WO2020034763A1 - 手势识别方法、手势处理方法及装置 - Google Patents

手势识别方法、手势处理方法及装置 Download PDF

Info

Publication number
WO2020034763A1
WO2020034763A1 PCT/CN2019/092559 CN2019092559W WO2020034763A1 WO 2020034763 A1 WO2020034763 A1 WO 2020034763A1 CN 2019092559 W CN2019092559 W CN 2019092559W WO 2020034763 A1 WO2020034763 A1 WO 2020034763A1
Authority
WO
WIPO (PCT)
Prior art keywords
state
finger
hand
gesture
image
Prior art date
Application number
PCT/CN2019/092559
Other languages
English (en)
French (fr)
Inventor
杜天元
钱晨
Original Assignee
北京市商汤科技开发有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 北京市商汤科技开发有限公司 filed Critical 北京市商汤科技开发有限公司
Priority to KR1020217007082A priority Critical patent/KR20210040435A/ko
Priority to JP2021506277A priority patent/JP7266667B2/ja
Priority to SG11202101142PA priority patent/SG11202101142PA/en
Publication of WO2020034763A1 publication Critical patent/WO2020034763A1/zh
Priority to US17/166,238 priority patent/US20210158031A1/en

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G06F18/2155—Generating training patterns; Bootstrap methods, e.g. bagging or boosting characterised by the incorporation of unlabelled data, e.g. multiple instance learning [MIL], semi-supervised techniques using expectation-maximisation [EM] or naïve labelling
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
    • G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
    • G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
    • G06F3/017—Gesture based interaction, e.g. based on a set of recognized hand gestures
    • 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
    • G06N3/084—Backpropagation, e.g. using gradient descent
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
    • G06V10/462—Salient features, e.g. scale invariant feature transforms [SIFT]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
    • 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
    • G06V40/107—Static hand or arm
    • 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
    • G06V40/20—Movements or behaviour, e.g. gesture recognition
    • G06V40/28—Recognition of hand or arm movements, e.g. recognition of deaf sign language

Definitions

  • the present disclosure relates to the field of image processing technologies, and in particular, to a gesture recognition method, a gesture processing method, and a device.
  • Non-contact human-computer interaction scenarios are increasingly used in life. Users can use different gestures to easily express different human-computer interaction instructions.
  • This disclosure proposes a technical solution for gesture recognition.
  • a gesture recognition method including: detecting a state of a finger included in a hand in an image; determining a state vector of the hand according to the state of the finger; and according to the state of the hand The vector determines the gesture of the hand.
  • a gesture processing method includes: acquiring an image; using the gesture recognition method described above to recognize a gesture of a hand included in the image; and performing a control operation corresponding to a recognition result of the gesture.
  • a gesture recognition device includes: a state detection module for detecting a state of a finger included in a hand in an image; and a state vector acquisition module for detecting a state of the finger according to the state Determining a state vector of the hand; a gesture determination module, configured to determine a gesture of the hand according to the state vector of the hand.
  • a gesture processing device includes: an image acquisition module for acquiring an image; and a gesture acquisition module for using the gesture recognition device to identify a gesture of a hand included in the image ; An operation execution module, configured to execute a control operation corresponding to a recognition result of a gesture.
  • an electronic device including: a processor; a memory for storing processor-executable instructions; wherein, the processor implements the foregoing gesture recognition method by calling the executable instructions and / Or gesture processing methods.
  • a computer-readable storage medium having computer program instructions stored thereon, the computer program instructions implementing the above-mentioned gesture recognition method and / or gesture processing method when executed by a processor.
  • a computer program includes computer-readable code, and when the computer-readable code runs in an electronic device, a processor in the electronic device performs the gesture recognition described above. Method and / or gesture processing method.
  • a state of a hand included in an image is detected, a state vector of the hand is determined according to the state of the finger, and a gesture of the hand is determined according to the determined state vector of the hand.
  • a state vector is determined according to a state of each finger, and a gesture is determined according to the state vector. The recognition efficiency is high and the versatility is strong.
  • FIG. 1 shows a flowchart of a gesture recognition method according to an embodiment of the present disclosure
  • FIG. 2 is a schematic diagram showing a state of a finger in a gesture recognition method according to an embodiment of the present disclosure
  • FIG. 3 shows a flowchart of a gesture recognition method according to an embodiment of the present disclosure
  • FIG. 4 shows a flowchart of a gesture recognition method according to an embodiment of the present disclosure
  • FIG. 5 illustrates a flowchart of a gesture recognition method according to an embodiment of the present disclosure
  • FIG. 6 illustrates a data processing flowchart of a neural network in a gesture recognition method according to an embodiment of the present disclosure
  • FIG. 7 illustrates a flowchart of a gesture recognition method according to an embodiment of the present disclosure
  • FIG. 8 illustrates a flowchart of a gesture processing method according to an embodiment of the present disclosure
  • FIG. 9 illustrates a block diagram of a gesture recognition apparatus according to an embodiment of the present disclosure.
  • FIG. 10 illustrates a block diagram of gesture processing according to an embodiment of the present disclosure
  • Fig. 11 is a block diagram of an electronic device according to an exemplary embodiment
  • Fig. 12 is a block diagram showing an electronic device according to an exemplary embodiment.
  • exemplary means “serving as an example, embodiment, or illustration.” Any embodiment described herein as “exemplary” is not necessarily to be construed as superior to or better than other embodiments.
  • FIG. 1 illustrates a flowchart of a gesture recognition method according to an embodiment of the present disclosure.
  • the gesture recognition method may be executed by an electronic device such as a terminal device or a server.
  • the terminal device may be a user equipment (UE), a mobile device, a user terminal, a terminal, a cellular phone, a cordless phone, or personal digital processing (Personal).
  • Digital Assistant (PDA) handheld devices, computing devices, in-vehicle devices, wearable devices, etc.
  • the gesture recognition method may be implemented by a processor invoking computer-readable instructions stored in a memory.
  • the method includes:
  • step S10 a state of a finger included in the hand in the image is detected.
  • the image may be a static image or a frame image in a video stream.
  • the state of each finger in the hand can be obtained in the image using an image recognition method. You can get the status of five fingers in the hand, or you can get the status of the specified multiple or single fingers, for example, you can get only the status of the index finger.
  • the state of the finger indicates whether the finger is extended and / or extended to the palm root of the hand.
  • each finger is in a non-extended state with respect to the root of the palm.
  • the state of the finger can be further divided according to the position of the finger relative to the palm or the degree of bending of the finger.
  • the state of a finger can be divided into two states: non-extended state or extended state, or three states: non-extended state, semi-extended state, and extended state. Out state, half-extended state, and bent state.
  • the state of the finger includes one or more of the following: an extended state, a non-extended state, a semi-extended state, and a bent state.
  • the state of each finger can be: non-extended state, semi-extended state, and bent state in the process of the hand from the fist to the five fingers open. , Extended state. You can also divide the status of different fingers into different status levels according to your needs. The present disclosure does not limit the classification manner, number, and order of use of the states of each finger.
  • FIG. 2 is a schematic diagram of a finger state in a gesture recognition method according to an embodiment of the present disclosure.
  • the state of the thumb is a non-extended state
  • the state of the index finger is an extended state
  • the state of the middle finger is In the extended state
  • the state of the ring finger is a non-extended state
  • the state of the little finger is a non-extended state.
  • the state of all five fingers can be acquired in the image, or only the state of a specified finger (such as the index finger and the middle finger) can be acquired.
  • Step S20 Determine a state vector of the hand according to the state of the finger.
  • the determining a state vector of the hand according to a state of the finger includes: determining a state value of the finger according to a state of the finger, wherein the fingers corresponding to different states are The state values are different; the state vector of the hand is determined according to the state value of the finger.
  • corresponding state values may be determined for the states of different fingers, and a corresponding relationship between the states of the fingers and the state values may be established.
  • the state value of the finger can be one or any combination of numbers, letters, or symbols.
  • the state value of the finger can be determined according to the obtained state of the finger and the established correspondence, and then the state vector of the hand can be obtained by using the state value of the finger.
  • the state vector of the hand can include various forms such as array, list, or matrix.
  • the state values of the fingers can be combined according to a set finger order to obtain a state vector of the hand.
  • the state vector of the hand can be obtained based on the state values of five fingers.
  • the state vectors of the five fingers can be combined in the order of thumb, forefinger, middle finger, ring finger, little finger to obtain the state vector of the hand.
  • the state vector of the hand can also be obtained by combining the state values of the fingers in any other set order.
  • a state value A may be used to indicate a non-extended state
  • a state value B may be used to indicate an extended state.
  • the state value of the thumb is A
  • the state value of the index finger is B
  • the state value of the middle finger is B
  • the state value of the ring finger is A
  • the state value of the little finger is A.
  • the state vector of the hand can be (A, B, B, A, A).
  • Step S30 Determine a gesture of the hand according to a state vector of the hand.
  • the state of each finger in the hand can be used to determine the hand gesture.
  • the different states of the fingers can be determined according to the requirements
  • the state vectors of the hands can be determined according to the different states of the fingers
  • the hand gestures can be determined according to the state vectors of the hands.
  • the recognition process of the finger state is convenient and reliable, which makes the determination process of the gesture more convenient and reliable.
  • the correspondence between the state vector of the hand and the gesture can be established. By adjusting the correspondence between the state vector and the gesture, the gesture can be determined more flexibly based on the state vector, making the determination process of the gesture more flexible and able to adapt to different applications. surroundings.
  • the state vector 1 of the hand corresponds to gesture 1
  • the state vector 2 of the hand corresponds to gesture 2
  • the state vector 3 of the hand corresponds to gesture 3.
  • the correspondence between the state vector of the hand and the gesture can be determined according to requirements.
  • the state vectors of one hand may correspond to one gesture, or the state vectors of multiple hands may correspond to one gesture.
  • the state vector of the hand is (A, B, B, A, A), and the correspondence between the state vector of the hand and the gesture
  • the gesture corresponding to the state vector (A, B, B, A, A) can be "Number 2" or "Victory”.
  • a state of a finger included in a hand in an image is detected, a state vector of the hand is determined according to the state of the finger, and a hand gesture is determined according to the determined state vector of the hand.
  • a state vector is determined according to a state of each finger, and a gesture is determined according to the state vector. The recognition efficiency is high and the versatility is strong.
  • This embodiment has high recognition efficiency for recognizing the state of each finger in the image, so that this embodiment has high recognition efficiency for recognizing gestures. And in this embodiment, the corresponding relationship between the state of the finger and the gesture can be arbitrarily adjusted according to the needs, and different gestures defined under different needs can be identified according to the same image, so that the determined gestures are highly versatile.
  • the state of the finger includes an extended state or a non-extended state
  • determining the state vector of the hand according to the state of the finger includes:
  • a state vector of the hand is determined according to a state value of the finger.
  • first state value and the second state value may be two values representing opposite meanings, for example, the first state value may be valid, and the second state value may be invalid.
  • the first state value and the second state value may also be two numbers with different values.
  • the first state value may be 1 and the second state value may be 0.
  • the state value of the thumb is 0, the state value of the index finger is 1, the state value of the middle finger is 1, the state value of the ring finger is 0, the state value of the little finger is 0, and the state vector of the hand Is (0,1,1,0,0).
  • the first state value and the second state value may be used to determine the state vector of the hand.
  • the states of the fingers of the hand can be expressed simply and intuitively.
  • FIG. 3 shows a flowchart of a gesture recognition method according to an embodiment of the present disclosure. As shown in FIG. 3, the method further includes:
  • Step S40 Detect position information of a finger included in a hand in the image.
  • the position information of the finger may include position information of the finger in the image.
  • the position information of the finger may include coordinate position information of a pixel of the finger in the image.
  • the image may also be divided into grids, and the position information of the grid where the pixels of the finger are located is determined as the position information of the finger.
  • the position information of the grid may include the number of the grid.
  • the position information of the finger may also include position information of the finger relative to the target object in the image.
  • the picture in the image is a person playing the piano
  • the position information of the finger in the image may include the position information of the finger relative to the key.
  • the distance between finger 1 and the keys is 0, and the distance between finger 2 and the keys is 3 cm.
  • the position information of the finger may include one-dimensional or multi-dimensional position information. According to the position information of the fingers, the relative position relationship between the fingers can be obtained.
  • Step S50 Determine a position vector of the hand according to the position information of the finger.
  • the position vector of the hand can be obtained by combining the position information of different fingers according to the set order of the fingers.
  • the hand position vector can include various forms such as array, list, or matrix.
  • Step S30 includes:
  • Step S31 Determine a gesture of the hand according to a state vector of the hand and a position vector of the hand.
  • the state of the finger in the hand can be obtained according to the state vector of the hand, and a more accurate gesture can be determined in combination with the position of the finger in the position vector of the hand.
  • the state vector of the hand is (0, 1, 1, 0, 0) and the position vector is (L1, L2, L3, L4, L5). If only the state vector of the hand is used, it can be determined that the state of the index finger and the middle finger in the hand is the extended state, and the other fingers are in the non-extended state.
  • the state vector of the hand can determine that the hand gesture is "Number 2" or " victory".
  • the hand gesture can be "number 2" or "victory” . If it can be determined that the forefinger and the middle finger are extended and close together according to the state vector of the hand and the position vector of the hand (not shown in the figure), the gesture of the hand can be "number 2", not " victory".
  • the state vector of the hand and the position vector of the hand can be combined to obtain a combination vector, and then the corresponding relationship between the combination vector and the gesture is established.
  • Different combination vectors composed of the same state vector and different position vectors can correspond to different gestures, and can also correspond to the same gesture.
  • a hand gesture may be determined according to a state vector and a position vector of the hand. By combining the hand's position vector and state vector, a more accurate gesture can be obtained.
  • FIG. 4 shows a flowchart of a gesture recognition method according to an embodiment of the present disclosure. As shown in FIG. 4, step S40 in the method includes:
  • Step S41 Detect key points of a finger included in the hand in the image, and obtain position information of the key points of the finger.
  • the key point includes a fingertip and / or a knuckle, wherein the knuckle can include a metacarpophalangeal joint or an interphalangeal joint.
  • the position of the finger can be accurately represented by using the position of the fingertip and / or knuckle of the finger. For example, in the image shown in FIG.
  • the key point of the finger is the fingertip
  • the position information of the fingertip of each finger can be determined as: thumb (X 1 , Y 1 ), index finger (X 2 , Y 2 ), middle finger (X 3 , Y 3 ), ring finger (X 4 , Y 4 ), pinky finger (X 5 , Y 5 ), where the coordinate points of the fingertips of the thumb, ring finger and pinky are close.
  • Step S50 includes:
  • Step S51 Determine a position vector of the hand according to the position information of a key point of the finger.
  • the position vector of the hand may be (X 1 , Y 1 , X 2 , Y 2 , X 3 , Y 3 , X 4 , Y 4 , X 5 , Y 5 ).
  • the hand's state vector (0, 1 , 1 , 0, 0) and the hand's position vector (X 1 , Y 1 , X 2 , Y 2 , X 3 , Y 3 , X 4 , Y 4 , X 5 , Y 5 )
  • the position vector of the hand can be obtained according to the position information of the key points of the fingers of the hand. This makes the process of determining the position vector of the hand easier.
  • step S41 includes: detecting a key point of a finger other than a non-extended state included in the hand in the image, and obtaining position information of the key point.
  • a key point on the finger other than the non-extended state can be determined in the image, and the key can be obtained.
  • Point location information The position coordinates of the key point of the finger in a non-extended state can be determined as a coordinate value that does not exist in the image.
  • the upper edge of the image can be positive in the X axis, and the left edge can be positive in the Y axis.
  • Invalid coordinates can be (-1, -1).
  • the upper edge of the image can be positive in the X axis
  • the left edge can be positive in the Y axis
  • the key point of the finger is the fingertip.
  • the position information of the fingertip in the image is obtained as follows: thumb (-1, -1), index finger (X 2 , Y 2 ), middle finger (X 3 , Y 3 ), ring finger (- 1, -1), little finger (-1, -1).
  • the position vector of the hand may be (-1, -1, X 2 , Y 2 , X 3 , Y 3 , -1, -1, -1, -1). You can also zero the position coordinates of the key points of the fingers in the non-extended state.
  • the position vector of the hand can be obtained according to the position information of the key point of the finger other than the non-extended state. This makes the process of determining the position vector of the hand more efficient.
  • FIG. 5 shows a flowchart of a gesture recognition method according to an embodiment of the present disclosure. As shown in FIG. 5, step S10 in the method includes:
  • step S11 the image is input to a neural network, and the state of the fingers included in the hand in the image is detected via the neural network.
  • the neural network is a mathematical or computational model that mimics the structure and function of a biological neural network.
  • a neural network may include an input layer, an intermediate layer, and an output layer.
  • the input layer is responsible for receiving input data from the outside and passing the input data to the intermediate layer.
  • the middle layer is responsible for information exchange. According to the requirement of information change ability, the middle layer can be designed as a single hidden layer or multiple hidden layers.
  • the intermediate layer passes the output result to the output layer for further processing, and then obtains the output result of the neural network.
  • the input layer, the middle layer, and the output layer can all include several neurons, and each neuron can be connected with a variable weight.
  • the neural network through the repeated learning and training of known information, by gradually adjusting the method of changing the connection weight of neurons, to achieve the purpose of establishing a model of the relationship between analog input and output.
  • the trained neural network can use the simulated relationship model between input and output, detect input information, and give output information corresponding to the input information.
  • a neural network may include a convolutional layer, a pooling layer, a fully connected layer, and the like.
  • a neural network can be used to extract features in the image and determine the state of the fingers in the image based on the extracted features.
  • the powerful processing capability of the neural network can be used to quickly and accurately determine the state of the fingers included in the hand in the image.
  • the neural network includes multiple state branch networks
  • step S11 includes: detecting states of different fingers included in the hand in the image through different state branch networks of the neural network.
  • five state branch networks can be set in the neural network, and each state branch network is used to obtain a state of a finger in an image.
  • FIG. 6 illustrates a data processing flowchart of a neural network in a gesture recognition method according to an embodiment of the present disclosure.
  • the neural network may include a convolutional layer and a fully connected layer.
  • the convolution layer may include a first convolution layer, a second convolution layer, a third convolution layer, and a fourth convolution layer.
  • the first convolutional layer may include one convolutional layer “conv1_1”, and the second to fourth convolutional layers may have two convolutional layers, for example, “conv2_1” to “conv4_2”.
  • the first convolution layer, the second convolution layer, the third convolution layer, and the fourth convolution layer may be used to extract features in the image.
  • the fully-connected layer may include a first fully-connected layer "ip1_fingers", a second fully-connected layer “ip2_fingers”, and a third fully-connected layer “ip3_fingers”.
  • the first fully-connected layer, the second fully-connected layer, and the third fully-connected layer may be used to determine a state of a finger and obtain a state vector of the finger.
  • "ip3_fingers” can be divided into five state branch networks: first state branch network (loss_littlefinger), second state branch network (loss_ringfinger), third state branch network (loss_middlefinger), and fourth state branch network (loss_forefinger). ) And a fifth state branch network (loss_thumb).
  • Each state branch network corresponds to a finger, and each state branch network can be trained separately.
  • the fully-connected layer further includes a location branch network
  • step S40 may include:
  • the neural network further includes a location branch network
  • the location branch network may include a fifth fully-connected layer “ip1_points”, a sixth fully-connected layer “ip2_points”, and a seventh fully-connected layer “ip3_points”.
  • the fifth fully-connected layer, the sixth fully-connected layer, and the seventh fully-connected layer are used to obtain position information of a finger.
  • the convolutional layer in FIG. 6 may further include an activation function (relu_conv), a pooling layer (pool), a loss function (loss), and the like, which are not described again.
  • the position branch network can be used to determine the position information of the finger in the image
  • the position branch network can be used to determine the position information of the finger in the image. According to the state branch network and the position branch network, the state information and position information of the finger can be obtained quickly and accurately in the image.
  • the neural network is obtained by training with a sample image with annotation information in advance, and the annotation information includes first annotation information indicating a state of the finger, and / or, indicating the
  • the second labeling information is position information of a finger or position information of a key point.
  • the label information of the sample image may include first label information indicating a state of a finger.
  • the state of the detected finger can be compared with the first label information to determine the loss of the prediction result of the gesture.
  • the label information of the sample image may include second label information indicating position information of a finger or position information of a key point.
  • the position of each finger or the position of a key point can be obtained according to the second label information, and the state of each finger can be determined according to the position of each finger or the position of a key point.
  • the state of the detected finger can be compared with the state of the finger determined according to the second label information to determine the loss of the prediction result of the gesture.
  • the label information of the sample image may include first label information and second label information.
  • the state of the detected finger can be compared with the first label information, and the detected position information can be compared with the second label to determine the loss of the gesture prediction result.
  • the first labeling information includes a state vector composed of a first identification value indicating a status of each finger
  • the second labeling information includes a position information or a key point identifying each finger.
  • the second labeling information of a finger in a non-extended state is not labeled.
  • An invalid second identification value can be set for a non-extended finger, such as (-1, -1).
  • the identification value in the first labeling information may be determined according to the number of states of the fingers. For example, if the state of the finger is a non-extended state or an extended state, the first identification value in the first labeling information may include 0 (non-extended state) or 1 (extended state). The state of the finger is a non-extended state, a semi-extended state, a bent state, and an extended state. The first identification value may include 0 (non-extended state), 1 (semi-extended state), 2 (bent state), 3 (Extended state). The first label information of the hand can be obtained according to the first identification value of each finger, such as (0,1,1,0,0)
  • an image coordinate system may be established for the sample image, and a second identification value in the second labeling information is determined according to the established image coordinate system.
  • the second identification can be worth a second finger of each hand section of the label information, for example (-1, -1, X 2, Y 2, X 3, Y 3, -1, -1, -1, -1).
  • FIG. 7 shows a flowchart of a gesture recognition method according to an embodiment of the present disclosure.
  • the training steps of the neural network include:
  • step S1 a sample image of a hand is input to a neural network to obtain a state of a finger in the hand.
  • inputting the sample image of the hand into the neural network to obtain the state of the finger in the hand includes: inputting the sample image of the hand into the neural network to obtain the state and position information of the finger in the hand.
  • the sample image of the hand may be an image labeled with a state and position information of a finger.
  • the sample image of the hand can be input into a neural network, and the neural network extracts features in the image, and determines the state and position information of the finger according to the extracted features.
  • a hand gesture may be determined according to the determined state and position information of the finger.
  • Step S2 Determine the position weight of the finger according to the state of the finger.
  • different position weights can be set for different states of the finger. For example, a higher position weight may be set for a finger in the extended state, and a lower position weight may be set for a finger in the non-extended state.
  • determining the position weight of the finger according to the state of the finger includes: when the state of the finger is a non-extended state, determining that the position weight of the finger is zero weight.
  • the position weight of the finger when the state of the finger is the extended state, the position weight of the finger may be determined to be a non-zero weight; when the state of the finger is the non-extended state, the position of the finger may be determined The weight is zero.
  • the position information of the key points of the finger in the extended state can be obtained, and the position information of the hand is obtained according to the position information of the key points of the finger in the extended state, and then based on the hand
  • the position information and status information of the part determine the gesture of the hand.
  • the state vector of the hand is (0,1,1,0,0)
  • the position vector of the hand is (-1, -1, X 2 , Y 2 , X 3 , Y 3 , -1, -1, -1).
  • the position weight is set to 1 for the index finger and the middle finger, and the position weight is set to 0 for the remaining three fingers.
  • the position weight of the hand is (0,0,1,1,1,1,0 , 0, 0, 0).
  • the state vector of the hand is (0, 1, 0, 0, 0)
  • the position vector of the hand with the fingertip as the key point is (-1, -1, X 2 , Y 2 , -1, -1, -1, -1, -1, -1) and the position weight is (0, 0, 1, 1, 0, 0, 0, 0,0,0).
  • the state vector of the hand is (0, 0, 0, 0, 0)
  • the position vector of the hand with the fingertip as the key point is (-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1), and the position weight is (0, 0, 0, 0, 0, 0, 0, 0).
  • the state vector of the hand is (0,0,1,1,1)
  • the position vector of the hand with the fingertip as the key point Is (-1, -1, -1, -1, X 3 , Y 3 , X 4 , Y 4 , X 5 , Y 5 )
  • the position weight is (0, 0 , 0 , 0 , 1, 1, 1, 1 , 1,1,1).
  • Step S3 Determine the loss of the gesture prediction result of the neural network according to the state of the finger and the position weight.
  • determining the loss of the gesture prediction result of the neural network according to the state of the finger and the position weight includes: according to the state of the finger, the position information, and the position The weight determines the loss of the gesture prediction result of the neural network.
  • Step S4 Back-propagating the loss to the neural network to adjust network parameters of the neural network.
  • the value of the position vector of the non-extended finger in the position vector of the finger will affect the loss function of the back propagation of the neural network. Calculation results. For example, if the neural network is back-propagated based only on the state and position information of the fingers, in the image shown in Figure 2, the state vector of the hand is (0, 1, 1, 0, 0), and the hand The position vector of the part is (-1, -1, X 2 , Y 2 , X 3 , Y 3 , -1, -1, -1, -1).
  • the thumb, ring finger and The position vector of the little finger will approach -1, causing the back propagation of the neural network to be biased, and the recognition result of the trained neural network will be inaccurate. If the position weights of hands (0,0,1,1,1,1,0,0,0,0) are combined, the back-propagation of the neural network will not reverse the position vectors of the thumb, ring finger and little finger. For the calculation of the forward propagation, the recognition result of the trained neural network is accurate.
  • the back propagation of the neural network is performed according to the state, position information, and position weight of the finger, which can reduce the adverse effect of the position coordinate value on the back propagation in the position information of the finger, so that the training Neural network is more accurate.
  • FIG. 8 illustrates a flowchart of a gesture processing method according to an embodiment of the present disclosure.
  • the gesture processing method may be executed by an electronic device such as a terminal device or a server, where the terminal device may be a user equipment (UE), a mobile device, a user terminal, a terminal, a cellular phone, a cordless phone, or a personal digital processing (Personal Digital Assistant (PDA), handheld devices, computing devices, in-vehicle devices, wearable devices, etc.
  • the gesture processing method may be implemented by a processor invoking computer-readable instructions stored in a memory.
  • the method includes:
  • step S60 an image is acquired.
  • Step S70 Recognize a gesture of a hand included in the image by using any of the gesture recognition methods described above.
  • Step S80 Perform a control operation corresponding to the recognition result of the gesture.
  • the required image can be captured by the shooting device, and the image can also be directly received through various types of receiving methods.
  • a gesture of a hand included in an image may be recognized in the acquired image.
  • the corresponding control operation can be performed according to the gesture recognized in the image.
  • step S80 includes: obtaining a control instruction corresponding to a recognition result of the gesture according to a predetermined mapping relationship between the gesture and the control instruction; and controlling the electronic device to perform a corresponding operation according to the control instruction.
  • a mapping relationship between a gesture and a control instruction may be established according to requirements. For example, a control command of "forward" for gesture 1 and a control command of "stop” for gesture 2 may be set. After the hand gesture is determined in the image, the control instruction corresponding to the gesture is determined according to the gesture and the established mapping relationship.
  • electronic devices configured on devices such as robots, mechanical devices, and vehicles may be controlled to implement automatic control of the devices such as robots, mechanical devices, and vehicles.
  • a gesture recognition method in an embodiment of the present disclosure is used to recognize a gesture in the captured image, and a control instruction is determined according to the gesture, so as to finally realize automatic robot operation control.
  • the present disclosure does not limit the type of electronic device controlled by the control instruction.
  • control instruction may be determined according to a gesture, and a rich control instruction may be determined for a gesture in an image by establishing a mapping relationship between the gesture and the control instruction as required.
  • Electronic equipment can be controlled by control instructions to achieve the purpose of controlling various devices such as vehicles.
  • step S80 includes: determining a special effect corresponding to a recognition result of a gesture according to a predetermined mapping relationship between a gesture and a special effect; and drawing the image in a computer drawing manner on the image. Special effects.
  • a mapping relationship between gestures and special effects can be established.
  • Special effects can be used to emphasize the content of gestures, or to enhance the performance of gestures. For example, when the gesture is recognized as "victory”, special effects such as fireworks can be made.
  • Special effects may be drawn in a computer drawing manner, and the completed effects are displayed together with the content of the image.
  • Special effects can include two-dimensional sticker special effects, two-dimensional image special effects, three-dimensional special effects, example special effects, partial image deformation special effects, and the like. This disclosure does not limit the content, type, and implementation of special effects.
  • drawing the special effect in a computer drawing manner on the image includes:
  • the special effect is drawn in a computer drawing manner.
  • additional information such as text, symbols, or images may be added to the image according to the position information of the hand.
  • the additional information may include one or any combination of the following information: text, images, symbols, letters, numbers. For example, you can add symbols such as "exclamation point” or image information such as "lightning" to the tip of your finger to add information that the editor needs to express or emphasize, and enrich the expression ability of the image.
  • a special effect corresponding to the gesture may be determined according to the gesture, and the performance of the image is increased by adding the special effect to the image.
  • FIG. 9 shows a block diagram of a gesture recognition device according to an embodiment of the present disclosure. As shown in FIG. 9, the gesture recognition device includes:
  • a state detection module 10 configured to detect a state of a finger included in a hand in an image
  • a state vector acquisition module 20 configured to determine a state vector of the hand according to a state of the finger
  • the gesture determining module 30 is configured to determine a gesture of the hand according to a state vector of the hand.
  • a state of a finger included in a hand in an image is detected, a state vector of the hand is determined according to the state of the finger, and a hand gesture is determined according to the determined state vector of the hand.
  • a state vector is determined according to a state of each finger, and a gesture is determined according to the state vector. The recognition efficiency is high and the versatility is strong.
  • the state of the finger indicates whether the finger is extended and / or extended to the palm root of the hand.
  • each finger is in a non-extended state with respect to the root of the palm.
  • the state of the finger can be further divided according to the position of the finger relative to the palm or the degree of bending of the finger.
  • the state of a finger can be divided into two states: a non-extended state or an extended state, or a non-extended state, a semi-extended state, or an extended state. Out state, half-extended state, and bent state.
  • the state vector acquisition module includes: a state value acquisition submodule, configured to determine the state value of the finger according to the state of the finger, where the state values of the finger corresponding to different states Different; a first state vector acquisition submodule is configured to determine a state vector of the hand according to a state value of the finger.
  • corresponding state values may be determined for the states of different fingers, and a corresponding relationship between the states of the fingers and the state values may be established.
  • the state value of the finger can be one or any combination of numbers, letters, or symbols.
  • the state value of the finger can be determined according to the obtained state of the finger and the established correspondence, and then the state vector of the hand can be obtained by using the state value of the finger.
  • the state vector of the hand can include various forms such as array, list, or matrix.
  • the state of the finger includes one or more of the following: an extended state, a non-extended state, a semi-extended state, and a bent state.
  • the state of each finger can be: non-extended state, semi-extended state, and bent state in the process of the hand from the fist to the five fingers open. , Extended state. You can also divide the status of different fingers into different status levels according to your needs. The present disclosure does not limit the classification manner, number, and order of use of the states of each finger.
  • the device further includes: a position information acquisition module, configured to detect position information of a finger included in a hand in the image; a position vector acquisition module, configured to detect the position information of the finger according to the position information of the finger; To determine a position vector of the hand;
  • the gesture determination module includes a first gesture determination sub-module for determining a gesture of the hand according to a state vector of the hand and a position vector of the hand.
  • a hand gesture may be determined according to a state vector and a position vector of the hand. By combining the hand's position vector and state vector, a more accurate gesture can be obtained.
  • the position information acquisition module includes: a key point detection sub-module for detecting a key point of a finger included in the hand in the image, and obtaining a key point of the finger. location information;
  • the position vector acquisition module includes a first position vector acquisition sub-module for determining a position vector of the hand according to position information of a key point of the finger.
  • the position vector of the hand can be obtained according to the position information of the key points of the fingers of the hand. This makes the process of determining the position vector of the hand easier.
  • the keypoint detection submodule is configured to detect keypoints of a finger other than a non-extended state included in the hand in the image, and obtain the keypoints. location information.
  • the position vector of the hand can be obtained according to the position information of the key point of the finger other than the non-extended state. This makes the process of determining the position vector of the hand more efficient.
  • the key points include fingertips and / or knuckles.
  • the knuckle joint may include a metacarpophalangeal joint or an interphalangeal joint.
  • the position of the finger can be accurately represented by using the position of the fingertip and / or knuckle of the finger.
  • the state detection module includes: a first state detection sub-module, configured to input the image into a neural network, and detect the finger included in the hand in the image through the neural network. status.
  • the powerful processing capability of the neural network can be used to quickly and accurately determine the state of the fingers included in the hand in the image.
  • the neural network includes a plurality of state branch networks, and the first state detection sub-module is configured to separately detect hands in the image through different state branch networks of the neural network. Status of different fingers included.
  • five state branch networks can be set in the neural network, and each state branch network is used to obtain a state of a finger in an image.
  • the neural network further includes a location branch network
  • the location information acquisition module includes: a first location information acquisition submodule, configured to detect a location through the location branch network of the neural network. Position information of a finger included in the hand in the image.
  • the position branch network can be used to determine the position information of the finger in the image
  • the position branch network can be used to determine the position information of the finger in the image. According to the state branch network and the position branch network, the state information and position information of the finger can be obtained quickly and accurately in the image.
  • the neural network is obtained by training with a sample image with annotation information in advance, and the annotation information includes first annotation information indicating a state of the finger, and / or, indicating the
  • the second labeling information is position information of a finger or position information of a key point.
  • the second labeling information of a finger in a non-extended state is not labeled.
  • An invalid second identification value may be set for a finger in a non-extended state.
  • the first labeling information includes a state vector composed of a first identification value indicating a status of each finger
  • the second labeling information includes a position information or a key point identifying each finger.
  • the neural network includes a training module
  • the training module includes: a state acquisition sub-module for inputting a sample image of a hand into the neural network to obtain a state of a finger in the hand; and determining a position weight A sub-module for determining a position weight of a finger according to a state of the finger; a loss determination sub-module for determining a loss of a gesture prediction result of the neural network according to the state of the finger and the position weight; A propagation sub-module configured to back-propagate the loss to the neural network to adjust network parameters of the neural network.
  • the state acquisition submodule is configured to: input a sample image of a hand into a neural network to obtain state and position information of a finger in the hand; and the loss determination submodule is configured to: The state of the finger, the position information, and the position weight determine a loss of a gesture prediction result of the neural network.
  • the back propagation of the neural network is performed according to the state, position information, and position weight of the finger, which can reduce the adverse effect of the position coordinate value on the back propagation in the position information of the finger, so that the training Neural network is more accurate.
  • the position weight determining sub-module is configured to: when the state of a finger is a non-extended state, determine that the position weight of the finger is zero weight.
  • the position weight of the finger when the state of the finger is the extended state, the position weight of the finger may be determined to be a non-zero weight; when the state of the finger is the non-extended state, the position of the finger may be determined The weight is zero.
  • FIG. 10 shows a block diagram of a gesture processing apparatus according to an embodiment of the present disclosure. As shown in FIG. 10, the apparatus includes:
  • a gesture acquisition module 2 for recognizing a gesture of a hand included in the image according to any one of the above gesture recognition devices;
  • the operation execution module 3 is configured to perform a control operation corresponding to a recognition result of a gesture.
  • the required image can be captured by the shooting device, and the image can also be directly received through various types of receiving methods.
  • a gesture of a hand included in an image may be recognized in the acquired image.
  • the corresponding control operation can be performed according to the gesture recognized in the image.
  • the operation execution module includes: a control instruction acquisition submodule, configured to acquire a control instruction corresponding to a recognition result of a gesture according to a mapping relationship between a predetermined gesture and the control instruction; An operation execution sub-module is configured to control the electronic device to perform a corresponding operation according to the control instruction.
  • control instruction may be determined according to a gesture, and a rich control instruction may be determined for a gesture in an image by establishing a mapping relationship between the gesture and the control instruction as required.
  • Electronic equipment can be controlled by control instructions to achieve the purpose of controlling various devices such as vehicles.
  • the operation execution module includes: a special effect determination sub-module for determining a special effect corresponding to a recognition result of a gesture according to a predetermined mapping relationship between a gesture and a special effect; a special effect executor A module for drawing the special effect on the image by means of computer drawing.
  • the special effect execution sub-module is configured to draw the special effect in a computer drawing manner based on a hand or a finger key point of the hand included in the image.
  • a special effect corresponding to the gesture may be determined according to the gesture, and the performance of the image is increased by adding the special effect to the image.
  • the present disclosure also provides the above-mentioned device, electronic device, computer-readable storage medium, and program, which can be used to implement any of the gesture recognition methods or gesture processing methods provided by the present disclosure.
  • the method section The corresponding records are not repeated here.
  • An embodiment of the present disclosure further provides a computer-readable storage medium having computer program instructions stored thereon, and the computer program instructions implement any of the foregoing method embodiments when executed by a processor.
  • the computer-readable storage medium may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium.
  • An embodiment of the present disclosure further provides an electronic device including a processor and a memory for storing processor-executable instructions; wherein the processor implements any method embodiment of the present disclosure by calling the executable instructions, specifically For the working process and the setting method, reference may be made to the specific description of the foregoing corresponding method embodiments of the present disclosure, which is limited in space and will not be repeated here.
  • An embodiment of the present disclosure also provides a computer program, where the computer program includes computer-readable code, and when the computer-readable code runs in an electronic device, a processor in the electronic device executes any method implementation of the present disclosure. example.
  • Fig. 11 is a block diagram of an electronic device 800 according to an exemplary embodiment.
  • the electronic device 800 may be a terminal such as a mobile phone, a computer, a digital broadcasting terminal, a messaging device, a game console, a tablet device, a medical device, a fitness device, a personal digital assistant, and the like.
  • the electronic device 800 may include one or more of the following components: a processing component 802, a memory 804, a power component 806, a multimedia component 808, an audio component 810, an input / output (I / O) interface 812, and a sensor component 814 , And communication component 816.
  • the processing component 802 generally controls overall operations of the electronic device 800, such as operations associated with display, telephone calls, data communications, camera operations, and recording operations.
  • the processing component 802 may include one or more processors 820 to execute instructions to complete all or part of the steps of the method described above.
  • the processing component 802 may include one or more modules to facilitate the interaction between the processing component 802 and other components.
  • the processing component 802 may include a multimedia module to facilitate the interaction between the multimedia component 808 and the processing component 802.
  • the memory 804 is configured to store various types of data to support operation at the electronic device 800. Examples of such data include instructions for any application or method for operating on the electronic device 800, contact data, phone book data, messages, pictures, videos, and the like.
  • the memory 804 may be implemented by any type of volatile or non-volatile storage devices, or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), Programming read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.
  • SRAM static random access memory
  • EEPROM electrically erasable programmable read-only memory
  • EPROM Programming read-only memory
  • PROM programmable read-only memory
  • ROM read-only memory
  • magnetic memory flash memory
  • flash memory magnetic disk or optical disk.
  • the power component 806 provides power to various components of the electronic device 800.
  • the power component 806 may include a power management system, one or more power sources, and other components associated with generating, managing, and distributing power for the electronic device 800.
  • the multimedia component 808 includes a screen that provides an output interface between the electronic device 800 and a user.
  • the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive an input signal from a user.
  • the touch panel includes one or more touch sensors to sense touch, swipe, and gestures on the touch panel. The touch sensor may not only sense the boundary of a touch or slide action, but also detect the duration and pressure associated with the touch or slide operation.
  • the multimedia component 808 includes a front camera and / or a rear camera. When the electronic device 800 is in an operation mode, such as a shooting mode or a video mode, the front camera and / or the rear camera can receive external multimedia data. Each front camera and rear camera can be a fixed optical lens system or have focal length and optical zoom capabilities.
  • the audio component 810 is configured to output and / or input audio signals.
  • the audio component 810 includes a microphone (MIC).
  • the microphone is configured to receive an external audio signal.
  • the received audio signal may be further stored in the memory 804 or transmitted via the communication component 816.
  • the audio component 810 further includes a speaker for outputting audio signals.
  • the I / O interface 812 provides an interface between the processing component 802 and a peripheral interface module.
  • the peripheral interface module may be a keyboard, a click wheel, a button, or the like. These buttons may include, but are not limited to: a home button, a volume button, a start button, and a lock button.
  • the sensor component 814 includes one or more sensors for providing various aspects of the state evaluation of the electronic device 800.
  • the sensor component 814 may detect the on / off state of the electronic device 800, and the relative positioning of the components, such as the display and keypad of the electronic device 800.
  • the sensor component 814 may also detect the electronic device 800 or an electronic device 800 The position of the component changes, the presence or absence of the user's contact with the electronic device 800, the orientation or acceleration / deceleration of the electronic device 800, and the temperature change of the electronic device 800.
  • the sensor component 814 may include a proximity sensor configured to detect the presence of nearby objects without any physical contact.
  • the sensor component 814 may also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications.
  • the sensor component 814 may further include an acceleration sensor, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.
  • the communication component 816 is configured to facilitate wired or wireless communication between the electronic device 800 and other devices.
  • the electronic device 800 can access a wireless network based on a communication standard, such as WiFi, 2G, or 3G, or a combination thereof.
  • the communication component 816 receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel.
  • the communication component 816 further includes a near field communication (NFC) module to facilitate short-range communication.
  • the NFC module can be implemented based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra wideband (UWB) technology, Bluetooth (BT) technology, and other technologies.
  • RFID radio frequency identification
  • IrDA infrared data association
  • UWB ultra wideband
  • Bluetooth Bluetooth
  • the electronic device 800 may be implemented by one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), Implementation of a programming gate array (FPGA), controller, microcontroller, microprocessor, or other electronic component to perform the above method.
  • ASICs application specific integrated circuits
  • DSPs digital signal processors
  • DSPDs digital signal processing devices
  • PLDs programmable logic devices
  • FPGA programming gate array
  • controller microcontroller, microprocessor, or other electronic component to perform the above method.
  • a non-volatile computer-readable storage medium such as a memory 804 including computer program instructions, and the computer program instructions may be executed by the processor 820 of the electronic device 800 to complete the above method.
  • Fig. 12 is a block diagram of an electronic device 1900 according to an exemplary embodiment.
  • the electronic device 1900 may be provided as a server.
  • the electronic device 1900 includes a processing component 1922, which further includes one or more processors, and a memory resource represented by a memory 1932, for storing instructions executable by the processing component 1922, such as an application program.
  • the application program stored in the memory 1932 may include one or more modules each corresponding to a set of instructions.
  • the processing component 1922 is configured to execute instructions to perform the method described above.
  • the electronic device 1900 may further include a power supply component 1926 configured to perform power management of the electronic device 1900, a wired or wireless network interface 1950 configured to connect the electronic device 1900 to a network, and an input / output (I / O) interface 1958 .
  • the electronic device 1900 can operate based on an operating system stored in the memory 1932, such as Windows ServerTM, Mac OSXTM, UnixTM, LinuxTM, FreeBSDTM, or the like.
  • a non-volatile computer-readable storage medium is also provided, such as a memory 1932 including computer program instructions, which can be executed by the processing component 1922 of the electronic device 1900 to complete the above method.
  • the present disclosure may be a system, method, and / or computer program product.
  • the computer program product may include a computer-readable storage medium having computer-readable program instructions for causing a processor to implement various aspects of the present disclosure.
  • the computer-readable storage medium may be a tangible device that can hold and store instructions used by the instruction execution device.
  • the computer-readable storage medium may be, for example, but not limited to, an electric storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing.
  • Non-exhaustive list of computer-readable storage media include: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM) Or flash memory), static random access memory (SRAM), portable compact disc read only memory (CD-ROM), digital versatile disc (DVD), memory stick, floppy disk, mechanical encoding device, such as a printer with instructions stored thereon A protruding structure in the hole card or groove, and any suitable combination of the above.
  • RAM random access memory
  • ROM read-only memory
  • EPROM erasable programmable read-only memory
  • flash memory flash memory
  • SRAM static random access memory
  • CD-ROM compact disc read only memory
  • DVD digital versatile disc
  • memory stick floppy disk
  • mechanical encoding device such as a printer with instructions stored thereon A protruding structure in the hole card or groove, and any suitable combination of the above.
  • Computer-readable storage media used herein are not to be interpreted as transient signals per se, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through waveguides or other transmission media (for example, light pulses through fiber optic cables), or via electrical wires Electrical signal transmitted.
  • the computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to various computing / processing devices, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network.
  • the network may include copper transmission cables, fiber optic transmission, wireless transmission, routers, firewalls, switches, gateway computers, and / or edge servers.
  • the network adapter card or network interface in each computing / processing device receives computer-readable program instructions from the network and forwards the computer-readable program instructions for storage in a computer-readable storage medium in each computing / processing device .
  • Computer program instructions for performing the operations of the present disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-related instructions, microcode, firmware instructions, state setting data, or in one or more programming languages.
  • the programming languages include object-oriented programming languages—such as Smalltalk, C ++, and the like—and conventional procedural programming languages—such as the "C" language or similar programming languages.
  • Computer-readable program instructions may be executed entirely on a user's computer, partly on a user's computer, as a stand-alone software package, partly on a user's computer, partly on a remote computer, or entirely on a remote computer or server carried out.
  • the remote computer can be connected to the user's computer through any kind of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (such as through an Internet service provider using the Internet connection).
  • LAN local area network
  • WAN wide area network
  • an electronic circuit such as a programmable logic circuit, a field-programmable gate array (FPGA), or a programmable logic array (PLA), can be personalized by using state information of computer-readable program instructions.
  • FPGA field-programmable gate array
  • PDA programmable logic array
  • the electronic circuit can Computer-readable program instructions are executed to implement various aspects of the present disclosure.
  • These computer-readable program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, or other programmable data processing device, thereby producing a machine such that when executed by a processor of a computer or other programmable data processing device , Means for implementing the functions / actions specified in one or more blocks in the flowcharts and / or block diagrams.
  • These computer-readable program instructions may also be stored in a computer-readable storage medium, and these instructions cause a computer, a programmable data processing apparatus, and / or other devices to work in a specific manner. Therefore, a computer-readable medium storing instructions includes: An article of manufacture that includes instructions to implement various aspects of the functions / acts specified in one or more blocks in the flowcharts and / or block diagrams.
  • Computer-readable program instructions can also be loaded onto a computer, other programmable data processing device, or other device, so that a series of operating steps can be performed on the computer, other programmable data processing device, or other device to produce a computer-implemented process , So that the instructions executed on the computer, other programmable data processing apparatus, or other equipment can implement the functions / actions specified in one or more blocks in the flowchart and / or block diagram.
  • each block in the flowchart or block diagram may represent a module, a program segment, or a part of an instruction, which contains one or more components for implementing a specified logical function.
  • Executable instructions may also occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved.
  • each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts can be implemented in a dedicated hardware-based system that performs the specified function or action. , Or it can be implemented with a combination of dedicated hardware and computer instructions.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • General Health & Medical Sciences (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Data Mining & Analysis (AREA)
  • Multimedia (AREA)
  • Computing Systems (AREA)
  • Software Systems (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Human Computer Interaction (AREA)
  • Biophysics (AREA)
  • Molecular Biology (AREA)
  • Biomedical Technology (AREA)
  • Mathematical Physics (AREA)
  • Computational Linguistics (AREA)
  • Medical Informatics (AREA)
  • Databases & Information Systems (AREA)
  • Social Psychology (AREA)
  • Psychiatry (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • User Interface Of Digital Computer (AREA)
  • Image Analysis (AREA)

Abstract

本公开涉及一种手势识别方法、手势处理方法及装置。所述手势识别方法包括：检测图像中手部包括的手指的状态；根据所述手指的状态确定所述手部的状态向量；根据所述手部的状态向量确定所述手部的手势。本公开实施例根据各手指的状态确定状态向量，根据状态向量确定手势，识别效率高，通用性强。

Description

手势识别方法、手势处理方法及装置
本申请要求在2018年8月17日提交中国专利局、申请号为201810942882.1、发明名称为“手势识别方法、手势处理方法及装置”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本公开涉及图像处理技术领域，尤其涉及一种手势识别方法、手势处理方法及装置。
背景技术
非接触式人机交互场景在生活中的应用越来越广泛。用户可以利用不同的手势方便地表达不同的人机交互指令。
发明内容
本公开提出了一种手势识别技术方案。
根据本公开的一方面，提供了一种手势识别方法，包括：检测图像中手部包括的手指的状态；根据所述手指的状态确定所述手部的状态向量；根据所述手部的状态向量确定所述手部的手势。
根据本公开的一方面，提供了一种手势处理方法，所述方法包括：获取图像；采用上述手势识别方法识别所述图像包括的手部的手势；执行与手势的识别结果相应的控制操作。
根据本公开的一方面，提供了一种手势识别装置，所述装置包括：状态检测模块，用于检测图像中手部包括的手指的状态；状态向量获取模块，用于根据所述手指的状态确定所述手部的状态向量；手势确定模块，用于根据所述手部的状态向量确定所述手部的手势。
根据本公开的一方面，提供了一种手势处理装置，所述装置包括：图像获取模块，用于获取图像；手势获取模块，用于采用上述手势识别装置识别所述图像包括的手部的手势；操作执行模块，用于执行与手势的识别结果相应的控制操作。
根据本公开的一方面，提供了一种电子设备，包括：处理器；用于存储处理器可执行指令的存储器；其中，所述处理器通过调用所述可执行指令实现上述手势识别方法和/或手势处理方法。
根据本公开的一方面，提供了一种计算机可读存储介质，其上存储有计算机程序指令，所述计算机程序指令被处理器执行时实现上述手势识别方法和/或手势处理方法。
根据本公开的一方面，提供了一种计算机程序，所述计算机程序包括计算机可读代码，当所述计算机可读代码在电子设备中运行时，所述电子设备中的处理器执行上述手势识别方法和/或手势处理方法。
在本公开实施例中，通过检测图像中手部包括的手指的状态，根据所述手指的状态确定所述手部的状态向量，并根据确定出的手部的状态向量确定手部的手势。本公开实施例根据各手指的状态确定状态向量，根据状态向量确定手势，识别效率高，通用性强。
根据下面参考附图对示例性实施例的详细说明，本公开的其它特征及方面将变得清楚。
附图说明
包含在说明书中并且构成说明书的一部分的附图与说明书一起示出了本公开的示例性实施例、特征和方面，并且用于解释本公开的原理。
图1示出根据本公开实施例的手势识别方法的流程图；
图2示出根据本公开实施例的手势识别方法中手指的状态示意图；
图3示出根据本公开实施例的手势识别方法的流程图；
图4示出根据本公开实施例的手势识别方法的流程图；
图5示出根据本公开实施例的手势识别方法的流程图；
图6示出根据本公开实施例的手势识别方法中神经网络的数据处理流程图；
图7示出根据本公开实施例的手势识别方法的流程图；
图8示出根据本公开实施例的手势处理方法的流程图；
图9示出根据本公开实施例的手势识别装置的框图；
图10示出根据本公开实施例的手势处理的框图；
图11是根据示例性实施例示出的一种电子设备的框图；
图12是根据示例性实施例示出的一种电子设备的框图。
具体实施方式
以下将参考附图详细说明本公开的各种示例性实施例、特征和方面。附图中相同的附图标记表示功能相同或相似的元件。尽管在附图中示出了实施例的各种方面，但是除非特别指出，不必按比例绘制附图。
在这里专用的词“示例性”意为“用作例子、实施例或说明性”。这里作为“示例性”所说明的任何实施例不必解释为优于或好于其它实施例。
另外，为了更好的说明本公开，在下文的具体实施方式中给出了众多的具体细节。本领域技术人员应当理解，没有某些具体细节，本公开同样可以实施。在一些实例中，对于本领域技术人员熟知的方法、手段、元件和电路未作详细描述，以便于凸显本公开的主旨。下面这几个具体的实施例可以相互结合，对于相同或相似的概念或过程可能在某些实施例不再赘述。可以理解，以下实施例仅为本公开的可选实现方式，不应理解为对本公开保护范围的实质限制，本领域技术人员可以在此基础上采用其他实现方式，均在本公开保护范围之内。
图1示出根据本公开实施例的手势识别方法的流程图。所述手势识别方法可以由终端设备或服务器等电子设备执行，其中，终端设备可以为用户设备(User Equipment，UE)、移动设备、用户终端、终端、蜂窝电话、无绳电话、个人数字处理(Personal Digital Assistant，PDA)、手持设备、计算设备、车载设备、可穿戴设备等。在一些可能的实现方式中，所述手势识别方法可以通过处理器调用存储器中存储的计算机可读指令的方式来实现。
如图1所示，所述方法包括：
步骤S10，检测图像中手部包括的手指的状态。
在一种可能的实现方式中，图像可以是静态的图像，也可以是视频流中的帧图像。可以利用图像识别方法，在图像中获取手部中各手指的状态。可以获取手部中五个手指的状态，也可以获取指定的多个或单个手指的状态，例如可以只获取食指的状态。
在一种可能的实现方式中，所述手指的状态表示所述手指相对于所述手部的掌根是否伸出和/或伸出程度的状态。当手部的手势是握拳时，各手指相对于掌根为非伸出状态。当手指相对于掌根为伸出状态时，根据手指相对于掌部的位置或自身的弯曲程度，又可以对手指的状态进行进一步的划分。例如，手指的状态可以分为非伸出状态或伸出状态两种状态，也可以分为非伸出状态、半伸出状态、 伸出状态三种，还可以分为伸出状态、非伸出状态、半伸出状态、弯曲状态等多种状态。
在一种可能的实现方式中，所述手指的状态包括以下一种或多种：伸出状态、非伸出状态、半伸出状态、弯曲状态。其中，按照手指与掌部的位置关系和手指自身的弯曲程度，手部从握拳到五指全部张开的过程中，各手指的状态可以依次为：非伸出状态、半伸出状态、弯曲状态、伸出状态。还可以根据需求，为不同的手指的状态划分不同的状态级别。本公开不限定各手指的状态的分类方式、数量和使用顺序。
图2示出根据本公开实施例的手势识别方法中手指的状态示意图，如图2所示的图像中，大拇指的状态为非伸出状态、食指的状态为伸出状态、中指的状态为伸出状态，无名指的状态为非伸出状态，小指的状态为非伸出状态。可以在图像中获取所有五个手指的状态，也可以只获取指定手指(例如食指和中指)的状态。
步骤S20，根据所述手指的状态确定所述手部的状态向量。
在一种可能的实现方式中，所述根据所述手指的状态确定所述手部的状态向量，包括：根据所述手指的状态确定所述手指的状态值，其中，不同状态对应的手指的状态值不同；根据所述手指的状态值确定所述手部的状态向量。
在一种可能的实现方式中，可以为不同的手指的状态确定相应的状态值，并建立手指的状态和状态值之间的对应关系。手指的状态值可以是数字、字母或符号的其中一种或任意组合。可以根据获取到的手指的状态和所建立的对应关系，确定手指的状态值，再利用手指的状态值得到手部的状态向量。手部的状态向量可以包括数组、列表或矩阵等各种形式。
在一种可能的实现方式中，可以将手指的状态值按照设定的手指顺序组合后得到手部的状态向量。例如，可以根据五个手指的状态值得到手部的状态向量。可以按照大拇指、食指、中指、无名指、小指的顺序，将五个手指的状态值组合后得到手部的状态向量。也可以按照其它任意设定的顺序将手指的状态值组合后得到手部的状态向量。
例如，如图2所示的图像中，可以利用状态值A表示非伸出状态，用状态值B表示伸出状态。如图2所示，大拇指的状态值为A、食指的状态值为B、中指的状态值为B，无名指的状态值为A，小指的状态值为A。则手部的状态向量可以为(A，B，B，A，A)。
步骤S30，根据所述手部的状态向量确定所述手部的手势。
在一种可能的实现方式中，可以利用手部中各手指的状态确定手部的手势。可以根据需求确定手指的不同状态，根据手指的不同状态确定手部的状态向量，再根据手部的状态向量确定手部的手势。手指状态的识别过程便捷可靠，使得手势的确定过程也更加便捷可靠。可以建立手部的状态向量和手势之间的对应关系，通过调整状态向量与手势之间的对应关系，可以更加灵活的根据状态向量确定手势，使得手势的确定过程更加灵活，能够适应不同的应用环境。例如，手部的状态向量1对应手势1、手部的状态向量2对应手势2、手部的状态向量3对应手势3。可以根据需求确定手部的状态向量和手势之间的对应关系。可以一个手部的状态向量对应一个手势，也可以多个手部的状态向量对应一个手势。
在一种可能的实现方式中，例如，如图2所示的图像中，手部的状态向量为(A，B，B，A，A)，在手部的状态向量和手势之间的对应关系中，状态向量为(A，B，B，A，A)对应的手势可以为“数字2”或“胜利”。
在本实施例中，通过检测图像中手部包括的手指的状态，根据所述手指的状态确定所述手部的状态向量，并根据确定出的手部的状态向量确定手部的手势。本公开实施例根据各手指的状态确定状态向量，根据状态向量确定手势，识别效率高，通用性强。
本实施例在图像中识别各手指的状态的识别效率高，使得本实施例识别手势的识别效率高。且本实施例可以根据需求任意调整手指的状态和手势之间的对应关系，可以根据相同的图像识别不同需求下定义的不同手势，使得确定出的手势的通用性强。
在一种可能的实现方式中，所述手指的状态包括伸出状态或非伸出状态，根据所述手指的状态确定所述手部的状态向量，包括：
当手指的状态为伸出状态时，确定所述手指的状态值为第一状态值；或
当手指的状态为非伸出状态时，确定所述手指的状态值为第二状态值；
根据所述手指的状态值确定所述手部的状态向量。
在一种可能的实现方式中，可以利用数字、字母或符号的其中一种或任意组合来表示第一状态值和第二状态值。第一状态值和第二状态值可以是表示相反含义的两个值，例如第一状态值可以为有效，第二状态值可以为无效。第一状态值和第二状态值也可以是两个不同数值的数字，例如第一状态值可以为1，第二状态值可以为0。如图2所示的图像中，大拇指的状态值为0、食指的状态值为1、中指的状态值为1，无名指的状态值为0，小指的状态值为0，手部的状态向量为(0，1，1，0，0)。
在本实施例中，可以利用第一状态值和第二状态值来确定手部的状态向量。利用两个状态值组成的手部的状态向量，可以简单直观的表达出手部各手指的状态。
图3示出根据本公开实施例的手势识别方法的流程图，如图3所示，所示方法还包括：
步骤S40，检测所述图像中手部包括的手指的位置信息。
在一种可能的实现方式中，手指的位置信息可以包括手指在图像中的位置信息。手指的位置信息可以包括手指的像素在图像中的坐标位置信息。也可以将图像分割成网格，并将手指的像素所在的网格的位置信息，确定为手指的位置信息。网格的位置信息可以包括网格的编号。
在一种可能的实现方式中，手指的位置信息也可以包括手指相对于图像中的目标对象的位置信息。例如，图像中的画面为一个人在弹钢琴，图像中手指的位置信息可以包括手指相对于琴键的位置信息。例如，手指1距离琴键的距离为0，手指2距离琴键的距离为3厘米等。
在一种可能的实现方式中，手指的位置信息可以包括一维或多维的位置信息。根据手指的位置信息，可以得到手指之间的相对位置关系。
步骤S50，根据所述手指的位置信息，确定所述手部的位置向量。
在一种可能的实现方式中，可以根据设定的手指的顺序，将不同手指的位置信息组合后得到手部的位置向量。手部的位置向量可以包括数组、列表或矩阵等各种形式。
步骤S30，包括：
步骤S31，根据所述手部的状态向量和所述手部的位置向量，确定所述手部的手势。
在一种可能的实现方式中，根据手部的状态向量可以得到手部中手指的状态，结合手部的位置向量中手指的位置，可以确定出更加精准的手势。例如，如图2所示的图像中，手部的状态向量为(0，1，1，0，0)，位置向量为(L1，L2，L3，L4，L5)。如果只根据手部的状态向量，可以确定手部中食指和中指的状态为伸出状态，其它手指为非伸出状态，手部的状态向量可以确定手部的手势为“数字2”或“胜利”。
如果将手部的位置向量与手部的状态向量结合，可以确定食指和中指伸出且分开了一定的角度，如图2所示，则手部的手势可以是“数字2”或“胜利”。如果根据手部的状态向量和手部的位置向量，可以确定食指和中指伸出且是并拢在一起的(图中未示出)，则手部的手势可以是“数字2”，不能是“胜利”。
可以根据需求将手部的状态向量和手部的位置向量进行组合，得到组合向量，再建立组合向量和手势之间的对应关系。相同的状态向量和不同的位置向量组成的不同组合向量，可以对应不同的手势，也可以对应相同的手势。
在本实施例中，可以根据手部的状态向量和位置向量确定手部的手势。将手部的位置向量和状态向量相结合，可以得到更加精准的手势。
图4示出根据本公开实施例的手势识别方法的流程图，如图4所示，所述方法中步骤S40，包括：
步骤S41，检测所述图像中所述手部包括的手指的关键点，获得所述手指的关键点的位置信息。
在一种可能的实现方式中，所述关键点包括指尖和/或指关节，其中，指关节可以包括掌指关节或指间关节。可以利用手指的指尖和/或指关节的位置准确的表示出手指的位置信息。例如，如图2所示的图像中，手指的关键点为指尖，可以确定各手指的指尖的位置信息为：拇指(X 1，Y 1)、食指(X 2，Y 2)、中指(X 3，Y 3)、无名指(X 4，Y 4)、小指(X 5，Y 5)，其中，拇指、无名指和小指的指尖的坐标点比较接近。
步骤S50，包括：
步骤S51，根据所述手指的关键点的位置信息，确定所述手部的位置向量。
在一种可能的实现方式中，例如，如图2所示的图像中，手部的位置向量可以为(X 1，Y 1，X 2，Y 2，X 3，Y 3，X 4，Y 4，X 5，Y 5)。
根据手部的状态向量(0，1，1，0，0)和手部的位置向量(X 1，Y 1，X 2，Y 2，X 3，Y 3，X 4，Y 4，X 5，Y 5)，可以确定出手部中食指和中指伸出且指尖间隔一定的距离，其余三指收拢在手掌位置，手部的手势为“胜利”。
在本实施例中，可以根据手部的手指的关键点的位置信息，得到手部的位置向量。使得手部的位置向量的确定过程更加简单。
在一种可能的实现方式中，步骤S41包括：检测所述图像中所述手部包括的状态为非伸出状态以外的手指的关键点，获得所述关键点的位置信息。
在一种可能的实现方式中，由于手势的确定可以根据状态为非伸出状态以外的手指来确定，因此可以在图像中确定状态为非伸出状态以外的手指上的关键点，并获取关键点的位置信息。可以将状态为非伸出状态的手指关键点的位置坐标，确定为一个在图像中不存在的坐标值，例如可以以图像的上边缘为X轴正向，左侧边缘为Y轴正向，无效坐标可以为(-1，-1)。
如图2所示的图像中，可以以图像的上边缘为X轴正向，左侧边缘为Y轴正向，手指的关键点为指尖，可以根据手部的状态向量(0，1，1，0，0)，在图像中获取手指的指尖的位置信息为：拇指(-1，-1)、食指(X 2，Y 2)、中指(X 3，Y 3)、无名指(-1，-1)、小指(-1，-1)。手部的位置向量可以为(-1，-1，X 2，Y 2，X 3，Y 3，-1，-1，-1，-1)。也可以将状态为非伸出状态的手指关键点的位置坐标填零。
根据手部的状态向量(0，1，1，0，0)和手部的位置向量(-1，-1，X 2，Y 2，X 3，Y 3，-1，-1，-1，-1)，可以确定出手部中食指和中指伸出且指尖间隔一定的距离，其余三指重叠收拢在手掌位置，手部的手势为“胜利”。
在本实施例中，可以根据状态为非伸出状态以外的手指的关键点的位置信息，得到手部的位置向量。使得手部的位置向量的确定过程更加高效。
图5示出根据本公开实施例的手势识别方法的流程图，如图5所示，所述方法中步骤S10，包括：
步骤S11，将所述图像输入神经网络，经所述神经网络检测所述图像中手部包括的手指的状态。
在一种可能的实现方式中，神经网络是一种模仿生物神经网络的结构和功能的数学模型或者计算 模型。神经网络可以包括输入层、中间层和输出层。输入层负责接收来自外部的输入数据，并将输入数据传递给中间层。中间层负责信息交换，根据信息变化能力的需求，中间层可以设计为单隐藏层或多隐藏层。中间层将输出结果传递到输出层进行进一步处理后，得到神经网络的输出结果。输入层、中间层和输出层都可以包括若干神经元，各神经元之间可以用带可变权重的有向连接。神经网络通过对已知信息的反复学习训练，通过逐步调整改变神经元连接权重的方法，达到建立模拟输入输出之间关系模型的目的。训练好的神经网络可以利用模拟好的输入输出之间的关系模型，检测输入信息，并给出与输入信息对应的输出信息。例如，神经网络可以包括卷积层、池化层和全连接层等。可以利用神经网络提取图像中的特征，并根据提取到的特征确定图像中的手指的状态。
在本实施例中，可以利用神经网络强大的处理能力迅速、准确地确定出图像中手部包括的手指的状态。
在一种可能的实现方式中，所述神经网络包括多个状态分支网络，步骤S11包括：经所述神经网络的不同状态分支网络分别检测所述图像中手部包括的不同手指的状态。
在一种可能的实现方式中，可以在神经网络中设置五个状态分支网络，每个状态分支网络用于在图像中获取一个手指的状态。
在一种可能的实现方式中，图6示出根据本公开实施例的手势识别方法中神经网络的数据处理流程图。在图6中，神经网络可以包括卷积层和全连接层。其中，卷积层可以包括第一卷积层、第二卷积层、第三卷积层和第四卷积层。第一卷积层可以包括一层卷积层“conv1_1”，第二卷积层至第四卷积层可以分别有两层卷积层，例如可以为“conv2_1”至“conv4_2”。第一卷积层、第二卷积层、第三卷积层和第四卷积层可以用于提取图像中的特征。
全连接层可以包括第一全连接层“ip1_fingers”、第二全连接层“ip2_fingers”和第三全连接层“ip3_fingers”。第一全连接层、第二全连接层和第三全连接层可以用于确定手指的状态，获取手指的状态向量。其中，“ip3_fingers”可以分割为五个状态分支网络，分别为第一状态分支网络(loss_littlefinger)、第二状态分支网络(loss_ringfinger)、第三状态分支网络(loss_middlefinger)、第四状态分支网络(loss_forefinger)和第五状态分支网络(loss_thumb)。每个状态分支网络对应一个手指，且每个状态分支网络可以单独训练。
在一种可能的实现方式中，所述全连接层还包括位置分支网络，步骤S40可包括：
经所述神经网络的所述位置分支网络检测所述图像中所述手部包括的手指的位置信息。
在图6中神经网络还包括位置分支网络，位置分支网络可以包括第五全连接层“ip1_points”、第六全连接层“ip2_points”和第七全连接层“ip3_points”。第五全连接层、第六全连接层和第七全连接层用于获取手指的位置信息。
此外，图6中卷积层还可以包括激活函数(relu_conv)，池化层(pool)，损失函数(loss)等，不再赘述。
在本实施例中，可以利用位置分支网络在图像中确定手指的位置信息，以及利用所述位置分支网络在所述图像中确定所述手指的位置信息。可以根据状态分支网络和位置分支网络，在图像中快捷准确地获取手指的状态信息和位置信息。
在一种可能的实现方式中，所述神经网络预先采用带有标注信息的样本图像训练而得，所述标注信息包括表示所述手指的状态的第一标注信息，和/或，表示所述手指的位置信息或关键点的位置信息的第二标注信息。
在一种可能的实现方式中，样本图像的标注信息可以包括表示手指的状态的第一标注信息。在神 经网络的训练过程中，可以将检测出的手指的状态，与第一标注信息进行比对，确定手势预测结果的损失。
在一种可能的实现方式中，样本图像的标注信息可以包括表示手指的位置信息或关键点的位置信息的第二标注信息。可以根据第二标注信息，得到各手指的位置或关键点的位置，并可以根据各手指的位置或关键点的位置确定各手指的状态。在神经网络的训练过程中，可以将检测出的手指的状态，与根据第二标注信息确定的手指的状态进行比对，确定手势预测结果的损失。
在一种可能的实现方式中，样本图像的标注信息可以包括第一标注信息和第二标注信息。在神经网络的训练过程中，可以将检测出的手指的状态与第一标注信息进行比对，将检测出的位置信息与第二标注进行比对，确定手势预测结果的损失。
在一种可能的实现方式中，所述第一标注信息包括由表示各手指的状态的第一标识值组成的状态向量；所述第二标注信息包括由标识各手指的位置信息或关键点的位置信息的第二标识值组成的位置向量。
在一种可能的实现方式中，所述样本图像中，不标注非伸出状态的手指的第二标注信息。可以为非伸出状态的手指设置无效的第二标识值，例如(-1，-1)。
在一种可能的实现方式中，可以根据手指的状态的数量，确定第一标注信息中的标识值。例如，手指的状态为非伸出状态或伸出状态，则第一标注信息中的第一标识值可以包括0(非伸出状态)或1(伸出状态)。手指的状态为非伸出状态、半伸出状态、弯曲状态和伸出状态，则第一标识值可以包括0(非伸出状态)、1(半伸出状态)、2(弯曲状态)、3(伸出状态)。可以根据各手指的第一标识值得到手部的第一标注信息，例如(0，1，1，0，0)
在一种可能的实现方式中，可以为样本图像建立图像坐标系，并根据所建立的图像坐标系，确定第二标注信息中的第二标识值。可以根据各手指的第二标识值得到手部的第二标注信息，例如(-1，-1，X 2，Y 2，X 3，Y 3，-1，-1，-1，-1)。
图7示出根据本公开实施例的手势识别方法的流程图，如图7所示，所述神经网络的训练步骤包括：
步骤S1，将手部的样本图像输入神经网络得到手部中手指的状态。
在一种可能的实现方式中，所述将手部的样本图像输入神经网络得到手部中手指的状态，包括：将手部的样本图像输入神经网络得到手部中手指的状态和位置信息。
在一种可能的实现方式中，手部的样本图像可以是标注了手指的状态和位置信息图像。可以将手部的样本图像输入神经网络，由神经网络提取图像中的特征，并根据提取到的特征确定手指的状态和位置信息。在后续的手势识别的步骤中，可以根据确定出的手指的状态和位置信息，确定手部的手势。
步骤S2，根据所述手指的状态确定手指的位置权重。
在一种可能的实现方式中，可以为手指的不同状态设置不同的位置权重。例如，可以为状态为伸出状态的手指设置较高的位置权重，为状态为非伸出状态的手指设置较低的位置权重。
在一种可能的实现方式中，所述根据所述手指的状态确定所述手指的位置权重，包括：当手指的状态为非伸出状态时，确定所述手指的位置权重为零权重。
在一种可能的实现方式中，当手指的状态为伸出状态时，可以确定所述手指的位置权重为非零权重；当手指的状态为非伸出状态时，可以确定所述手指的位置权重为零权重。
在一种可能的实现方式中，可以获取状态为伸出状态的手指的关键点的位置信息，并根据状态为伸出状态的手指的关键点的位置信息得到手部的位置信息，再根据手部的位置信息和状态信息确定手部的手势。例如，如图2所示的图像中，手部的状态向量为(0，1，1，0，0)，手部的位置向量为(-1， -1，X 2，Y 2，X 3，Y 3，-1，-1，-1，-1)。可以根据手部的状态向量，为食指和中指设置位置权重为1，为其余三指设置位置权重为0，可以得到手部的位置权重为(0，0，1，1，1，1，0，0，0，0)。
在一种可能的实现方式中，对于食指伸出另外四指收拢的手势，手部的状态向量为(0，1，0，0，0)，以指尖为关键点手部的位置向量为(-1，-1，X 2，Y 2，-1，-1，-1，-1，-1，-1)，位置权重为(0，0，1，1，0，0，0，0，0，0)。对于拳头的手势，手部的状态向量为(0，0，0，0，0)，以指尖为关键点手部的位置向量为(-1，-1，-1，-1，-1，-1，-1，-1，-1，-1)，位置权重为(0，0，0，0，0，0，0，0，0，0)。对于中指、无名指和小指伸出、拇指和和食指捏起的“OK”手势，手部的状态向量为(0，0，1，1，1)，以指尖为关键点手部的位置向量为(-1，-1，-1，-1，X 3，Y 3，X 4，Y 4，X 5，Y 5)，位置权重为(0，0，0，0，1，1，1，1，1，1)。
步骤S3，根据所述手指的状态和所述位置权重，确定所述神经网络的手势预测结果的损失。
在一种可能的实现方式中，根据所述手指的状态和所述位置权重，确定所述神经网络的手势预测结果的损失，包括：根据所述手指的状态、所述位置信息和所述位置权重，确定所述神经网络的手势预测结果的损失。
步骤S4，向所述神经网络反向传播所述损失，以调整所述神经网络的网络参数。
在一种可能的实现方式中，在神经网络的反向传播过程中，手指的位置向量中非伸出状态的手指的位置向量的取值，会影响到神经网络的反向传播中损失函数的计算结果。例如，如果只根据手指的状态和位置信息对所述神经网络进行反向传播，在如图2所示的图像中，手部的状态向量为(0，1，1，0，0)，手部的位置向量为(-1，-1，X 2，Y 2，X 3，Y 3，-1，-1，-1，-1)，在神经网络的反向传播中，拇指、无名指和小指的位置向量将趋近于-1，导致神经网络的反向传播出现偏差，训练出的神经网络的识别结果不准确。如果结合手部的位置权重(0，0，1，1，1，1，0，0，0，0)，在神经网络的反向传播中，将不对拇指、无名指和小指的位置向量进行反向传播的计算，训练出的神经网络的识别结果准确。
在本实施例中，根据手指的状态、位置信息和位置权重，对神经网络进行反向传播，可以减小手指的位置信息中位置坐标的取值对反向传播产生的不利影响，使得训练出的神经网络更加准确。
图8示出根据本公开实施例的手势处理方法的流程图。所述手势处理方法可以由终端设备或服务器等电子设备执行，其中，终端设备可以为用户设备(User Equipment，UE)、移动设备、用户终端、终端、蜂窝电话、无绳电话、个人数字处理(Personal Digital Assistant，PDA)、手持设备、计算设备、车载设备、可穿戴设备等。在一些可能的实现方式中，所述手势处理方法可以通过处理器调用存储器中存储的计算机可读指令的方式来实现。
如图8所示，所述方法包括：
步骤S60，获取图像。
步骤S70，采用上述任一项手势识别方法识别所述图像包括的手部的手势。
步骤S80，执行与手势的识别结果相应的控制操作。
在一种可能的实现方式中，可以通过拍摄装置拍摄所需要的图像，也可以通过各种类型的接收方式直接接收图像。可以根据本公开实施例中任一项所述的手势识别方法，在获取的图像中识别图像中包括的手部的手势。可以根据在图像中识别出的手势进行相应的控制操作。
在一种可能的实现方式中，步骤S80，包括：根据预先确定的手势与控制指令之间的映射关系，获取与手势的识别结果相应的控制指令；根据所述控制指令控制电子设备执行相应操作。
在一种可能的实现方式中，可以根据需求建立手势和控制指令之间的映射关系。例如，可以为手 势1设定“前进”的控制指令，为手势2设定“停止”的控制指令。可以在图像中确定出手部的手势后，根据手势和建立的映射关系，确定与手势对应的控制指令。
在一种可能的实现方式中，可以根据确定出的手势的控制指令，控制机器人、机械设备、车辆等装置上配置的电子设备，以实现对机器人、机械设备、车辆等装置的自动控制。例如，可以利用机器人配置的拍摄设备拍摄控制者的手部图像后，利用本公开实施例中的手势识别方法识别所拍摄的图像中的手势，并根据手势确定控制指令，最终实现对机器人的自动控制。本公开不限定控制指令所控制的电子设备的类型。
在本实施例中，可以根据手势确定控制指令，可以根据需求通过建立手势与控制指令之间的映射关系，为图像中的手势确定丰富的控制指令。可以通过控制指令控制电子设备，达到控制车辆等各种装置的目的。
在一种可能的实现方式中，步骤S80，包括：根据预先确定的手势与特效之间的映射关系，确定与手势的识别结果相应的特效；在所述图像上采用计算机绘图的方式绘制所述特效。
在一种可能的实现方式中，可以建立手势与特效之间的映射关系。特效可以用于强调手势的内容，或加强手势的表现能力等。例如，当识别到手势为“胜利”时，可以做出放烟花的特效等。
在一种可能的实现方式中，可以采用计算机绘图的方式绘制特效，并将绘制完成的特效与图像的内容一起进行显示。特效可以包括二维贴纸特效、二维图像特效、三维特效、例子特效、局部图像变形特效等。本公开不限定特效的内容、类型及实现方式。
在一种可能的实现方式中，在所述图像上采用计算机绘图的方式绘制所述特效，包括：
基于所述图像包括的手部或者手部的手指关键点，采用计算机绘图的方式绘制所述特效。
在一种可能的实现方式中，播放图像时，可以根据手部的位置信息，在图像中增加文字、符号或图像等附加信息。附加信息可以包括以下信息中的其中一种或任意组合：文字、图像、符号、字母、数字。例如，可以在手指的指尖部位，增加“感叹号”等符号，或增加“闪电”等图像信息，用来在图像中增加编辑者需要表达或强调的信息，丰富图像的表达能力。
在本实施例中，可以根据手势确定与之对应的特效，通过在图像上增加特效，增加图像的表现能力。
图9示出根据本公开实施例的手势识别装置的框图，如图9所示，所述手势识别装置，包括：
状态检测模块10，用于检测图像中手部包括的手指的状态；
状态向量获取模块20，用于根据所述手指的状态确定所述手部的状态向量；
手势确定模块30，用于根据所述手部的状态向量确定所述手部的手势。
本实施例中，通过检测图像中手部包括的手指的状态，根据所述手指的状态确定所述手部的状态向量，并根据确定出的手部的状态向量确定手部的手势。本公开实施例根据各手指的状态确定状态向量，根据状态向量确定手势，识别效率高，通用性强。
在一种可能的实现方式中，所述手指的状态表示所述手指相对于所述手部的掌根是否伸出和/或伸出程度的状态。当手部的手势是握拳时，各手指相对于掌根为非伸出状态。当手指相对于掌根为伸出状态时，根据手指相对于掌部的位置或自身的弯曲程度，又可以对手指的状态进行进一步的划分。例如，手指的状态可以分为非伸出状态或伸出状态两种状态，也可以分为非伸出状态、半伸出状态、伸出状态三种，还可以分为伸出状态、非伸出状态、半伸出状态、弯曲状态等多种状态。
在一种可能的实现方式中，所述状态向量获取模块，包括：状态值获取子模块，用于根据所述手指的状态确定所述手指的状态值，其中，不同状态对应的手指的状态值不同；第一状态向量获取子模 块，用于根据所述手指的状态值确定所述手部的状态向量。
在一种可能的实现方式中，可以为不同的手指的状态确定相应的状态值，并建立手指的状态和状态值之间的对应关系。手指的状态值可以是数字、字母或符号的其中一种或任意组合。可以根据获取到的手指的状态和所建立的对应关系，确定手指的状态值，再利用手指的状态值得到手部的状态向量。手部的状态向量可以包括数组、列表或矩阵等各种形式。
在一种可能的实现方式中，所述手指的状态包括以下一种或多种：伸出状态、非伸出状态、半伸出状态、弯曲状态。其中，按照手指与掌部的位置关系和手指自身的弯曲程度，手部从握拳到五指全部张开的过程中，各手指的状态可以依次为：非伸出状态、半伸出状态、弯曲状态、伸出状态。还可以根据需求，为不同的手指的状态划分不同的状态级别。本公开不限定各手指的状态的分类方式、数量和使用顺序。
在一种可能的实现方式中，所述装置还包括：位置信息获取模块，用于检测所述图像中手部包括的手指的位置信息；位置向量获取模块，用于根据所述手指的位置信息，确定所述手部的位置向量；
所述手势确定模块，包括：第一手势确定子模块，用于根据所述手部的状态向量和所述手部的位置向量，确定所述手部的手势。
在本实施例中，可以根据手部的状态向量和位置向量确定手部的手势。将手部的位置向量和状态向量相结合，可以得到更加精准的手势。
在一种可能的实现方式中，所述位置信息获取模块，包括：关键点检测子模块，用于检测所述图像中所述手部包括的手指的关键点，获得所述手指的关键点的位置信息；
所述位置向量获取模块，包括：第一位置向量获取子模块，用于根据所述手指的关键点的位置信息，确定所述手部的位置向量。
在本实施例中，可以根据手部的手指的关键点的位置信息，得到手部的位置向量。使得手部的位置向量的确定过程更加简单。
在一种可能的实现方式中，所述关键点检测子模块，用于：检测所述图像中所述手部包括的状态为非伸出状态以外的手指的关键点，获得所述关键点的位置信息。
在本实施例中，可以根据状态为非伸出状态以外的手指的关键点的位置信息，得到手部的位置向量。使得手部的位置向量的确定过程更加高效。
在一种可能的实现方式中，所述关键点包括指尖和/或指关节。其中，指关节可以包括掌指关节或指间关节。可以利用手指的指尖和/或指关节的位置准确的表示出手指的位置信息。
在一种可能的实现方式中，所述状态检测模块，包括：第一状态检测子模块，用于将所述图像输入神经网络，经所述神经网络检测所述图像中手部包括的手指的状态。
在本实施例中，可以利用神经网络强大的处理能力迅速、准确地确定出图像中手部包括的手指的状态。
在一种可能的实现方式中，所述神经网络包括多个状态分支网络，所述第一状态检测子模块，用于：经所述神经网络的不同状态分支网络分别检测所述图像中手部包括的不同手指的状态。
在一种可能的实现方式中，可以在神经网络中设置五个状态分支网络，每个状态分支网络用于在图像中获取一个手指的状态。
在一种可能的实现方式中，所述神经网络还包括位置分支网络，所述位置信息获取模块包括：第一位置信息获取子模块，用于经所述神经网络的所述位置分支网络检测所述图像中所述手部包括的手指的位置信息。
在本实施例中，可以利用位置分支网络在图像中确定手指的位置信息，以及利用所述位置分支网络在所述图像中确定所述手指的位置信息。可以根据状态分支网络和位置分支网络，在图像中快捷准确地获取手指的状态信息和位置信息。
在一种可能的实现方式中，所述神经网络预先采用带有标注信息的样本图像训练而得，所述标注信息包括表示所述手指的状态的第一标注信息，和/或，表示所述手指的位置信息或关键点的位置信息的第二标注信息。
在一种可能的实现方式中，所述样本图像中，不标注非伸出状态的手指的第二标注信息。可以为非伸出状态的手指设置无效的第二标识值。
在一种可能的实现方式中，所述第一标注信息包括由表示各手指的状态的第一标识值组成的状态向量；所述第二标注信息包括由标识各手指的位置信息或关键点的位置信息的第二标识值组成的位置向量。
在一种可能的实现方式中，所述神经网络包括训练模块，所述训练模块包括：状态获取子模块，用于将手部的样本图像输入神经网络得到手部中手指的状态；位置权重确定子模块，用于根据所述手指的状态确定手指的位置权重；损失确定子模块，用于根据所述手指的状态和所述位置权重，确定所述神经网络的手势预测结果的损失；反向传播子模块，用于向所述神经网络反向传播所述损失，以调整所述神经网络的网络参数。
在一种可能的实现方式中，所述状态获取子模块，用于：将手部的样本图像输入神经网络得到手部中手指的状态和位置信息；所述损失确定子模块，用于：根据所述手指的状态、所述位置信息和所述位置权重，确定所述神经网络的手势预测结果的损失。
在本实施例中，根据手指的状态、位置信息和位置权重，对神经网络进行反向传播，可以减小手指的位置信息中位置坐标的取值对反向传播产生的不利影响，使得训练出的神经网络更加准确。
在一种可能的实现方式中，所述位置权重确定子模块，用于：当手指的状态为非伸出状态时，确定所述手指的位置权重为零权重。
在一种可能的实现方式中，当手指的状态为伸出状态时，可以确定所述手指的位置权重为非零权重；当手指的状态为非伸出状态时，可以确定所述手指的位置权重为零权重。
图10示出根据本公开实施例的手势处理装置的框图，如图10所示，所述装置包括：
图像获取模块1，用于获取图像；
手势获取模块2，用于上述手势识别装置中任一项所述的装置识别所述图像包括的手部的手势；
操作执行模块3，用于执行与手势的识别结果相应的控制操作。
在一种可能的实现方式中，可以通过拍摄装置拍摄所需要的图像，也可以通过各种类型的接收方式直接接收图像。可以根据本公开实施例中任一项所述的手势识别方法，在获取的图像中识别图像中包括的手部的手势。可以根据在图像中识别出的手势进行相应的控制操作。
在一种可能的实现方式中，所述操作执行模块，包括：控制指令获取子模块，用于根据预先确定的手势与控制指令之间的映射关系，获取与手势的识别结果相应的控制指令；操作执行子模块，用于根据所述控制指令控制电子设备执行相应操作。
在本实施例中，可以根据手势确定控制指令，可以根据需求通过建立手势与控制指令之间的映射关系，为图像中的手势确定丰富的控制指令。可以通过控制指令控制电子设备，达到控制车辆等各种装置的目的。
在一种可能的实现方式中，所述操作执行模块，包括：特效确定子模块，用于根据预先确定的手 势与特效之间的映射关系，确定与手势的识别结果相应的特效；特效执行子模块，用于在所述图像上采用计算机绘图的方式绘制所述特效。
在一种可能的实现方式中，所述特效执行子模块，用于：基于所述图像包括的手部或者手部的手指关键点，采用计算机绘图的方式绘制所述特效。
在本实施例中，可以根据手势确定与之对应的特效，通过在图像上增加特效，增加图像的表现能力。
可以理解，本公开提及的上述各个方法实施例，在不违背原理逻辑的情况下，均可以彼此相互结合形成结合后的实施例，限于篇幅，本公开不再赘述。
此外，本公开还提供了上述装置、电子设备、计算机可读存储介质、程序，上述均可用来实现本公开提供的任一种手势识别方法或手势处理方法，相应技术方案和描述和参见方法部分的相应记载，不再赘述。
本公开实施例还提出一种计算机可读存储介质，其上存储有计算机程序指令，所述计算机程序指令被处理器执行时实现上述任一方法实施例。计算机可读存储介质可以是非易失性计算机可读存储介质或易失性计算机可读存储介质。
本公开实施例还提出一种电子设备，包括：处理器和用于存储处理器可执行指令的存储器；其中，所述处理器通过调用所述可执行指令实现本公开任一方法实施例，具体工作过程以及设置方式均可以参照本公开上述相应方法实施例的具体描述，限于篇幅，在此不再赘述。
本公开实施例还提出一种计算机程序，所述计算机程序包括计算机可读代码，当所述计算机可读代码在电子设备中运行时，所述电子设备中的处理器执行本公开任一方法实施例。
图11是根据示例性实施例示出的一种电子设备800的框图。例如，电子设备800可以是移动电话，计算机，数字广播终端，消息收发设备，游戏控制台，平板设备，医疗设备，健身设备，个人数字助理等终端。
参照图11，电子设备800可以包括以下一个或多个组件：处理组件802，存储器804，电源组件806，多媒体组件808，音频组件810，输入/输出(I/O)的接口812，传感器组件814，以及通信组件816。
处理组件802通常控制电子设备800的整体操作，诸如与显示，电话呼叫，数据通信，相机操作和记录操作相关联的操作。处理组件802可以包括一个或多个处理器820来执行指令，以完成上述的方法的全部或部分步骤。此外，处理组件802可以包括一个或多个模块，便于处理组件802和其他组件之间的交互。例如，处理组件802可以包括多媒体模块，以方便多媒体组件808和处理组件802之间的交互。
存储器804被配置为存储各种类型的数据以支持在电子设备800的操作。这些数据的示例包括用于在电子设备800上操作的任何应用程序或方法的指令，联系人数据，电话簿数据，消息，图片，视频等。存储器804可以由任何类型的易失性或非易失性存储设备或者它们的组合实现，如静态随机存取存储器(SRAM)，电可擦除可编程只读存储器(EEPROM)，可擦除可编程只读存储器(EPROM)，可编程只读存储器(PROM)，只读存储器(ROM)，磁存储器，快闪存储器，磁盘或光盘。
电源组件806为电子设备800的各种组件提供电力。电源组件806可以包括电源管理系统，一个或多个电源，及其他与为电子设备800生成、管理和分配电力相关联的组件。
多媒体组件808包括在所述电子设备800和用户之间的提供一个输出接口的屏幕。在一些实施例中，屏幕可以包括液晶显示器(LCD)和触摸面板(TP)。如果屏幕包括触摸面板，屏幕可以被实现为触摸屏，以接收来自用户的输入信号。触摸面板包括一个或多个触摸传感器以感测触摸、滑动和触摸面板上的手势。所述触摸传感器可以不仅感测触摸或滑动动作的边界，而且还检测与所述触摸或滑 动操作相关的持续时间和压力。在一些实施例中，多媒体组件808包括一个前置摄像头和/或后置摄像头。当电子设备800处于操作模式，如拍摄模式或视频模式时，前置摄像头和/或后置摄像头可以接收外部的多媒体数据。每个前置摄像头和后置摄像头可以是一个固定的光学透镜系统或具有焦距和光学变焦能力。
音频组件810被配置为输出和/或输入音频信号。例如，音频组件810包括一个麦克风(MIC)，当电子设备800处于操作模式，如呼叫模式、记录模式和语音识别模式时，麦克风被配置为接收外部音频信号。所接收的音频信号可以被进一步存储在存储器804或经由通信组件816发送。在一些实施例中，音频组件810还包括一个扬声器，用于输出音频信号。
I/O接口812为处理组件802和外围接口模块之间提供接口，上述外围接口模块可以是键盘，点击轮，按钮等。这些按钮可包括但不限于：主页按钮、音量按钮、启动按钮和锁定按钮。
传感器组件814包括一个或多个传感器，用于为电子设备800提供各个方面的状态评估。例如，传感器组件814可以检测到电子设备800的打开/关闭状态，组件的相对定位，例如所述组件为电子设备800的显示器和小键盘，传感器组件814还可以检测电子设备800或电子设备800一个组件的位置改变，用户与电子设备800接触的存在或不存在，电子设备800方位或加速/减速和电子设备800的温度变化。传感器组件814可以包括接近传感器，被配置用来在没有任何的物理接触时检测附近物体的存在。传感器组件814还可以包括光传感器，如CMOS或CCD图像传感器，用于在成像应用中使用。在一些实施例中，该传感器组件814还可以包括加速度传感器，陀螺仪传感器，磁传感器，压力传感器或温度传感器。
通信组件816被配置为便于电子设备800和其他设备之间有线或无线方式的通信。电子设备800可以接入基于通信标准的无线网络，如WiFi，2G或3G，或它们的组合。在一个示例性实施例中，通信组件816经由广播信道接收来自外部广播管理系统的广播信号或广播相关信息。在一个示例性实施例中，所述通信组件816还包括近场通信(NFC)模块，以促进短程通信。例如，在NFC模块可基于射频识别(RFID)技术，红外数据协会(IrDA)技术，超宽带(UWB)技术，蓝牙(BT)技术和其他技术来实现。
在示例性实施例中，电子设备800可以被一个或多个应用专用集成电路(ASIC)、数字信号处理器(DSP)、数字信号处理设备(DSPD)、可编程逻辑器件(PLD)、现场可编程门阵列(FPGA)、控制器、微控制器、微处理器或其他电子元件实现，用于执行上述方法。
在示例性实施例中，还提供了一种非易失性计算机可读存储介质，例如包括计算机程序指令的存储器804，上述计算机程序指令可由电子设备800的处理器820执行以完成上述方法。
图12是根据示例性实施例示出的一种电子设备1900的框图。例如，电子设备1900可以被提供为一服务器。参照图12，电子设备1900包括处理组件1922，其进一步包括一个或多个处理器，以及由存储器1932所代表的存储器资源，用于存储可由处理组件1922的执行的指令，例如应用程序。存储器1932中存储的应用程序可以包括一个或一个以上的每一个对应于一组指令的模块。此外，处理组件1922被配置为执行指令，以执行上述方法。
电子设备1900还可以包括一个电源组件1926被配置为执行电子设备1900的电源管理，一个有线或无线网络接口1950被配置为将电子设备1900连接到网络，和一个输入输出(I/O)接口1958。电子设备1900可以操作基于存储在存储器1932的操作系统，例如Windows ServerTM，Mac OS XTM，UnixTM,LinuxTM，FreeBSDTM或类似。
在示例性实施例中，还提供了一种非易失性计算机可读存储介质，例如包括计算机程序指令的存 储器1932，上述计算机程序指令可由电子设备1900的处理组件1922执行以完成上述方法。
本公开可以是系统、方法和/或计算机程序产品。计算机程序产品可以包括计算机可读存储介质，其上载有用于使处理器实现本公开的各个方面的计算机可读程序指令。
计算机可读存储介质可以是可以保持和存储由指令执行设备使用的指令的有形设备。计算机可读存储介质例如可以是――但不限于――电存储设备、磁存储设备、光存储设备、电磁存储设备、半导体存储设备或者上述的任意合适的组合。计算机可读存储介质的更具体的例子(非穷举的列表)包括：便携式计算机盘、硬盘、随机存取存储器(RAM)、只读存储器(ROM)、可擦式可编程只读存储器(EPROM或闪存)、静态随机存取存储器(SRAM)、便携式压缩盘只读存储器(CD-ROM)、数字多功能盘(DVD)、记忆棒、软盘、机械编码设备、例如其上存储有指令的打孔卡或凹槽内凸起结构、以及上述的任意合适的组合。这里所使用的计算机可读存储介质不被解释为瞬时信号本身，诸如无线电波或者其他自由传播的电磁波、通过波导或其他传输媒介传播的电磁波(例如，通过光纤电缆的光脉冲)、或者通过电线传输的电信号。
这里所描述的计算机可读程序指令可以从计算机可读存储介质下载到各个计算/处理设备，或者通过网络、例如因特网、局域网、广域网和/或无线网下载到外部计算机或外部存储设备。网络可以包括铜传输电缆、光纤传输、无线传输、路由器、防火墙、交换机、网关计算机和/或边缘服务器。每个计算/处理设备中的网络适配卡或者网络接口从网络接收计算机可读程序指令，并转发该计算机可读程序指令，以供存储在各个计算/处理设备中的计算机可读存储介质中。
用于执行本公开操作的计算机程序指令可以是汇编指令、指令集架构(ISA)指令、机器指令、机器相关指令、微代码、固件指令、状态设置数据、或者以一种或多种编程语言的任意组合编写的源代码或目标代码，所述编程语言包括面向对象的编程语言—诸如Smalltalk、C++等，以及常规的过程式编程语言—诸如“C”语言或类似的编程语言。计算机可读程序指令可以完全地在用户计算机上执行、部分地在用户计算机上执行、作为一个独立的软件包执行、部分在用户计算机上部分在远程计算机上执行、或者完全在远程计算机或服务器上执行。在涉及远程计算机的情形中，远程计算机可以通过任意种类的网络—包括局域网(LAN)或广域网(WAN)—连接到用户计算机，或者，可以连接到外部计算机(例如利用因特网服务提供商来通过因特网连接)。在一些实施例中，通过利用计算机可读程序指令的状态信息来个性化定制电子电路，例如可编程逻辑电路、现场可编程门阵列(FPGA)或可编程逻辑阵列(PLA)，该电子电路可以执行计算机可读程序指令，从而实现本公开的各个方面。
这里参照根据本公开实施例的方法、装置(系统)和计算机程序产品的流程图和/或框图描述了本公开的各个方面。应当理解，流程图和/或框图的每个方框以及流程图和/或框图中各方框的组合，都可以由计算机可读程序指令实现。
这些计算机可读程序指令可以提供给通用计算机、专用计算机或其它可编程数据处理装置的处理器，从而生产出一种机器，使得这些指令在通过计算机或其它可编程数据处理装置的处理器执行时，产生了实现流程图和/或框图中的一个或多个方框中规定的功能/动作的装置。也可以把这些计算机可读程序指令存储在计算机可读存储介质中，这些指令使得计算机、可编程数据处理装置和/或其他设备以特定方式工作，从而，存储有指令的计算机可读介质则包括一个制造品，其包括实现流程图和/或框图中的一个或多个方框中规定的功能/动作的各个方面的指令。
也可以把计算机可读程序指令加载到计算机、其它可编程数据处理装置、或其它设备上，使得在计算机、其它可编程数据处理装置或其它设备上执行一系列操作步骤，以产生计算机实现的过程，从而使得在计算机、其它可编程数据处理装置、或其它设备上执行的指令实现流程图和/或框图中的一 个或多个方框中规定的功能/动作。
附图中的流程图和框图显示了根据本公开的多个实施例的系统、方法和计算机程序产品的可能实现的体系架构、功能和操作。在这点上，流程图或框图中的每个方框可以代表一个模块、程序段或指令的一部分，所述模块、程序段或指令的一部分包含一个或多个用于实现规定的逻辑功能的可执行指令。在有些作为替换的实现中，方框中所标注的功能也可以以不同于附图中所标注的顺序发生。例如，两个连续的方框实际上可以基本并行地执行，它们有时也可以按相反的顺序执行，这依所涉及的功能而定。也要注意的是，框图和/或流程图中的每个方框、以及框图和/或流程图中的方框的组合，可以用执行规定的功能或动作的专用的基于硬件的系统来实现，或者可以用专用硬件与计算机指令的组合来实现。
在不违背逻辑的情况下，本申请不同实施例之间可以相互结合，不同实施例描述有所侧重，为侧重描述的部分可以参见其他实施例的记载。
以上已经描述了本公开的各实施例，上述说明是示例性的，并非穷尽性的，并且也不限于所披露的各实施例。在不偏离所说明的各实施例的范围和精神的情况下，对于本技术领域的普通技术人员来说许多修改和变更都是显而易见的。本文中所用术语的选择，旨在最好地解释各实施例的原理、实际应用或对市场中的技术的改进，或者使本技术领域的其它普通技术人员能理解本文披露的各实施例。

Claims (45)

  1. 一种手势识别方法，其特征在于，所述方法包括：
    检测图像中手部包括的手指的状态；
    根据所述手指的状态确定所述手部的状态向量；
    根据所述手部的状态向量确定所述手部的手势。
  2. 根据权利要求1所述的方法，其特征在于，所述手指的状态表示所述手指相对于所述手部的掌根是否伸出和/或伸出程度的状态。
  3. 根据权利要求1或2所述的方法，其特征在于，所述根据所述手指的状态确定所述手部的状态向量，包括：
    根据所述手指的状态确定所述手指的状态值，其中，不同状态对应的手指的状态值不同；
    根据所述手指的状态值确定所述手部的状态向量。
  4. 根据权利要求1至3中任一项所述的方法，其特征在于，所述手指的状态包括以下一种或多种：伸出状态、非伸出状态、半伸出状态、弯曲状态。
  5. 根据权利要求1至4中任一项所述的方法，其特征在于，所述方法还包括：
    检测所述图像中手部包括的手指的位置信息；
    根据所述手指的位置信息，确定所述手部的位置向量；
    根据所述手部的状态向量确定所述手部的手势，包括：
    根据所述手部的状态向量和所述手部的位置向量，确定所述手部的手势。
  6. 根据权利要求5所述的方法，其特征在于，所述检测所述图像中手部包括的手指的位置信息，包括：
    检测所述图像中所述手部包括的手指的关键点，获得所述手指的关键点的位置信息；
    根据所述手指的位置信息，确定所述手部的位置向量，包括：
    根据所述手指的关键点的位置信息，确定所述手部的位置向量。
  7. 根据权利要求6所述的方法，其特征在于，所述检测所述图像中所述手部包括的手指的关键点，获得所述手指的关键点的位置信息，包括：
    检测所述图像中所述手部包括的状态为非伸出状态以外的手指的关键点，获得所述关键点的位置信息。
  8. 根据权利要求7所述的方法，其特征在于，所述关键点包括指尖和/或指关节。
  9. 根据权利要求1至8中任一项所述的方法，其特征在于，所述检测图像中手部包括的手指的状态，包括：
    将所述图像输入神经网络，经所述神经网络检测所述图像中手部包括的手指的状态。
  10. 根据权利要求9所述的方法，其特征在于，所述神经网络包括多个状态分支网络，所述经所述神经网络检测所述图像中手部包括的手指的状态，包括：
    经所述神经网络的不同状态分支网络分别检测所述图像中手部包括的不同手指的状态。
  11. 根据权利要求9或10所述的方法，其特征在于，所述神经网络还包括位置分支网络，所述检测所述图像中手部包括的手指的位置信息，包括：
    经所述神经网络的所述位置分支网络检测所述图像中所述手部包括的手指的位置信息。
  12. 根据权利要求9至11中任一项所述的方法，其特征在于，所述神经网络预先采用带有标注信息的样本图像训练而得，所述标注信息包括表示所述手指的状态的第一标注信息，和/或，表示所述手指的位置信息或关键点的位置信息的第二标注信息。
  13. 根据权利要求12所述的方法，其特征在于，所述样本图像中，不标注非伸出状态的手指的第二标注信息。
  14. 根据权利要求12或13所述的方法，其特征在于，所述第一标注信息包括由表示各手指的状态的第一标识值组成的状态向量；
    所述第二标注信息包括由标识各手指的位置信息或关键点的位置信息的第二标识值组成的位置向量。
  15. 根据权利要求9至14中任一项所述的方法，其特征在于，所述神经网络的训练步骤包括：
    将手部的样本图像输入神经网络得到手部中手指的状态；
    根据所述手指的状态确定手指的位置权重；
    根据所述手指的状态和所述位置权重，确定所述神经网络的手势预测结果的损失；
    向所述神经网络反向传播所述损失，以调整所述神经网络的网络参数。
  16. 根据权利要求15中所述的方法，其特征在于，所述将手部的样本图像输入神经网络得到手部中手指的状态，包括：
    将手部的样本图像输入神经网络得到手部中手指的状态和位置信息；
    根据所述手指的状态和所述位置权重，确定所述神经网络的手势预测结果的损失，包括：
    根据所述手指的状态、所述位置信息和所述位置权重，确定所述神经网络的手势预测结果的损失。
  17. 根据权利要求15或16所述的方法，其特征在于，所述根据所述手指的状态确定所述手指的位置权重，包括：
    当手指的状态为非伸出状态时，确定所述手指的位置权重为零权重。
  18. 一种手势处理方法，其特征在于，所述方法包括：
    获取图像；
    采用如权利要求1至17中任一项所述的方法识别所述图像包括的手部的手势；
    执行与手势的识别结果相应的控制操作。
  19. 根据权利要求18所述的方法，其特征在于，执行与手势的识别结果相应的操作控制，包括：
    根据预先确定的手势与控制指令之间的映射关系，获取与手势的识别结果相应的控制指令；
    根据所述控制指令控制电子设备执行相应操作。
  20. 根据权利要求18所述的方法，其特征在于，执行与手势的识别结果相应的操作控制，包括：
    根据预先确定的手势与特效之间的映射关系，确定与手势的识别结果相应的特效；
    在所述图像上采用计算机绘图的方式绘制所述特效。
  21. 根据权利要求20所述的方法，其特征在于，在所述图像上采用计算机绘图的方式绘制所述特效，包括：
    基于所述图像包括的手部或者手部的手指关键点，采用计算机绘图的方式绘制所述特效。
  22. 一种手势识别装置，其特征在于，所述装置包括：
    状态检测模块，用于检测图像中手部包括的手指的状态；
    状态向量获取模块，用于根据所述手指的状态确定所述手部的状态向量；
    手势确定模块，用于根据所述手部的状态向量确定所述手部的手势。
  23. 根据权利要求22所述的装置，其特征在于，所述手指的状态表示所述手指相对于所述手部的掌根是否伸出和/或伸出程度的状态。
  24. 根据权利要求22或23所述的装置，其特征在于，所述状态向量获取模块，包括：
    状态值获取子模块，用于根据所述手指的状态确定所述手指的状态值，其中，不同状态对应的手指的状态值不同；
    第一状态向量获取子模块，用于根据所述手指的状态值确定所述手部的状态向量。
  25. 根据权利要求22至24中任一项所述的装置，其特征在于，所述手指的状态包括以下一种或多种：伸出状态、非伸出状态、半伸出状态、弯曲状态。
  26. 根据权利要求22至25中任一项所述的装置，其特征在于，所述装置还包括：
    位置信息获取模块，用于检测所述图像中手部包括的手指的位置信息；
    位置向量获取模块，用于根据所述手指的位置信息，确定所述手部的位置向量；
    所述手势确定模块，包括：
    第一手势确定子模块，用于根据所述手部的状态向量和所述手部的位置向量，确定所述手部的手势。
  27. 根据权利要求26所述的装置，其特征在于，所述位置信息获取模块，包括：
    关键点检测子模块，用于检测所述图像中所述手部包括的手指的关键点，获得所述手指的关键点的位置信息；
    所述位置向量获取模块，包括：
    第一位置向量获取子模块，用于根据所述手指的关键点的位置信息，确定所述手部的位置向量。
  28. 根据权利要求27所述的装置，其特征在于，所述关键点检测子模块，用于：
    检测所述图像中所述手部包括的状态为非伸出状态以外的手指的关键点，获得所述关键点的位置信息。
  29. 根据权利要求28所述的装置，其特征在于，所述关键点包括指尖和/或指关节。
  30. 根据权利要求22至29中任一项所述的方法，其特征在于，所述状态检测模块，包括：
    第一状态检测子模块，用于将所述图像输入神经网络，经所述神经网络检测所述图像中手部包括的手指的状态。
  31. 根据权利要求30所述的装置，其特征在于，所述神经网络包括多个状态分支网络，所述第一状态检测子模块，用于：
    经所述神经网络的不同状态分支网络分别检测所述图像中手部包括的不同手指的状态。
  32. 根据权利要求30或31所述的装置，其特征在于，所述神经网络还包括位置分支网络，所述位置信息获取模块包括：
    第一位置信息获取子模块，用于经所述神经网络的所述位置分支网络检测所述图像中所述手部包括的手指的位置信息。
  33. 根据权利要求30至32中任一项所述的装置，其特征在于，所述神经网络预先采用带有标注信息的样本图像训练而得，所述标注信息包括表示所述手指的状态的第一标注信息，和/或，表示所述手指的位置信息或关键点的位置信息的第二标注信息。
  34. 根据权利要求33所述的装置，其特征在于，所述样本图像中，不标注非伸出状态的手指的第二标注信息。
  35. 根据权利要求33或34所述的装置，其特征在于，所述第一标注信息包括由表示各手指的状态的第一标识值组成的状态向量；
    所述第二标注信息包括由标识各手指的位置信息或关键点的位置信息的第二标识值组成的位置向量。
  36. 根据权利要求30至35中任一项所述的装置，其特征在于，所述神经网络包括训练模块，所述训练模块包括：
    状态获取子模块，用于将手部的样本图像输入神经网络得到手部中手指的状态；
    位置权重确定子模块，用于根据所述手指的状态确定手指的位置权重；
    损失确定子模块，用于根据所述手指的状态和所述位置权重，确定所述神经网络的手势预测结果的损失；
    反向传播子模块，用于向所述神经网络反向传播所述损失，以调整所述神经网络的网络参数。
  37. 根据权利要求36中所述的装置，其特征在于，所述状态获取子模块，用于：
    将手部的样本图像输入神经网络得到手部中手指的状态和位置信息；
    所述损失确定子模块，用于：
    根据所述手指的状态、所述位置信息和所述位置权重，确定所述神经网络的手势预测结果的损失。
  38. 根据权利要求36或37所述的装置，其特征在于，所述位置权重确定子模块，用于：
    当手指的状态为非伸出状态时，确定所述手指的位置权重为零权重。
  39. 一种手势处理装置，其特征在于，所述装置包括：
    图像获取模块，用于获取图像；
    手势获取模块，用于采用如权利要求22至38中任一项所述的装置识别所述图像包括的手部的手势；
    操作执行模块，用于执行与手势的识别结果相应的控制操作。
  40. 根据权利要求39所述的装置，其特征在于，所述操作执行模块，包括：
    控制指令获取子模块，用于根据预先确定的手势与控制指令之间的映射关系，获取与手势的识别结果相应的控制指令；
    操作执行子模块，用于根据所述控制指令控制电子设备执行相应操作。
  41. 根据权利要求39所述的装置，其特征在于，所述操作执行模块，包括：
    特效确定子模块，用于根据预先确定的手势与特效之间的映射关系，确定与手势的识别结果相应的特效；
    特效执行子模块，用于在所述图像上采用计算机绘图的方式绘制所述特效。
  42. 根据权利要求41所述的装置，其特征在于，所述特效执行子模块，用于：
    基于所述图像包括的手部或者手部的手指关键点，采用计算机绘图的方式绘制所述特效。
  43. 一种电子设备，其特征在于，包括：
    处理器；
    用于存储处理器可执行指令的存储器；
    其中，所述处理器通过调用所述可执行指令实现如权利要求1至21中任意一项所述的方法。
  44. 一种计算机可读存储介质，其上存储有计算机程序指令，其特征在于，所述计算机程序指令被处理器执行时实现权利要求1至21中任意一项所述的方法。
  45. 一种计算机程序，其特征在于，所述计算机程序包括计算机可读代码，当所述计算机可读代码在电子设备中运行时，所述电子设备中的处理器执行用于实现权利要求1至21中的任意一项所述的方法。
PCT/CN2019/092559 2018-08-17 2019-06-24 手势识别方法、手势处理方法及装置 WO2020034763A1 (zh)

Priority Applications (4)

Application Number Priority Date Filing Date Title
KR1020217007082A KR20210040435A (ko) 2018-08-17 2019-06-24 제스처 인식 방법, 제스처 처리 방법 및 장치
JP2021506277A JP7266667B2 (ja) 2018-08-17 2019-06-24 ジェスチャー認識方法、ジェスチャー処理方法及び装置
SG11202101142PA SG11202101142PA (en) 2018-08-17 2019-06-24 Gesture recognition method, and gesture processing method and apparatus
US17/166,238 US20210158031A1 (en) 2018-08-17 2021-02-03 Gesture Recognition Method, and Electronic Device and Storage Medium

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201810942882.1A CN110837766B (zh) 2018-08-17 2018-08-17 手势识别方法、手势处理方法及装置
CN201810942882.1 2018-08-17

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US17/166,238 Continuation US20210158031A1 (en) 2018-08-17 2021-02-03 Gesture Recognition Method, and Electronic Device and Storage Medium

Publications (1)

Publication Number Publication Date
WO2020034763A1 true WO2020034763A1 (zh) 2020-02-20

Family

ID=69525088

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/092559 WO2020034763A1 (zh) 2018-08-17 2019-06-24 手势识别方法、手势处理方法及装置

Country Status (6)

Country Link
US (1) US20210158031A1 (zh)
JP (1) JP7266667B2 (zh)
KR (1) KR20210040435A (zh)
CN (1) CN110837766B (zh)
SG (1) SG11202101142PA (zh)
WO (1) WO2020034763A1 (zh)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112527113A (zh) * 2020-12-09 2021-03-19 北京地平线信息技术有限公司 手势识别及手势识别网络的训练方法和装置、介质和设备
CN112947755A (zh) * 2021-02-24 2021-06-11 Oppo广东移动通信有限公司 手势控制方法与装置、电子设备及存储介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101577062A (zh) * 2008-12-30 2009-11-11 浙江工业大学 一种基于空间编码的手语运动信息与文本信息相互转换的实现方法
CN102368290A (zh) * 2011-09-02 2012-03-07 华南理工大学 一种基于手指高级特征的手势识别方法
US20150346831A1 (en) * 2014-05-28 2015-12-03 Kyocera Corporation Mobile electronic device and method
CN105868715A (zh) * 2016-03-29 2016-08-17 苏州科达科技股份有限公司 一种手势识别方法、装置及手势学习系统
CN108229277A (zh) * 2017-03-31 2018-06-29 北京市商汤科技开发有限公司 手势识别、控制及神经网络训练方法、装置及电子设备

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101536494B (zh) 2005-02-08 2017-04-26 奥布隆工业有限公司 用于基于姿势的控制系统的系统和方法
JP2014182662A (ja) 2013-03-19 2014-09-29 Stanley Electric Co Ltd 操作装置及び操作方法
CN106295612A (zh) * 2016-08-23 2017-01-04 广西科技大学 一种手部康复训练中手指运动的视觉监测方法
CN106709461B (zh) * 2016-12-28 2019-09-17 中国科学院深圳先进技术研究院 基于视频的行为识别方法及装置
CN108230352B (zh) * 2017-01-24 2021-02-26 北京市商汤科技开发有限公司 目标对象的检测方法、装置和电子设备
CN107563494B (zh) * 2017-08-01 2020-08-18 华南理工大学 一种基于卷积神经网络和热图的第一视角指尖检测方法
CN107808143B (zh) * 2017-11-10 2021-06-01 西安电子科技大学 基于计算机视觉的动态手势识别方法
CN108227912B (zh) * 2017-11-30 2021-05-11 北京市商汤科技开发有限公司 设备控制方法和装置、电子设备、计算机存储介质

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101577062A (zh) * 2008-12-30 2009-11-11 浙江工业大学 一种基于空间编码的手语运动信息与文本信息相互转换的实现方法
CN102368290A (zh) * 2011-09-02 2012-03-07 华南理工大学 一种基于手指高级特征的手势识别方法
US20150346831A1 (en) * 2014-05-28 2015-12-03 Kyocera Corporation Mobile electronic device and method
CN105868715A (zh) * 2016-03-29 2016-08-17 苏州科达科技股份有限公司 一种手势识别方法、装置及手势学习系统
CN108229277A (zh) * 2017-03-31 2018-06-29 北京市商汤科技开发有限公司 手势识别、控制及神经网络训练方法、装置及电子设备

Also Published As

Publication number Publication date
SG11202101142PA (en) 2021-03-30
JP2021534482A (ja) 2021-12-09
KR20210040435A (ko) 2021-04-13
CN110837766A (zh) 2020-02-25
CN110837766B (zh) 2023-05-05
US20210158031A1 (en) 2021-05-27
JP7266667B2 (ja) 2023-04-28

Similar Documents

Publication Publication Date Title
TWI747325B (zh) 目標對象匹配方法及目標對象匹配裝置、電子設備和電腦可讀儲存媒介
JP7238141B2 (ja) 顔と手を関連付けて検出する方法及び装置、電子機器、記憶媒体及びコンピュータプログラム
WO2021135601A1 (zh) 辅助拍照方法、装置、终端设备及存储介质
CN110889382A (zh) 虚拟形象渲染方法及装置、电子设备和存储介质
CN110991327A (zh) 交互方法及装置、电子设备和存储介质
WO2020078105A1 (zh) 姿势检测方法、装置、设备及存储介质
CN111241887B (zh) 目标对象关键点识别方法及装置、电子设备和存储介质
WO2023045207A1 (zh) 任务处理方法及装置、电子设备、存储介质和计算机程序
WO2020038108A1 (zh) 动态动作检测方法、动态动作控制方法及装置
US12008167B2 (en) Action recognition method and device for target object, and electronic apparatus
JP2022533065A (ja) 文字認識方法及び装置、電子機器並びに記憶媒体
CN112597944B (zh) 关键点检测方法及装置、电子设备和存储介质
CN111242303A (zh) 网络训练方法及装置、图像处理方法及装置
WO2020034763A1 (zh) 手势识别方法、手势处理方法及装置
CN110929616B (zh) 一种人手识别方法、装置、电子设备和存储介质
WO2022111458A1 (zh) 图像拍摄方法和装置、电子设备及存储介质
JP2024516570A (ja) データ処理方法、装置、コンピュータ機器及びコンピュータプログラム
CN111382598B (zh) 一种识别方法、装置和电子设备
CN113657173B (zh) 一种数据处理方法、装置和用于数据处理的装置
KR20230168568A (ko) 정보 입력 방법 및 장치, 전자 기기 및 저장 매체
CN115035440A (zh) 时序动作提名的生成方法及装置、电子设备和存储介质
CN114266305A (zh) 对象识别方法及装置、电子设备和存储介质
CN114020264A (zh) 算子处理方法及装置、电子设备和存储介质
CN111626398B (zh) 运算方法、装置及相关产品
CN113553946A (zh) 信息提示方法及装置、电子设备和存储介质

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: 19849873

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2021506277

Country of ref document: JP

Kind code of ref document: A

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS (EPO FORM 1205A DATED 26.05.2021)

122 Ep: pct application non-entry in european phase

Ref document number: 19849873

Country of ref document: EP

Kind code of ref document: A1