CN111046792A - Face detection method and device, electronic equipment and computer readable storage medium - Google Patents
Face detection method and device, electronic equipment and computer readable storage medium Download PDFInfo
- Publication number
- CN111046792A CN111046792A CN201911268526.7A CN201911268526A CN111046792A CN 111046792 A CN111046792 A CN 111046792A CN 201911268526 A CN201911268526 A CN 201911268526A CN 111046792 A CN111046792 A CN 111046792A
- Authority
- CN
- China
- Prior art keywords
- face
- image
- detection model
- detected
- full convolution
- 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.)
- Pending
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 131
- 238000000034 method Methods 0.000 claims description 40
- 238000004590 computer program Methods 0.000 claims description 7
- 238000004422 calculation algorithm Methods 0.000 description 8
- 238000010586 diagram Methods 0.000 description 8
- 230000008569 process Effects 0.000 description 8
- 238000004364 calculation method Methods 0.000 description 7
- 238000004891 communication Methods 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 238000004080 punching Methods 0.000 description 3
- 230000008859 change Effects 0.000 description 2
- 230000002411 adverse Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000000802 evaporation-induced self-assembly Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000001629 suppression Effects 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR 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
- G06V40/168—Feature extraction; Face representation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G07—CHECKING-DEVICES
- G07C—TIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE
- G07C1/00—Registering, indicating or recording the time of events or elapsed time, e.g. time-recorders for work people
- G07C1/10—Registering, indicating or recording the time of events or elapsed time, e.g. time-recorders for work people together with the recording, indicating or registering of other data, e.g. of signs of identity
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Human Computer Interaction (AREA)
- Multimedia (AREA)
- Image Analysis (AREA)
Abstract
The embodiment of the invention provides a face detection method, a face detection device, electronic equipment and a computer readable storage medium, and relates to the field of computer vision, wherein the face detection method comprises the following steps: acquiring an image of a snapshot machine, wherein the image comprises a face to be detected; and inputting the image into a pre-trained full convolution network detection model to obtain a face boundary box of the face to be detected. Therefore, according to the technical scheme provided by the embodiment, the full convolution network detection model based on the FCN network structure is directly adopted for face detection aiming at the image shot by the snapshot machine, so that the problem of high time consumption caused by the dependence of a face detection mode based on the FPN network on a characteristic pyramid structure and a multi-anchor strategy in the prior art is solved, and the detection speed is improved.
Description
Technical Field
The invention relates to the field of computer vision, in particular to a face detection method, a face detection device, electronic equipment and a computer readable storage medium.
Background
In the staff attendance system applying the face recognition, the face detection is the basis of the face recognition, and the accuracy of the face detection directly determines the accuracy of the face recognition.
At present, an existing staff attendance system firstly acquires an image by using a snapshot machine, and then performs face detection based on a face detection algorithm of a FPN (feature pyramid Network) to complete card punching.
However, in the face attendance system, most of the images acquired by the snapshot machine are 1 face, and under the condition that the face scale change is small, the feature pyramid structure and the multi-anchor strategy of the FPN-based face detection algorithm cause more time consumption, and the detection speed is low.
Disclosure of Invention
In view of the above, the present invention provides a face detection method, a face detection apparatus, an electronic device, and a computer-readable storage medium.
In order to achieve the above purpose, the embodiment of the present invention adopts the following technical solutions:
in a first aspect, an embodiment of the present invention provides a face detection method, including:
acquiring an image of a snapshot machine, wherein the image comprises a face to be detected;
and determining the face boundary frame of the face to be detected according to the image by applying a pre-trained full convolution network detection model.
In an optional embodiment, the determining, by using a pre-trained full convolution network detection model, a face bounding box of the face to be detected according to the image includes:
applying the pre-trained full convolution network detection model and outputting bias characteristics according to the image; the bias feature is used for representing the bias of the corner points of the face boundary box of the face to be detected relative to the center of the image;
determining the face bounding box based on the biased features.
In an alternative embodiment, the bias feature comprises a bias coordinate; the offset coordinates are used for representing the offset of the corner point coordinates of the face boundary box of the face to be detected relative to the central coordinates of the image;
alternatively, the first and second electrodes may be,
the offset feature comprises an offset distance; the offset distance is used for representing the distance between the boundary line of the face boundary frame of the face to be detected and the center of the image.
In an optional embodiment, the determining, by using a pre-trained full convolution network detection model, a face bounding box of the face to be detected according to the image includes:
detecting the image by applying a pre-trained full convolution network detection model to obtain a human face boundary box to be confirmed;
determining the face key points of the face to be detected according to the image by applying a pre-trained key point detection model;
and if the face key point is in the face boundary box to be confirmed, determining the face boundary box to be confirmed as the face boundary box of the face to be detected.
In an alternative embodiment, the method further comprises:
and carrying out duplication removal processing on the face boundary frame of the face to be detected to obtain a target face boundary frame.
In a second aspect, an embodiment of the present invention provides a face detection method, including:
acquiring training sample data; the training sample data comprises image samples of a plurality of snapshot machines, and the image samples comprise a predetermined face boundary box and pixel-level labels of corner points of the face boundary box;
and applying the training sample data to train the initial full convolution network detection model to obtain the trained full convolution network detection model.
In an alternative embodiment, the image sample further comprises a predetermined pixel-level annotation of a face keypoint; the method further comprises the following steps:
and applying the training sample data to train the initial key point detection model to obtain the trained key point detection model.
In a third aspect, an embodiment of the present invention provides a face detection apparatus, including:
the acquisition module is used for acquiring an image of the snapshot machine, wherein the image comprises a face to be detected;
and the output module is used for applying a pre-trained full convolution network detection model and determining the face boundary frame of the face to be detected according to the image.
In a fourth aspect, an embodiment of the present invention provides an electronic device, including: a processor and a memory; the processor is connected with the memory;
the memory is used for storing programs;
the processor is configured to execute a program stored in the memory to perform the method of any of the preceding embodiments.
In a fifth aspect, the present invention provides a computer-readable storage medium, on which a computer program is stored, and the computer program, when executed by a processor, implements the method according to any one of the foregoing embodiments.
The embodiment of the invention has the following beneficial effects: the embodiment of the invention provides a face detection method, a face detection device, electronic equipment and a computer readable storage medium, wherein the face detection method comprises the following steps: acquiring an image of a snapshot machine, wherein the image comprises a face to be detected; and determining the face boundary frame of the face to be detected according to the image by applying a pre-trained full convolution network detection model. Therefore, according to the technical scheme provided by the embodiment, the full convolution network detection model based on the FCN network structure is directly adopted for face detection aiming at the image shot by the snapshot machine, so that the problem of high time consumption caused by the dependence of a face detection mode based on the FPN network on a characteristic pyramid structure and a multi-anchor strategy in the prior art is solved, and the detection speed is improved.
In order to make the aforementioned and other objects, features and advantages of the present invention comprehensible, preferred embodiments accompanied with figures are described in detail below.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the embodiments will be briefly described below, it should be understood that the following drawings only illustrate some embodiments of the present invention and therefore should not be considered as limiting the scope, and for those skilled in the art, other related drawings can be obtained according to the drawings without inventive efforts.
Fig. 1 shows a flowchart of a face detection method according to an embodiment of the present invention;
FIG. 2 is a flow chart of another face detection method provided by the embodiment of the invention;
FIG. 3 is a schematic diagram illustrating a face detection method according to an embodiment of the present invention;
fig. 4 is a schematic diagram illustrating a face detection apparatus according to an embodiment of the present invention;
fig. 5 is a flow chart showing an actual process of a face detection apparatus according to an embodiment of the present invention;
FIG. 6 is a schematic diagram of another face detection apparatus according to an embodiment of the present invention;
fig. 7 shows a schematic diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. The components of embodiments of the present invention generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations.
Thus, the following detailed description of the embodiments of the present invention, presented in the figures, is not intended to limit the scope of the invention, as claimed, but is merely representative of selected embodiments of the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments of the present invention without making any creative effort, shall fall within the protection scope of the present invention.
It is noted that relational terms such as "first" and "second," and the like, may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
The face detection is the basis of face recognition, and the accuracy of face detection directly determines the accuracy of face recognition. At present, an existing employee face card-punching attendance system firstly acquires an image by using a snapshot machine, then acquires a face image and face key points based on a face detection algorithm of an FPN network, finally aligns the face image by using the key points, inputs the face image into an identification network to extract face features, retrieves similarity with the face features of a base, and completes face identification card punching.
However, in the face attendance system, the image of the snapshot machine is obtained by outwardly expanding a certain multiple of the face area, and only a single face is arranged on each image, so that the size of the image obtained by the snapshot machine is not changed greatly and is in a certain proportion to the size of the face.
Based on this, the face detection method, the face detection device, the electronic device and the computer-readable storage medium provided by the embodiment of the invention combine the characteristic that the face size of the image of the snapshot machine is relatively fixed, and perform face detection by adopting the full convolution network detection model based on the network structure of the FCN, so that the problem of much time consumption caused by dependence on the feature pyramid structure and the multi-anchor strategy in the face detection mode based on the FPN network in the prior art is solved, and the detection speed is improved.
The term anchor mentioned above means various bounding boxes that can be acquired in advance to better fit the detected object according to the kind and size of the object to be detected.
Referring to fig. 1, an embodiment of the present application provides a face detection method, which belongs to the field of computer vision and is executed by electronic equipment in the field.
Specifically, the method is applied to a face detection stage and comprises the following steps:
step S102, acquiring an image of a snapshot machine, wherein the image comprises a face to be detected;
and step S104, determining a face boundary frame of the face to be detected according to the image by applying a pre-trained full convolution network detection model.
In step S102, the image refers to an image to be detected or a target image, and the image captured by the capturing machine is used as a network input.
For step S104, after the image shot by the snapshot machine is acquired, the image shot by the snapshot machine is input into a pre-trained full convolution network detection model, and the full convolution network detection model can detect the face to be detected therein according to the image, and output a face boundary box of the face to be detected, wherein the face boundary box indicates a detected face region, and defines a face detection range.
The face detection method provided by the embodiment of the invention comprises the steps of obtaining an image of a snapshot machine, wherein the image comprises a face to be detected; and determining the face boundary frame of the face to be detected according to the image by applying a pre-trained full convolution network detection model. According to the method, the image of the snapshot machine is obtained, and the characteristic that the face size of the snapshot machine is fixed is combined, the trained full convolution network detection model based on the FCN network structure is directly adopted to detect the input image of the snapshot machine, so that time consumption caused by a characteristic pyramid structure and a multi-anchor strategy is avoided, and the detection speed is improved.
Alternatively, the step S104 may be performed by:
a, applying the pre-trained full convolution network detection model and outputting bias characteristics according to the image; the bias feature is used for representing the bias of the corner points of the face boundary box of the face to be detected relative to the center of the image;
and B, determining the face bounding box based on the bias characteristics.
Wherein the bias feature comprises a bias coordinate; the offset coordinates are used for representing the offset of the corner point coordinates of the face boundary box of the face to be detected relative to the central coordinates of the image;
alternatively, the first and second electrodes may be,
the offset feature comprises an offset distance; the offset distance is used for representing the distance between the boundary line of the face boundary frame of the face to be detected and the center of the image.
Accordingly, step a may be performed by one of the following:
the first method is as follows:
1. applying the pre-trained full convolution network detection model, and outputting a bias coordinate according to the image, wherein the bias coordinate is used for representing the bias of the corner point coordinate of the face boundary box of the face to be detected relative to the center coordinate of the image;
2. and determining the corner coordinates of the face bounding box and the face bounding box based on the offset coordinates and the central coordinates of the image, wherein the corner coordinates of the face bounding box and the face bounding box can be determined in two directions.
The second method comprises the following steps:
1) and outputting an offset distance according to the image by using the pre-trained full convolution network detection model, wherein the offset distance is used for representing the distance between the boundary line of the face boundary frame of the face to be detected and the center of the image.
2) Determining the boundary line of the face boundary frame and the face boundary frame based on the offset distance and the center of the image, wherein the boundary line of the face boundary frame and the face boundary frame can also be determined in two directions.
In consideration of the fact that the center of the face of the image of the snapshot machine is located at the center of the image, and the size of the image is in a certain proportion to the size of the face of the person, the face bounding box of the embodiment is rectangular and is provided with four corner points and four boundary lines; therefore, the offset change of the four corner points of the face bounding box in each picture relative to the image center is small, so that the offset of the four corner points of the face bounding box relative to the image center is directly used as the output of the network, the selection of an anchor is further avoided, and the detection speed is further improved.
Optionally, the step S104 includes:
(1) detecting the image by applying a pre-trained full convolution network detection model to obtain a human face boundary box to be confirmed;
(2) determining the face key points of the face to be detected according to the image by applying a pre-trained key point detection model;
(3) and if the face key point is in the face boundary box to be confirmed, determining the face boundary box to be confirmed as the face boundary box of the face to be detected.
The human face key points comprise a left eye, a right eye, a nose tip, a left mouth corner and a right mouth corner.
The face boundary box of the face to be detected can be determined through the face key points and the face boundary box to be confirmed, wherein the face key points are required to fall into the range of the face boundary box, and the detection precision of the face image can be ensured through mutual comparison of the face key points and the face boundary box to be confirmed.
According to the scheme, through the steps (1) - (3), the detection efficiency can be improved on the premise of ensuring the detection precision of the face image of the snapshot machine.
Optionally, the method further includes:
and a duplication removing step, namely carrying out duplication removing treatment on the face boundary frame of the face to be detected to obtain a target face boundary frame.
For example, the face bounding box may be subjected to deduplication processing by an image deduplication algorithm, so as to obtain a target face bounding box.
Specifically, the deduplication step may be performed by:
and A, carrying out deduplication processing on the face boundary box by using an NMS (Non-Maximum Suppression) algorithm, and eliminating repeated and overlapped boundary boxes to obtain a target face boundary box.
And B, acquiring the corner coordinates of the target face boundary box according to the image size (image center point coordinates) and a preset calculation formula, thereby determining and obtaining the accurate position of the corner of the face boundary box. Taking the bias characteristic as a bias coordinate as an example, the calculation formula is as follows:
in the formula (t)x,ty) Representing offset coordinates, (x, y) representing corner coordinates of a face bounding box, (c)x,cy) Representing the center coordinates of the image.
In the duplication removing step, after the target face boundary frame is obtained, the formula (1) is used for operation, and then the corner point coordinates of the face boundary frame can be determined, so that the calculation amount is reduced, the problems of resource waste and high hardware cost requirement caused by occupation of calculation resources are solved, and the detection efficiency is improved.
The application of the trained full-convolution network detection model in the actual use process is introduced above, namely the application of the trained full-convolution network detection model in the acquisition of the human face bounding box. It can be understood that, before the image to be detected (also referred to as a target image) of the snapshot machine is input into the trained full convolution network detection model, the trained full convolution network detection model is obtained first, that is, the full convolution network detection model is trained first, that is, the full convolution network detection model is applied to a network training stage to obtain the full convolution network detection model.
Referring to fig. 2, an embodiment of the present invention further provides a face detection method applied in a network training phase, where the method includes:
step S202, acquiring training sample data; the training sample data comprises image samples of a plurality of snapshot machines, and the image samples comprise a predetermined face boundary box and pixel-level labels of corner points of the face boundary box;
and step S204, applying the training sample data to train the initial full convolution network detection model to obtain the trained full convolution network detection model.
For step S202, the training sample data includes, but is not limited to, images of a snapshot machine in which a face bounding box and corner coordinates of the face bounding box are marked in advance; the image samples in the training phase may also be referred to as sample images.
It should be explained that in training the full convolution network detection model, the input data for the model is the image of the snapshot machine that has been previously labeled by the training person with the face bounding box (e.g., the solid line portion in fig. 3) and the corner coordinates of the face bounding box (the corner coordinates schematically identifying the upper left corner of the face bounding box in fig. 3), and the output data is the face bounding box on the image. The purpose of training is to obtain a proper weight between input and output, so that the full convolution network detection model can accurately output the face bounding box.
Optionally, step S204 is to apply the training sample data to train the initial full convolution network detection model to obtain a trained full convolution network detection model, and may be implemented by the following steps:
a, obtaining a bias coordinate based on the corner coordinates of the face bounding box and the central coordinates of the image, wherein the bias coordinate is used for representing the bias of the corner coordinates of the face bounding box relative to the central coordinates of the image;
and b, inputting the offset coordinates as image labels into the FCN for training to obtain a trained full convolution network detection model.
In step a, the offset coordinates are used to represent the offset of the corner coordinates of the face bounding box with respect to the center coordinates of the image.
Optionally, the step a is performed by:
and calculating the corner coordinates of the face bounding box and the center coordinates of the image to obtain offset coordinates by substituting the following formula (1):
in the formula (t)x,ty) Representing offset coordinates, (x, y) representing corner coordinates of a face bounding box, (c)x,cy) Representing the center coordinates of the image.
Alternatively to this, the first and second parts may,
the image sample also comprises a predetermined pixel level label of the key point of the human face; the method further comprises the following steps:
and applying the training sample data to train the initial key point detection model to obtain the trained key point detection model.
The above-described key point detection model may be an object detection model based on the FCN network structure.
In other words, the training sample data further comprises face key point coordinates marked in advance on the image of the snapshot machine; the method further comprises the following steps: and inputting the coordinates of the key points of the human face as the image labels into the FCN for training to obtain a key point detection model.
Of course, it should be understood that the face key points may be obtained by using an existing face key point detection algorithm or a face key point recognition algorithm.
Optionally, in order to eliminate an adverse error caused by the heterogeneous sample data, after the training sample data is acquired, the method further includes: the training sample data is normalized so that the pixels of the image in the training sample data are in [0, 1 ].
According to the face detection method provided by the embodiment, the full convolution network detection model is obtained through the training of the steps, the face detection can be directly carried out on the image to be detected of the snapshot machine, namely the image to be detected is input, and the face boundary frame can be output.
For ease of understanding, the complete training process of the full convolution network detection model is described below with reference to fig. 3:
referring to fig. 3, for the network training phase, coordinates (x, y) of each of four corner points of the pre-marked face bounding box and center coordinates (c) of an image directly read from the captured image are usedx,cy) Dividing, calculating to obtain offset coordinates (t) of four corner pointsx,ty) And the angular point offset calculation formula is shown as a formula (1), and the angular point offset calculation formula and 5 pre-marked personal face key points are used as image labels and input into an FCN network structure for training to obtain a full convolution network detection model.
It should be noted that, in fig. 3, the face bounding box is schematically represented by a rectangle, and in other embodiments, the face bounding box may also be represented by a circle, an ellipse, or the like as needed, and the present application is not limited in particular.
Referring to fig. 4, based on the same inventive concept, an embodiment of the present application further provides a face detection apparatus 400, which is applied to a face attendance system, and includes an acquisition module 401 and an output module 402;
the acquisition module 401 is configured to acquire an image of a snapshot machine, where the image includes a face to be detected; the image here refers to an image to be detected.
The output module 402 is configured to apply a pre-trained full convolution network detection model, and determine a face bounding box of the face to be detected according to the image.
Optionally, the output module 402 is configured to apply the pre-trained full convolution network detection model, and output a bias feature according to the image; the bias feature is used for representing the bias of the corner points of the face boundary box of the face to be detected relative to the center of the image; determining the face bounding box based on the biased features.
Optionally, the bias feature comprises a bias coordinate; the offset coordinates are used for representing the offset of the corner point coordinates of the face boundary box of the face to be detected relative to the central coordinates of the image; alternatively, the offset feature comprises an offset distance; the offset distance is used for representing the distance between the boundary line of the face boundary frame of the face to be detected and the center of the image.
Optionally, the output module 402 is configured to apply a pre-trained full convolution network detection model to detect the image, so as to obtain a face bounding box to be confirmed;
determining the face key points of the face to be detected according to the image by applying a pre-trained key point detection model;
and if the face key point is in the face boundary box to be confirmed, determining the face boundary box to be confirmed as the face boundary box of the face to be detected.
Optionally, the apparatus further includes a duplicate removal module 403, configured to perform duplicate removal processing on the face bounding box of the to-be-detected face to obtain a target face bounding box.
The following describes an actual detection process of the face detection apparatus provided in the embodiment of the present invention with reference to fig. 5:
referring to fig. 5, for the trained network structure of the full convolution network detection model, the snapshot image obtained by the snapshot machine is used as the input of the model, the offset of the corner points representing the offset of the corner points of the face bounding box relative to the center of the image is output, the coordinates of 5 key points of the face are output, finally, the repeated overlapped bounding boxes are eliminated by the NMS algorithm, the target face bounding box is obtained, and the accurate position result of the corner points of the target face bounding box is obtained according to the image size and the calculation formula (1).
Referring to fig. 6, the present embodiment further provides another face detection apparatus 600, including:
a plurality of sample blocks 601 for acquiring training sample data; the training sample data comprises image samples of a plurality of snapshot machines, and the image samples comprise a predetermined face boundary box and pixel-level labels of corner points of the face boundary box;
a training module 602, configured to apply the training sample data to train the initial full convolution network detection model, so as to obtain a trained full convolution network detection model.
Optionally, the image sample further includes a predetermined pixel-level annotation of a face key point; a training module 602, configured to apply the training sample data to train an initial keypoint detection model, so as to obtain a trained keypoint detection model.
It should be noted that the coordinates mentioned in this embodiment are pixel coordinates.
Referring to fig. 7, based on the same inventive concept, an embodiment of the invention further provides an electronic device 100, including: a processor 40, a memory 41, a bus 42 and a communication interface 43, wherein the processor 40, the communication interface 43 and the memory 41 are connected through the bus 42; the processor 40 is arranged to execute executable modules, such as computer programs, stored in the memory 41.
The Memory 41 may include a high-speed Random Access Memory (RAM) and may also include a non-volatile Memory (non-volatile Memory), such as at least one disk Memory. The communication connection between the network element of the system and at least one other network element is realized through at least one communication interface 43 (which may be wired or wireless), and the internet, a wide area network, a local network, a metropolitan area network, etc. may be used.
The bus 42 may be an ISA bus, PCI bus, EISA bus, or the like. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, only one double-headed arrow is shown in FIG. 7, but this does not indicate only one bus or one type of bus.
The memory 41 is used for storing a program, the processor 40 executes the program after receiving an execution instruction, and the method executed by the apparatus defined by the flow process disclosed in any of the foregoing embodiments of the present invention may be applied to the processor 40, or implemented by the processor 40.
The processor 40 may be an integrated circuit chip having signal processing capabilities. In implementation, the steps of the above method may be performed by integrated logic circuits of hardware or instructions in the form of software in the processor 40. The Processor 40 may be a general-purpose Processor, and includes a Central Processing Unit (CPU), a Network Processor (NP), and the like; the device can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA) or other Programmable logic device, a discrete Gate or transistor logic device, or a discrete hardware component. The various methods, steps and logic blocks disclosed in the embodiments of the present invention may be implemented or performed. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The steps of the method disclosed in connection with the embodiments of the present invention may be directly implemented by a hardware decoding processor, or implemented by a combination of hardware and software modules in the decoding processor. The software module may be located in ram, flash memory, rom, prom, or eprom, registers, etc. storage media as is well known in the art. The storage medium is located in a memory 41, and the processor 40 reads the information in the memory 41 and completes the steps of the method in combination with the hardware thereof.
The embodiment of the present application further provides a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the steps of the face detection method provided in the foregoing embodiment are executed.
In the embodiments provided in the present application, it should be understood that the disclosed apparatus and method can be implemented in other ways. The apparatus embodiments described above are merely illustrative and, for example, the flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods and computer program products according to various embodiments of the present application. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
In addition, each functional module or unit in each embodiment of the present application may be integrated together to form an independent part, or each module may exist separately, or two or more modules may be integrated to form an independent part.
The functions, if implemented in the form of software functional modules and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present application or portions thereof that substantially contribute to the prior art may be embodied in the form of a software product stored in a storage medium and including instructions for causing a computer device (which may be a smart phone, a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present application. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.
The above description is only for the specific embodiments of the present application, but the scope of the present application is not limited thereto, and any person skilled in the art can easily conceive of the changes or substitutions within the technical scope of the present application, and shall be covered by the scope of the present application.
Claims (10)
1. A face detection method, comprising:
acquiring an image of a snapshot machine, wherein the image comprises a face to be detected;
and determining the face boundary frame of the face to be detected according to the image by applying a pre-trained full convolution network detection model.
2. The method according to claim 1, wherein the determining the face bounding box of the face to be detected according to the image by applying a pre-trained full convolution network detection model comprises:
applying the pre-trained full convolution network detection model and outputting bias characteristics according to the image; the bias feature is used for representing the bias of the corner points of the face boundary box of the face to be detected relative to the center of the image;
determining the face bounding box based on the biased features.
3. The method of claim 2, wherein the bias features comprise bias coordinates; the offset coordinates are used for representing the offset of the corner point coordinates of the face boundary box of the face to be detected relative to the central coordinates of the image;
alternatively, the first and second electrodes may be,
the offset feature comprises an offset distance; the offset distance is used for representing the distance between the boundary line of the face boundary frame of the face to be detected and the center of the image.
4. The method according to claim 1, wherein the determining the face bounding box of the face to be detected according to the image by applying a pre-trained full convolution network detection model comprises:
detecting the image by applying a pre-trained full convolution network detection model to obtain a human face boundary box to be confirmed;
determining the face key points of the face to be detected according to the image by applying a pre-trained key point detection model;
and if the face key point is in the face boundary box to be confirmed, determining the face boundary box to be confirmed as the face boundary box of the face to be detected.
5. The method of claim 1, further comprising:
and carrying out duplication removal processing on the face boundary frame of the face to be detected to obtain a target face boundary frame.
6. A face detection method, comprising:
acquiring training sample data; the training sample data comprises image samples of a plurality of snapshot machines, and the image samples comprise a predetermined face boundary box and pixel-level labels of corner points of the face boundary box;
and applying the training sample data to train the initial full convolution network detection model to obtain the trained full convolution network detection model.
7. The method of claim 6, wherein the image samples further comprise pixel-level labels for predetermined face keypoints; the method further comprises the following steps:
and applying the training sample data to train the initial key point detection model to obtain the trained key point detection model.
8. A face detection apparatus, comprising:
the acquisition module is used for acquiring an image of the snapshot machine, wherein the image comprises a face to be detected;
and the output module is used for applying a pre-trained full convolution network detection model and determining the face boundary frame of the face to be detected according to the image.
9. An electronic device, comprising: a processor and a memory; the processor is connected with the memory;
the memory is used for storing programs;
the processor is configured to execute a program stored in the memory to perform the method of any of claims 1-7.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the method according to any one of claims 1-7.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201911268526.7A CN111046792A (en) | 2019-12-11 | 2019-12-11 | Face detection method and device, electronic equipment and computer readable storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201911268526.7A CN111046792A (en) | 2019-12-11 | 2019-12-11 | Face detection method and device, electronic equipment and computer readable storage medium |
Publications (1)
Publication Number | Publication Date |
---|---|
CN111046792A true CN111046792A (en) | 2020-04-21 |
Family
ID=70235766
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201911268526.7A Pending CN111046792A (en) | 2019-12-11 | 2019-12-11 | Face detection method and device, electronic equipment and computer readable storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111046792A (en) |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104598936B (en) * | 2015-02-28 | 2018-07-27 | 北京畅景立达软件技术有限公司 | The localization method of facial image face key point |
CN109558864A (en) * | 2019-01-16 | 2019-04-02 | 苏州科达科技股份有限公司 | Face critical point detection method, apparatus and storage medium |
CN110276308A (en) * | 2019-06-25 | 2019-09-24 | 上海商汤智能科技有限公司 | Image processing method and device |
CN110309706A (en) * | 2019-05-06 | 2019-10-08 | 深圳市华付信息技术有限公司 | Face critical point detection method, apparatus, computer equipment and storage medium |
CN110490170A (en) * | 2019-08-27 | 2019-11-22 | 浙江中正智能科技有限公司 | A kind of face candidate frame extracting method |
-
2019
- 2019-12-11 CN CN201911268526.7A patent/CN111046792A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104598936B (en) * | 2015-02-28 | 2018-07-27 | 北京畅景立达软件技术有限公司 | The localization method of facial image face key point |
CN109558864A (en) * | 2019-01-16 | 2019-04-02 | 苏州科达科技股份有限公司 | Face critical point detection method, apparatus and storage medium |
CN110309706A (en) * | 2019-05-06 | 2019-10-08 | 深圳市华付信息技术有限公司 | Face critical point detection method, apparatus, computer equipment and storage medium |
CN110276308A (en) * | 2019-06-25 | 2019-09-24 | 上海商汤智能科技有限公司 | Image processing method and device |
CN110490170A (en) * | 2019-08-27 | 2019-11-22 | 浙江中正智能科技有限公司 | A kind of face candidate frame extracting method |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110827247B (en) | Label identification method and device | |
CN110705405B (en) | Target labeling method and device | |
CN108985199B (en) | Detection method and device for commodity taking and placing operation and storage medium | |
CN108875731B (en) | Target identification method, device, system and storage medium | |
CN109002820B (en) | License plate recognition method and device and related equipment | |
CN111382704B (en) | Vehicle line pressing violation judging method and device based on deep learning and storage medium | |
CN109086734B (en) | Method and device for positioning pupil image in human eye image | |
CN108268867B (en) | License plate positioning method and device | |
US9916513B2 (en) | Method for processing image and computer-readable non-transitory recording medium storing program | |
US9600736B2 (en) | Pose detection using depth camera | |
CN104615972B (en) | Intelligent identification method and device for pointer instrument | |
CN112434715B (en) | Target identification method and device based on artificial intelligence and storage medium | |
CN108875500B (en) | Pedestrian re-identification method, device and system and storage medium | |
CN112308046A (en) | Method, device, server and readable storage medium for positioning text region of image | |
CN111680680A (en) | Object code positioning method and device, electronic equipment and storage medium | |
CN111199198A (en) | Image target positioning method, image target positioning device and mobile robot | |
CN111046831B (en) | Poultry identification method, device and server | |
CN109858363B (en) | Dog nose print feature point detection method, device, system and storage medium | |
CN111985438A (en) | Static face processing method, device and equipment | |
CN109816709B (en) | Monocular camera-based depth estimation method, device and equipment | |
CN111046792A (en) | Face detection method and device, electronic equipment and computer readable storage medium | |
CN110610178A (en) | Image recognition method, device, terminal and computer readable storage medium | |
CN113903014B (en) | Lane line prediction method and device, electronic device and computer-readable storage medium | |
CN110782439B (en) | Method and device for auxiliary detection of image annotation quality | |
KR20210087494A (en) | Human body orientation detection method, apparatus, electronic device and computer storage medium |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
TA01 | Transfer of patent application right | ||
TA01 | Transfer of patent application right |
Effective date of registration: 20230612 Address after: 314506 room 116, building 4, No. 288, development avenue, Tongxiang Economic Development Zone, Tongxiang City, Jiaxing City, Zhejiang Province Applicant after: Shengjing Intelligent Technology (Jiaxing) Co.,Ltd. Address before: 102200 5th floor, building 6, No.8 Beiqing Road, Changping District, Beijing Applicant before: SANY HEAVY INDUSTRY Co.,Ltd. |
|
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200421 |