FIG. 1 is a block diagram showing the configuration of the face detection processing device 100 according to the first embodiment.
- the face detection processing device 100 outputs an image for the face detection device 110 to detect the face to the face detection device 110 that detects the face of the person in the image.
- the face detection device 110 has a function of detecting the face of a person in a search area set in at least a part of the sequentially input images. Further, the face detection device 110 has a function of changing the search area in the next input image based on the success or failure of the face detection.
- the face detection processing device 100 includes a validity determination unit 10 and an image insertion unit 20.
- the validity determination unit 10 determines that the face of the person detected by the face detection device 110 is the face of the person to be detected in advance based on the predetermined conditions regarding the face of the person detected by the face detection device 110. Determine if it is a face. In other words, the validity determination unit 10 determines the validity of the detected person's face.
- the predetermined conditions are, for example, a condition relating to the position of the face in the image or the position of the feature portion of the face in the image, a condition relating to the size of the face in the image, or a condition relating to the feature of the face. ..
- the predetermined condition may be, for example, a condition relating to the relationship between the traveling state of the vehicle on which the person is on board and the orientation of the person's face.
- the image insertion unit 20 outputs a dummy image to the face detection device 110 based on the determination result of the validity determination unit 10.
- the dummy image is an image for causing the face detection device 110 to fail to detect the face of the person previously detected in the search area and change the search area.
- the dummy image is an image for resetting the search area.
- the face detection device 110 When a dummy image is input to the face detection device 110, the face detection device 110 fails to detect the face of the person previously detected in the dummy image.
- the face detection device 110 changes the search area in the image input next to the dummy image. Then, the face detection device 110 detects the face of the person in the changed search area.
- the modified search area is likely to contain the face of the person to be detected. Therefore, the face detection device 110 can accurately detect the face of the person to be detected.
- FIG. 2 is a diagram showing an example of the configuration of the processing circuit 90 included in the face detection processing device 100.
- Each function of the validity determination unit 10 and the image insertion unit 20 is realized by the processing circuit 90. That is, the processing circuit 90 has a validity determination unit 10 and an image insertion unit 20.
- the processing circuit 90 may be, for example, a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an ASIC (Application Specific Integrated Circuit), or an FPGA (Field). -ProgrammableGateArray), or a circuit that combines these.
- Each function of the validity determination unit 10 and the image insertion unit 20 may be individually realized by a plurality of processing circuits, or may be collectively realized by one processing circuit.
- FIG. 3 is a diagram showing another example of the configuration of the processing circuit included in the face detection processing device 100.
- the processing circuit includes a processor 91 and a memory 92.
- each function of the validity determination unit 10 and the image insertion unit 20 is realized.
- each function is realized by executing software or firmware described as a program by the processor 91.
- the face detection processing device 100 has a memory 92 for storing the program and a processor 91 for executing the program.
- the face of the person detected by the face detection device 110 is defined in advance by the face detection processing device 100 based on the predetermined conditions regarding the face of the person detected by the face detection device 110. It is determined whether or not the face is the face of the person to be detected, and based on the determination result, the face detection device 110 fails to detect the face of the person in the search area and detects a dummy image for changing the search area. The function of outputting to the device 110 is described.
- the program causes the computer to execute the procedure or method of the validity determination unit 10 and the image insertion unit 20.
- the processor 91 is, for example, a CPU (Central Processing Unit), an arithmetic unit, a microprocessor, a microcomputer, a DSP (Digital Signal Processor), or the like.
- the memory 92 is, for example, non-volatile or volatile such as RAM (RandomAccessMemory), ROM (ReadOnlyMemory), flash memory, EPROM (ErasableProgrammableReadOnlyMemory), and EEPROM (ElectricallyErasableProgrammableReadOnlyMemory). It is a semiconductor memory.
- the memory 92 may be any storage medium used in the future, such as a magnetic disk, a flexible disk, an optical disk, a compact disk, a mini disk, or a DVD.
- Some of the functions of the validity determination unit 10 and the image insertion unit 20 described above may be realized by dedicated hardware, and some may be realized by software or firmware. In this way, the processing circuit realizes each of the above-mentioned functions by hardware, software, firmware, or a combination thereof.
- FIG. 4 is a flowchart showing the face detection processing method according to the first embodiment.
- step S1 the validity determination unit 10 determines the face of the person detected by the face detection device 110 in advance based on the predetermined conditions regarding the face of the person detected by the face detection device 110. It is determined whether or not the face is the face of the person to be detected. When the face of the detected person is a predetermined person to be detected, the face detection processing method ends. If the face of the detected person is not a predetermined face of the person to be detected, step S2 is executed.
- step S2 the image insertion unit 20 outputs a dummy image to the face detection device 110.
- the face detection device 110 When a dummy image is input to the face detection device 110, the face detection device 110 fails to detect the face of the person previously detected in the dummy image.
- the face detection device 110 changes the search area in the image input next to the dummy image. Then, the face detection device 110 detects the face of the person in the changed search area.
- the modified search area is likely to contain the face of the person to be detected. Therefore, the face detection device 110 can accurately detect the face of the person to be detected.
- the face detection processing device 100 in the first embodiment outputs an image to the face detection device 110.
- the face detection device 110 detects the face of the person in the search area set in at least a part of the sequentially input images, and is input next based on the success or failure of the detection of the face of the person. Change the search area in the image.
- the face detection processing device 100 includes a validity determination unit 10 and an image insertion unit 20.
- the validity determination unit 10 determines that the face of the person detected by the face detection device 110 is a predetermined detection target person based on the predetermined conditions regarding the face of the person detected by the face detection device 110. Determine if it is a face.
- the image insertion unit 20 outputs a dummy image to the face detection device 110 based on the determination result of the validity determination unit 10.
- the dummy image is an image for causing the face detection device 110 to fail to detect the face of a person in the search area and change the search area.
- Such a face detection processing device 100 enables the face detection device 110 to accurately detect the face of a person to be detected.
- the face detection processing method in the first embodiment outputs an image to the face detection device 110.
- the face detection device 110 detects the face of the person in the search area set in at least a part of the sequentially input images, and is input next based on the success or failure of the detection of the face of the person. Change the search area in the image.
- the face of the person detected by the face detection device 110 is the face of the person to be detected in advance based on the predetermined conditions regarding the face of the person detected by the face detection device 110. It is determined whether or not the face is, and a dummy image is output to the face detection device 110 based on the determination result.
- the dummy image is an image for causing the face detection device 110 to fail to detect the face of a person in the search area and change the search area.
- Such a face detection processing method enables the face detection device 110 to accurately detect the face of the person to be detected.
- the face detection processing apparatus and the face detection processing method according to the second embodiment will be described.
- the second embodiment is a subordinate concept of the first embodiment, and the face detection processing device according to the second embodiment includes each configuration of the face detection processing device 100 according to the first embodiment. The same configuration and operation as in the first embodiment will not be described.
- FIG. 5 is a block diagram showing the configurations of the face detection processing device 101 and the face detection processing system 200 according to the second embodiment.
- the face detection processing system 200 includes an image acquisition device 120, a face detection processing device 101, and a face detection device 110.
- the face detection processing device 101 includes a storage unit 30, an image insertion unit 20, and a validity determination unit 10.
- the face detection device 110 includes a feature portion detection unit 111.
- the image acquisition device 120 sequentially acquires images including the face of a person.
- the image acquisition device 120 is, for example, a camera mounted on a moving body and acquiring an image of a passenger of the moving body.
- the moving body is a vehicle, an airplane, a railroad, a bus, a motorcycle, or the like. Passengers include drivers.
- the camera is installed in the center of the front of the vehicle, such as the center console of the vehicle, and captures both the driver and the passenger in the passenger seat.
- the camera includes a lens, an aperture, a shutter and an image sensor.
- the storage unit 30 stores the images sequentially acquired by the image acquisition device 120 as time-series data. Further, the storage unit 30 stores a dummy image for causing the face detection device 110 to fail in detecting the face of a person in the search area.
- the dummy image is, for example, a uniform image having the same brightness of all pixels.
- the uniform image in the second embodiment is a black image in which the brightness of all pixels is zero.
- the image insertion unit 20 reads a dummy image or an image acquired by the image acquisition device 120 from the storage unit 30 and outputs the dummy image or the image acquired by the image acquisition device 120 to the face detection device 110 based on the determination result of the validity determination unit 10.
- the validity determination unit 10 determines that the face of the person detected by the face detection device 110 is a predetermined detection target person based on the predetermined conditions regarding the face of the person detected by the face detection device 110. Determine if it is a face. In other words, the validity determination unit 10 determines the validity of the detected person's face.
- Predetermined conditions include, for example, conditions relating to the position of the face or the position of the featured part of the face in the image.
- the validity determination unit 10 determines whether or not the position of the face of the detected person or the position of the featured portion of the detected face is included in a predetermined range.
- the predetermined range includes the position of the face or the position of the facial feature of the person to be detected.
- the validity determination unit 10 statistically obtains the amount of variation between the position of the face detected in the previously input image and the position of the face detected in the image input next. It may be determined whether or not it is within the operating range of the person to be detected. In this case, it is premised that the position of the face detected in the previously input image is the position of the face of the person to be detected.
- the operating range is statistically determined based on, for example, information on the seat layout in the vehicle and information on the physique of the driver.
- Predetermined conditions include, for example, conditions related to face size.
- the validity determination unit 10 determines whether or not the size of the face of the detected person is included in the range of the predetermined size.
- the predetermined size range is determined based on, for example, the average face size of the person to be detected in the image.
- the validity determination unit 10 determines that the amount of variation between the face size detected in the previously input image and the face size detected in the next input image is the same as described above. , It may be determined whether or not it is within the operating range of the person to be detected that is statistically determined. In this case, it is premised that the size of the face detected in the previously input image is the size of the face of the person to be detected.
- Predetermined conditions include, for example, conditions related to facial features.
- the validity determination unit 10 determines whether or not the facial features of the detected person match the facial features of the person to be detected that are registered in advance.
- the predetermined conditions shown above were the conditions regarding the relationship between the face of the person detected by the face detection device 110 and the face of the person to be detected in advance, but the determination of validity was made. Does not necessarily require predetermined information on the face of the person to be detected. An example is shown below.
- the predetermined conditions include, for example, a condition relating to the relationship between the running state of the vehicle on which the person detected by the face detection device 110 is on board and the orientation of the face of the detected person.
- the validity determination unit 10 detects the face of the detected person when the angle formed by the traveling direction of the vehicle and the direction of the face of the detected person is included in a predetermined angle range. Judge that it is the face of a person. At this time, the validity determination unit acquires information on the steering angle of the vehicle from the vehicle control device or the like as information on the traveling direction of the vehicle.
- the validity determination unit 10 does not use the predetermined face information of the person to be detected, but obtains the information on the running state of the vehicle and the information on the face of the person detected by the face detection device 110. The validity is judged based on the predetermined conditions regarding.
- the validity determination unit 10 determines that the face of the person to be detected has been detected. When the face of the person to be detected is detected, the validity determination unit 10 outputs the face detection result by the face detection device 110 to the outside. When the face of the non-detection target person is detected, the validity determination unit 10 rejects the face detection result by the face detection device 110 and outputs the detection failure to the outside.
- the face detection processing device 101 has a counter for determining whether or not the face of the person to be detected has been detected.
- the face detection processing device 101 resets the counter.
- the face detection processing device 101 counts up the counter.
- the feature site detection unit 111 detects the face of a person in the search area set in at least a part of the image sequentially input as time series data. At that time, the feature region detection unit 111 detects the face or the feature region of the face by the pattern matching process. As a pattern matching process, the feature region detection unit 111 learns, for example, a face at a specific angle or an image of a feature region of a specific face as a model for each of a plurality of identifiers. The feature portion detection unit 111 evaluates the degree of coincidence between the image of the learning model and the image input from the image insertion unit 20. Based on the evaluation result, the feature portion detection unit 111 detects the face of a person or the feature portion.
- the feature site detection unit 111 performs pattern matching processing in the search area even when the image input from the image insertion unit 20 is a dummy image. However, the feature portion detection unit 111 fails to detect the face of a person in the dummy image.
- the feature site detection unit 111 changes the search area in the image to be input next as time series data based on the success or failure of face detection. For example, when the feature portion detection unit 111 succeeds in detecting the face, the search area in the next input image is set based on the movement amount of the person expected from the time interval of the time series data. For example, when the time series data is composed of 10 fps, the time interval of each image is 100 msec. At that time interval, the range in which the human face is expected to be movable is statistically determined. The feature portion detection unit 111 sets a face existence range estimated based on the expected movement range and the distance from the face position detected in the previous image to the camera as a search area.
- the distance from the face position detected in the previous image to the camera can be estimated from the previous face detection result.
- the feature portion detection unit 111 may set a new search area obtained by expanding the previous search area by a certain amount based on the position of the face detected in the previous image.
- the feature site detection unit 111 when the feature site detection unit 111 fails to detect the face, the feature site detection unit 111 sets the entire area in the next input image as the search area. As described above, when the image input from the image insertion unit 20 is a dummy image, the feature portion detection unit 111 fails to detect the face of the person. Therefore, the search area in the image input next to the dummy image is set to the entire area of the input image.
- the face detection processing device 101 includes the processing circuit shown in FIG. 2 or 3, and the functions of the validity determination unit 10 and the image insertion unit 20 are realized by the processing circuit. Further, the function of the storage unit 30 is realized by, for example, the memory shown in FIG. Similarly, the face detection device 110 also includes the processing circuit shown in FIG. 2 or FIG. 3, and the function of the feature portion detection unit 111 is realized by the processing circuit.
- FIG. 6 is a flowchart showing the face detection processing method according to the second embodiment.
- the flowchart shown in FIG. 6 shows a case where the face detection processing system 200 executes a face detection operation on one image in the time series data.
- the image captured by the image acquisition device 120 is already stored in the storage unit 30.
- step S10 the image insertion unit 20 determines whether or not the value of the counter is equal to or greater than a predetermined value. If the value of the counter is equal to or greater than a predetermined value, step S20 is executed. If the value of the counter is less than a predetermined value, step S30 is executed.
- step S20 the image insertion unit 20 reads a dummy image from the storage unit 30 and outputs it to the feature portion detection unit 111 of the face detection device 110.
- step S30 the image insertion unit 20 outputs the image captured by the image acquisition device 120 from the storage unit 30 to the feature portion detection unit 111 of the face detection device 110. At this time, the image insertion unit 20 outputs to the feature portion detection unit 111 the image next to the image previously output to the face detection device 110 among the images constituting the time series data.
- step S40 the feature site detection unit 111 determines the success or failure of face detection in the previously input image. If the face detection is successful, step S50 is executed. If the face detection was unsuccessful, step S60 is executed.
- step S50 the feature site detection unit 111 sets the search area in the image input from the image insertion unit 20 this time based on the face detection result in the previous image.
- step S60 the feature site detection unit 111 sets the entire image input from the image insertion unit 20 in the search area this time.
- step S70 the feature site detection unit 111 detects the face of the person in the set search area by using the pattern matching process.
- step S80 the validity determination unit 10 determines whether or not the face detection device 110 succeeds in detecting the face of the person. If the face detection device 110 succeeds in detecting the face of the person, step S90 is executed. If the face detection device 110 fails to detect the face of a person, step S120 is executed.
- step S90 the validity determination unit 10 compares the face of the person detected by the face detection device 110 with a predetermined condition and confirms the validity. For example, the validity determination unit 10 compares the position of the detected person's face in the image with a predetermined range in the image.
- step S100 the validity determination unit 10 determines whether or not the face of the person detected by the face detection device 110 satisfies a predetermined condition. In other words, the validity determination unit 10 determines whether or not the face detection result is valid. If the face of the detected person satisfies a predetermined condition, that is, the detection result is valid, step S120 is executed. If the face of the detected person does not meet the predetermined conditions, that is, the detection result is not valid, step S110 is executed.
- step S110 the face detection processing device 101 counts up the counter. In this state, the face detection device 110 succeeds in detecting the face of the person, but fails to detect the face of the person to be detected.
- step S120 the face detection processing device 101 resets the counter. In this state, the face detection device 110 succeeds in detecting the face of the person to be detected, or the face detection device 110 fails to detect the face of the person in the first place.
- step S10 to S120 are processing for one image in the time series data. Therefore, in practice, after step S110 or S120, step S10 is executed again as a process for the next image.
- the image acquisition device 120 is a camera mounted on the vehicle and images the passengers in the vehicle, and the predetermined detection target person is a driver among the passengers of the vehicle.
- FIG. 7 is an example of an image captured by the image acquisition device 120, and is a diagram showing an image output by the image insertion unit 20 to the feature portion detection unit 111. The image shows four passengers 2, including the driver of the vehicle. Further, in the following, the case where the face detection device 110 succeeds in detecting the face in the image immediately before the image of FIG. 7 will be described.
- step S40 the feature site detection unit 111 determines that the face has been successfully detected in the previous image.
- step S50 is executed.
- step S50 the feature site detection unit 111 sets the search area based on the face detection result in the previous image.
- the feature portion detection unit 111 sets the search region 3 in the region on the left side of the image corresponding to the driver's seat of the vehicle, as shown in FIG.
- step S70 the feature site detection unit 111 detects the face of the person in the search area 3.
- a part of the face of the passenger 2A in the front seat, which is the driver, and the entire face of the passenger 2B in the rear seat are shown in the search area 3.
- the state as shown in FIG. 7 can be obtained.
- the feature portion detecting unit 111 preferentially detects the face of the passenger 2B in the rear seat.
- step S80 the validity determination unit 10 determines that the feature site detection unit 111 has succeeded in detecting the face.
- Step S90 is executed.
- step S90 the validity determination unit 10 compares the coordinates in the image of the face of the passenger 2B in the rear seat detected by the feature site detection unit 111 with the coordinate range in which the driver's face should exist. And confirm the validity.
- step S100 the validity determination unit 10 determines that the coordinates of the face of the passenger 2B in the rear seat are not included in the coordinate range in which the driver's face should exist. That is, the validity determination unit 10 determines that the detection of the face of the driver to be detected has failed, and the step S110 is executed.
- step S110 the face detection processing device 101 counts up the counter. Subsequently, step S10 is executed.
- step S10 the image insertion unit 20 determines whether or not the value of the counter is equal to or greater than a predetermined value.
- a predetermined value the case where the value of the counter reaches a predetermined value or more will be described.
- step S20 the image insertion unit 20 reads a dummy image from the storage unit 30 and outputs it to the feature portion detection unit 111 of the face detection device 110.
- the dummy image is a black image.
- step S40 the feature site detection unit 111 determines that the face of the passenger 2B in the rear seat has been successfully detected in the image of FIG. 7, that is, the previous image.
- step S50 is executed.
- step S50 the feature region detection unit 111 sets the search region 3 in the region on the left side of the dummy image based on the face detection result in the image of FIG. 7.
- step S70 the feature site detection unit 111 attempts to detect the face of the person in the search area 3 in the dummy image. However, since the dummy image is a black image, the feature portion detection unit 111 fails to detect the face of the person in the dummy image.
- step S80 the validity determination unit 10 determines that the feature region detection unit 111 has failed to detect the face.
- step S120 is executed.
- step S120 the face detection processing device 101 resets the counter. Step S10 is executed again.
- step S10 the image insertion unit 20 determines whether or not the value of the counter is equal to or greater than a predetermined value.
- the value of the counter is less than a predetermined value because it has been reset.
- step S30 is executed.
- step S30 the image insertion unit 20 reads the image next to the image previously output to the face detection device 110 among the time series data stored in the storage unit 30, and outputs the image to the face detection device 110. ..
- the image insertion unit 20 may skip the frame in which the dummy image is read out of the time-series data stored in the storage unit 30 and read the next image.
- FIG. 8 is an example of an image captured by the image acquisition device 120, and is a diagram showing an image output by the image insertion unit 20 to the feature portion detection unit 111 in this step S30.
- the image shown in FIG. 8 is an image captured by the image acquisition device 120 after the image shown in FIG. 7.
- step S40 the feature portion detection unit 111 determines that the face detection in the previous dummy image has failed.
- step S60 is executed.
- step S60 the feature region detection unit 111 sets the entire image shown in FIG. 8 in the search region 3.
- step S70 the feature site detection unit 111 detects the faces of the four passengers 2 shown in the search area 3.
- step S80 the validity determination unit 10 determines that the feature site detection unit 111 has succeeded in detecting the face.
- step S90 is executed.
- step S90 the validity determination unit 10 compares the coordinates in the image of the faces of the four passengers 2 detected by the feature site detection unit 111 with the coordinate range in which the driver's face should exist. Check the validity.
- step S100 the validity determination unit 10 determines that the coordinates of the face of one passenger 2A in the front seat among the four passengers 2 are included in the coordinate range in which the driver's face should exist. to decide. That is, the validity determination unit 10 determines that the face of the driver to be detected has been successfully detected, and the step S120 is executed.
- step S120 the face detection processing device 101 resets the counter.
- the face detection device 110 limits the search area 3 and detects the face of the person to be detected. Therefore, the amount of calculation for face detection is reduced, and the processing speed is improved.
- the face detection processing device 101 rejects the face detection result of the non-detection target person and sends a dummy image to the face detection device 110. Output. Since the dummy image in the second embodiment is a black image, the face detection device 110 fails to detect the face in the input dummy image. Therefore, the face detection device 110 changes the search area 3 in the image input thereafter.
- the face detection device 110 Since the face detection device 110 performs a face detection operation in the changed search area 3, it can detect the face of a person to be detected that satisfies a predetermined condition.
- the face detection device 110 and the face detection processing system 200 of the second embodiment improve the face detection speed of the face detection device 110 and also improve the face detection accuracy of the person to be detected.
- the predetermined conditions in the second embodiment are the condition regarding the position of the face in the image or the position of the feature portion of the face in the image, the condition regarding the size of the face in the image, and the condition regarding the feature of the face. Or, the condition regarding the relationship between the running state of the vehicle on which the person is on board and the orientation of the face is included.
- the face detection processing device 101 enables the face detection device 110 to accurately detect the face of the person to be detected even under various conditions.
- the face detection processing device 101 executes a determination process regarding the validity of the detection result of the face detection device 110, but the determination process can also be incorporated into the face detection device 110. .. However, its incorporation requires modification or new design of the face detection device 110.
- the face detection processing device 101 according to the second embodiment has the above-mentioned effect only by connecting to the existing image acquisition device 120 and the face detection device 110. That is, the face detection processing device 101 enables the face detection device 110 to accurately detect the face of the person to be detected without changing the input / output interfaces of the existing image acquisition device 120 and the face detection device 110. To.
- the face detection processing device 101 and the face detection processing system 200 can be incorporated into a driver monitoring system mounted on a moving body.
- the driver monitoring system can output a warning regarding dozing and inattentiveness to the driver based on the detection result output by the face detection processing device 101.
- the driving support device for supporting the driving of the vehicle or the automatic driving control device for controlling the automatic driving of the vehicle is linked with the driver monitoring system, the portion is based on the detection result output by the face detection processing device 101. It is possible to accurately determine the transfer of driving initiative from a vehicle to a human in a self-driving vehicle.
- the face detection processing device 101 and the face detection processing system 200 can be incorporated into a personal authentication system that authenticates based on the face of a person.
- the personal authentication system can accurately perform personal authentication based on the detection result output by the face detection processing device 101.
- the storage unit 30 may store the processed image as a dummy image.
- the processed image is an image in which the face of a person whose validity has been denied by the validity determination unit 10 is filled.
- the processed image is an image in which the face of the passenger 2B in the rear seat shown in FIG. 7 is filled.
- step S20 of the second embodiment the image insertion unit 20 outputs the processed image to the feature portion detection unit 111 of the face detection device 110 instead of the black image. Similar to the black image, in step S70, the feature portion detecting unit 111 fails to detect the face of the person in the search area 3. Other processing is the same as in the second embodiment. Such a face detection processing device 101 also has the same effect as that of the second embodiment.
- the storage unit 30 may store a mask image as a dummy image.
- the mask image is an image for being superimposed on the image acquired by the image acquisition device 120.
- the mask image makes the pixels corresponding to the face of the person whose validity is denied by the validity determination unit 10 opaque. For example, by superimposing the mask image on the image shown in FIG. 7, a state is generated in which the pixel corresponding to the face of the passenger 2B in the rear seat is not transmitted and the face is not detected by the face detection device 110.
- the image insertion unit 20 may output an image in which the mask image is superimposed on the next image acquired by the image acquisition device 120 instead of the black image. Similar to the black image, in step S70, the feature portion detecting unit 111 fails to detect the face of the person in the search area 3. Other processing is the same as in the second embodiment. Such a face detection processing device 101 also has the same effect as that of the second embodiment.
- the face detection processing device 101 and the face detection processing method according to the third embodiment will be described.
- the third embodiment is a subordinate concept of the first embodiment, and the face detection processing device 101 in the third embodiment includes each configuration of the face detection processing device 100 in the first embodiment.
- the description of the configuration and operation similar to those of the first or second embodiment will be omitted.
- the configuration of the face detection processing device 101 and the face detection processing system 200 in the third embodiment is the same as the block diagram shown in FIG.
- the storage unit 30 in the third embodiment stores a black image and a simulated image as dummy images.
- the simulated image is an image in which the face of a dummy person who satisfies a predetermined condition and is simulated by a predetermined person to be detected is drawn. For example, when the person to be detected is the driver of the vehicle, the face of the dummy person is drawn in the range where the driver's face should exist in the image acquired by the image acquisition device 120.
- FIG. 9 is a flowchart showing the face detection processing method according to the third embodiment. Step S115 and step S130 are added to the flowchart shown in FIG. 6 of the second embodiment. Based on the flowchart of FIG. 9, the detailed operation of the face detection device 110 resetting the search area 3 by the dummy image will be described. Since the same as in the second embodiment up to step S70 in which the feature portion detecting unit 111 fails to detect the face of the person in the search area 3 of the black image which is a dummy image, the description thereof will be omitted.
- step S80 the validity determination unit 10 determines that the feature region detection unit 111 has failed to detect the face.
- step S115 is executed.
- step S115 the validity determination unit 10 determines whether or not the image output to the feature portion detection unit 111 is a dummy image.
- step S130 is executed.
- step S130 the face detection processing device 101 holds the counter. Step S10 is executed again.
- step S10 the image insertion unit 20 determines whether or not the value of the counter is equal to or greater than a predetermined value. Here, since the value of the counter is held, it still reaches a predetermined value or more. Therefore, step S20 is executed.
- step S20 the image insertion unit 20 reads a dummy image from the storage unit 30 and outputs it to the feature portion detection unit 111 of the face detection device 110. At this time, the image insertion unit 20 remembers that the black image was read as the dummy image last time.
- the image insertion unit 20 reads a simulated image as a dummy image next to the black image and outputs the simulated image to the feature portion detection unit 111.
- FIG. 10 is a diagram showing an example of a simulated image according to the third embodiment. The position of the dummy passenger 2C's face in the simulated image corresponds to the position of the driver's face when the driver is seated in the driver's seat and is facing the front.
- step S40 the feature region detection unit 111 determines that the face detection in the previous black image has failed.
- step S60 is executed.
- step S60 the feature region detection unit 111 sets the entire image shown in FIG. 10 in the search region 3.
- step S70 the feature site detection unit 111 detects the face of the dummy passenger 2C shown in the search area 3.
- step S80 the validity determination unit 10 determines that the feature site detection unit 111 has succeeded in detecting the face.
- step S90 is executed.
- step S90 the validity determination unit 10 compares the coordinates in the image of the face of the dummy passenger 2C detected by the feature site detection unit 111 with the coordinate range in which the driver's face should exist. Check the validity.
- step S100 the validity determination unit 10 determines that the coordinates of the face of the dummy passenger 2C are included in the coordinate range in which the driver's face should exist. That is, the validity determination unit 10 determines that the face of the driver to be detected has been successfully detected, and the step S120 is executed.
- step S120 the face detection processing device 101 resets the counter. Step S10 is executed again.
- step S10 the image insertion unit 20 determines whether or not the value of the counter is equal to or greater than a predetermined value.
- the value of the counter is less than a predetermined value because it has been reset.
- step S30 is executed.
- step S30 the image insertion unit 20 reads the image next to the image previously output to the face detection device 110 among the time series data stored in the storage unit 30, and outputs the image to the face detection device 110.
- FIG. 11 is an example of an image captured by the image acquisition device 120, and is a diagram showing an image output by the image insertion unit 20 to the feature portion detection unit 111 in this step S30.
- step S40 the feature site detection unit 111 determines that the face has been successfully detected in the previous simulated image.
- step S50 is executed.
- step S50 the feature site detection unit 111 sets the search area 3 based on the face detection result in the previous simulated image.
- the feature region detection unit 111 sets the search region 3 in the region corresponding to the face of the dummy passenger 2C shown in FIG. 10 in the image of FIG.
- step S70 the feature site detection unit 111 detects the face of the passenger 2 in the search area 3.
- the search area 3 the entire face of the passenger 2A in the front seat and a part of the face of the passenger 2B in the rear seat are shown.
- the feature portion detecting unit 111 preferentially detects the face of the passenger 2A in the front seat.
- step S80 the validity determination unit 10 determines that the feature site detection unit 111 has succeeded in detecting the face.
- Step S90 is executed.
- step S90 the validity determination unit 10 compares the coordinates in the image of the face of the passenger 2A in the front seat detected by the feature site detection unit 111 with the coordinate range in which the driver's face should exist. Check the validity.
- step S100 the validity determination unit 10 determines that the coordinates of the face of the passenger 2A in the front seat are included in the coordinate range in which the driver's face should exist. That is, the validity determination unit 10 determines that the face of the driver to be detected has been successfully detected, and the step S120 is executed.
- step S120 the face detection processing device 101 resets the counter.
- the dummy image in the third embodiment is a black image and a simulated image in which the face of the dummy person is drawn.
- the face detection device 110 After failing to detect a face in the dummy image, the face detection device 110 changes the search area to an appropriate search area by using a simulated image. Since the face detection device 110 performs a face detection operation in the appropriately changed search area 3, it can detect the face of a person to be detected that satisfies a predetermined condition.
- the face detection device 110 and the face detection processing system 200 of the third embodiment improve the face detection speed of the face detection device 110 and also improve the face detection accuracy of the person to be detected.
- the face detection processing device shown in each of the above embodiments shall also be applied to a system constructed by appropriately combining a navigation device, a communication terminal, a server, and the functions of applications installed in the navigation device. Can be done.
- the navigation device includes, for example, a PND (Portable Navigation Device) and the like.
- Communication terminals include, for example, mobile terminals such as mobile phones, smartphones and tablets.
- FIG. 12 is a block diagram showing the configuration of the face detection processing device 100 and the device operating in connection therewith according to the fourth embodiment.
- the face detection processing device 100, the face detection device 110, and the communication device 130 are provided in the server 300.
- the face detection processing device 100 acquires an image of the inside of the vehicle 1 from the image acquisition device 120 provided in the vehicle 1 via the communication device 140 and the communication device 130.
- the face detection processing device 100 determines whether or not the face of the person detected by the face detection device 110 satisfies a predetermined condition.
- the face detection processing device 100 outputs a dummy image or an image acquired from the image acquisition device 120 to the face detection device 110 based on the determination result. Further, the face detection processing device 100 outputs the face detection result by the face detection device 110 to the control device 150 provided in the vehicle 1 via each communication device.
- the control device 150 provides notification and driving support for the vehicle 1 based on the detection result.
- the face detection processing device 100 By arranging the face detection processing device 100 on the server 300 in this way, the configuration of the in-vehicle device can be simplified.
- the face detection processing device 100 may be provided in the server 300, and the other part may be provided in the vehicle 1 in a distributed manner.
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Human Computer Interaction (AREA)
- Image Processing (AREA)
- Image Analysis (AREA)
Abstract
Description
図1は、実施の形態1における顔検出処理装置100の構成を示すブロック図である。
実施の形態2における顔検出処理装置および顔検出処理方法を説明する。実施の形態2は実施の形態1の下位概念であり、実施の形態2における顔検出処理装置は、実施の形態1における顔検出処理装置100の各構成を含む。なお、実施の形態1と同様の構成および動作については説明を省略する。
記憶部30は、ダミー画像として、加工画像を記憶していてもよい。加工画像とは、妥当性判断部10による妥当性が否定された人物の顔が塗りつぶされた画像である。例えば、加工画像とは、図7に示される後方の座席の搭乗者2Bの顔が塗りつぶされた画像である。
記憶部30は、ダミー画像として、マスク画像を記憶していてもよい。マスク画像とは、画像取得装置120によって取得された画像に重ね合わせるための画像である。マスク画像は、妥当性判断部10による妥当性が否定された人物の顔に対応する画素を不透過にする。例えば、マスク画像を、図7に示される画像に重ね合わせることにより、後方の座席の搭乗者2Bの顔に対応する画素が透過せず、その顔が顔検出装置110によって検出されない状態が生成される。
実施の形態3における顔検出処理装置101および顔検出処理方法を説明する。実施の形態3は実施の形態1の下位概念であり、実施の形態3における顔検出処理装置101は、実施の形態1における顔検出処理装置100の各構成を含む。なお、実施の形態1または2と同様の構成および動作については説明を省略する。
以上の各実施の形態に示された顔検出処理装置は、ナビゲーション装置と、通信端末と、サーバと、これらにインストールされるアプリケーションの機能とを適宜に組み合わせて構築されるシステムにも適用することができる。ここで、ナビゲーション装置とは、例えば、PND(Portable Navigation Device)などを含む。通信端末とは、例えば、携帯電話、スマートフォンおよびタブレットなどの携帯端末を含む。
Claims (5)
- 順次入力される画像の少なくとも一部の領域に設定される探索領域に写っている人物の顔を検出し、かつ、前記人物の前記顔の検出の成否に基づいて次に入力される画像における前記探索領域を変更する顔検出装置に対し、前記画像を出力する顔検出処理装置であって、
前記顔検出装置によって検出された前記人物の前記顔に関する予め定められた条件に基づいて、前記顔検出装置によって検出された前記人物の前記顔が、予め定められた検出対象の人物の顔であるか否かを判定する妥当性判断部と、
前記妥当性判断部の判定結果に基づいて、前記探索領域における前記人物の前記顔の前記検出を前記顔検出装置に失敗させて前記探索領域を変更させるためのダミー画像を、前記画像として、前記顔検出装置に出力する画像挿入部と、を備える顔検出処理装置。 - 前記ダミー画像は、全画素の輝度が同じ均一画像である、請求項1に記載の顔検出処理装置。
- 前記ダミー画像は、前記予め定められた条件を満たし、かつ、前記予め定められた検出対象の前記人物が模擬されたダミーの人物の顔が描かれた模擬画像である、請求項1に記載の顔検出処理装置。
- 前記予め定められた条件とは、前記画像内での前記顔の位置もしくは前記画像内での前記顔の特徴部位の位置に関する条件、前記画像内での前記顔の大きさに関する条件、前記顔の特徴に関する条件、または、前記人物が搭乗している車両の走行状態と前記顔の向きとの関係に関する条件を含む、請求項1に記載の顔検出処理装置。
- 順次入力される画像の少なくとも一部の領域に設定される探索領域に写っている人物の顔を検出し、かつ、前記人物の前記顔の検出の成否に基づいて次に入力される画像における前記探索領域を変更する顔検出装置に対し、前記画像を出力する顔検出処理方法であって、
前記顔検出装置によって検出された前記人物の前記顔に関する予め定められた条件に基づいて、前記顔検出装置によって検出された前記人物の前記顔が、予め定められた検出対象の人物の顔であるか否かを判定し、
判定結果に基づいて、前記探索領域における前記人物の前記顔の前記検出を前記顔検出装置に失敗させて前記探索領域を変更させるためのダミー画像を、前記画像として、前記顔検出装置に出力する、顔検出処理方法。
Priority Applications (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| DE112019007033.9T DE112019007033B4 (de) | 2019-07-09 | 2019-07-09 | Gesichtsdetektionsverarbeitungsvorrichtung und Gesichtsdetektionsverarbeitungsverfahren |
| JP2021530389A JP7051014B2 (ja) | 2019-07-09 | 2019-07-09 | 顔検出処理装置および顔検出処理方法 |
| PCT/JP2019/027103 WO2021005702A1 (ja) | 2019-07-09 | 2019-07-09 | 顔検出処理装置および顔検出処理方法 |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| PCT/JP2019/027103 WO2021005702A1 (ja) | 2019-07-09 | 2019-07-09 | 顔検出処理装置および顔検出処理方法 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2021005702A1 true WO2021005702A1 (ja) | 2021-01-14 |
Family
ID=74114457
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/JP2019/027103 Ceased WO2021005702A1 (ja) | 2019-07-09 | 2019-07-09 | 顔検出処理装置および顔検出処理方法 |
Country Status (3)
| Country | Link |
|---|---|
| JP (1) | JP7051014B2 (ja) |
| DE (1) | DE112019007033B4 (ja) |
| WO (1) | WO2021005702A1 (ja) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2024122476A1 (ja) * | 2022-12-06 | 2024-06-13 | キヤノン株式会社 | 情報処理装置、情報処理方法、及びプログラム |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2009026146A (ja) * | 2007-07-20 | 2009-02-05 | Canon Inc | 画像処理装置及び画像処理方法 |
| JP2011138388A (ja) * | 2009-12-28 | 2011-07-14 | Canon Inc | データ補正装置及び方法 |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP4367624B2 (ja) | 2004-01-20 | 2009-11-18 | オムロン株式会社 | 運転中における電話使用時の車輌制御装置とその方法 |
| DE112019007522T5 (de) | 2019-07-02 | 2022-05-05 | Mitsubishi Electric Corporation | Onboard-Bildverarbeitungsvorrichtung und Onboard-Bildverarbeitungsverfahren |
-
2019
- 2019-07-09 WO PCT/JP2019/027103 patent/WO2021005702A1/ja not_active Ceased
- 2019-07-09 JP JP2021530389A patent/JP7051014B2/ja active Active
- 2019-07-09 DE DE112019007033.9T patent/DE112019007033B4/de active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2009026146A (ja) * | 2007-07-20 | 2009-02-05 | Canon Inc | 画像処理装置及び画像処理方法 |
| JP2011138388A (ja) * | 2009-12-28 | 2011-07-14 | Canon Inc | データ補正装置及び方法 |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2024122476A1 (ja) * | 2022-12-06 | 2024-06-13 | キヤノン株式会社 | 情報処理装置、情報処理方法、及びプログラム |
Also Published As
| Publication number | Publication date |
|---|---|
| JPWO2021005702A1 (ja) | 2021-10-14 |
| DE112019007033T5 (de) | 2021-12-09 |
| JP7051014B2 (ja) | 2022-04-08 |
| DE112019007033B4 (de) | 2025-02-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| JP6259132B2 (ja) | 車載用カメラ装置 | |
| CN113784875B (zh) | 相机位置检测装置和方法、相机单元以及存储介质 | |
| US11189048B2 (en) | Information processing system, storing medium storing program, and information processing device controlling method for performing image processing on target region | |
| JP7109649B2 (ja) | 覚醒度推定装置、自動運転支援装置および覚醒度推定方法 | |
| JP2021189852A (ja) | 状態判定装置および状態判定方法 | |
| JP7051014B2 (ja) | 顔検出処理装置および顔検出処理方法 | |
| US11440407B2 (en) | Non-contact operating apparatus for vehicle and vehicle | |
| JP2008258778A (ja) | 撮像システム | |
| KR20230096530A (ko) | 주행 제어 장치 및 방법 | |
| JP2022143854A (ja) | 乗員状態判定装置および乗員状態判定方法 | |
| JP2023077916A (ja) | 車両乗員の移動距離測定装置、移動距離測定プログラム、及び移動距離測定方法 | |
| JP2019074965A (ja) | 運転不能状態検出装置及び運転不能状態検出システム | |
| CN119329552A (zh) | 监视装置、计算机程序产品以及方法 | |
| JP4539427B2 (ja) | 画像処理装置 | |
| US20230306707A1 (en) | Around view monitoring system and the method thereof | |
| JP2019114118A (ja) | 警告装置、自動車、警告方法、及びプログラム | |
| JP6847323B2 (ja) | 視線検出装置および視線検出方法 | |
| JP2018113622A (ja) | 画像処理装置、画像処理システム、及び、画像処理方法 | |
| JP7837412B2 (ja) | 操作判定装置及び操作判定方法 | |
| WO2021240668A1 (ja) | ジェスチャ検出装置およびジェスチャ検出方法 | |
| WO2021186710A1 (ja) | ジェスチャ検出装置及びジェスチャ検出方法 | |
| JPWO2020129146A1 (ja) | 表示制御装置及び表示制御方法 | |
| JP7483060B2 (ja) | 手検出装置、ジェスチャー認識装置および手検出方法 | |
| US20240174248A1 (en) | Vehicle warning apparatus | |
| JP2020160201A (ja) | 表示制御装置、表示装置、表示制御方法およびプログラム |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 19936915 Country of ref document: EP Kind code of ref document: A1 |
|
| ENP | Entry into the national phase |
Ref document number: 2021530389 Country of ref document: JP Kind code of ref document: A |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 19936915 Country of ref document: EP Kind code of ref document: A1 |
|
| WWG | Wipo information: grant in national office |
Ref document number: 112019007033 Country of ref document: DE |