Image processing apparatus, method, processor, electronic device, and storage medium
By performing head detection and pose detection in parallel and combining convolutional neural network models, the computational efficiency and accuracy of object pose detection are improved in closed scenes, solving the problems of low computational efficiency and insufficient accuracy in existing technologies, and making it suitable for deployment on edge devices.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- BEIJING ESWIN COMPUTING TECH CO LTD
- Filing Date
- 2023-05-18
- Publication Date
- 2026-07-03
AI Technical Summary
In existing technologies, object pose detection has low computational efficiency and insufficient accuracy in video surveillance, security and video retrieval, especially in closed scenarios such as smart classrooms and smart recording studios.
We employ parallel execution of head detection and pose detection as independent operations, utilize a convolutional neural network model to perform head pose and intermediate pose detection separately, combine head and pose detection information to determine target pose information, and improve detection accuracy through overlap and matching information.
It improves the computational efficiency and accuracy of pose detection, meets the deployment requirements of edge devices, and can efficiently and accurately detect object pose in closed scenes.
Smart Images
Figure CN116543468B_ABST
Abstract
Description
Technical Field
[0001] This disclosure relates to the field of artificial intelligence technology, particularly to the fields of computer vision and deep learning technology, and can be applied to scenarios such as smart classrooms, smart recording and broadcasting, video monitoring, video retrieval, and visual security. More specifically, it relates to an image processing device, method, processor, electronic device, storage medium, and program product. Background Technology
[0002] Object pose detection has wide applications in video surveillance, security, and video retrieval. For example, object pose detection can include standing pose detection. However, pose detection suffers from low computational efficiency and low accuracy. Summary of the Invention
[0003] In view of the above problems, this disclosure provides an image processing apparatus, method, processor, electronic device, storage medium, and program product.
[0004] According to a first aspect of this disclosure, an image processing apparatus is provided, comprising:
[0005] The target video acquisition circuit is configured to acquire a target video, wherein the target video includes N target video frames, and N is an integer greater than or equal to 2;
[0006] The head pose information determination circuit is configured to perform head detection on the nth target video frame to obtain the nth head pose information, wherein the nth head pose information represents the first probability that a preset pose exists in the nth target video frame, n is an integer greater than 1 and less than or equal to N, and m is an integer greater than or equal to 1 and less than n.
[0007] The intermediate attitude information determination circuit is configured to perform attitude detection on the nth target video frame to obtain the nth intermediate attitude information, wherein the nth intermediate attitude information represents the second probability that a preset attitude exists in the nth target video frame;
[0008] The target pose information determination circuit is configured to determine the nth target pose information based on the nth head pose information and the nth intermediate pose information, wherein the nth target pose information represents the probability that a preset pose exists in the nth target video frame.
[0009] According to embodiments of this disclosure, the head pose information determination circuit is further configured to perform head detection on the nth target video frame to obtain the nth head pose information by:
[0010] Perform head detection on the nth target video frame to obtain the nth head position information;
[0011] Based on the nth head position information and at least one mth head position information, determine the nth position change information, and based on the nth position change information, determine the nth head pose information.
[0012] According to embodiments of this disclosure, the nth position change information includes the nth position change value;
[0013] The head pose information determination circuit is also configured to determine the nth head pose information based on the nth position change information through the following operations:
[0014] If the change value at the nth position is determined to be greater than or equal to the first preset threshold, the nth head pose information is determined to indicate that a preset pose exists in the nth target video frame.
[0015] According to embodiments of this disclosure, the intermediate pose information determination circuit is further configured to perform pose detection on the nth target video frame to obtain the nth intermediate pose information by:
[0016] Perform pose detection on the nth target video frame to obtain at least one nth intermediate pose detection box, and determine the nth intermediate pose information based on at least one nth intermediate pose detection box.
[0017] According to embodiments of this disclosure, the target pose information determination circuit is further configured to determine the nth intermediate pose information based on at least one nth intermediate pose detection box by the following operation:
[0018] Given that the nth head pose information indicates that a preset pose exists in the nth target video frame and the nth intermediate pose information indicates that a preset pose exists in the nth target video frame, the nth target pose information is determined based on the nth intermediate pose detection box and the nth head detection box corresponding to the nth head position information.
[0019] According to embodiments of this disclosure, the target pose information determination circuit is further configured to determine the nth target pose information based on the nth intermediate pose detection box and the nth head detection box corresponding to the nth head position information by the following operation:
[0020] If the overlap between the nth head detection box and the nth intermediate pose detection box is greater than or equal to the second preset threshold, the nth target pose information indicates that a preset pose exists in the nth target video frame.
[0021] According to embodiments of this disclosure, the target pose information determination circuit is further configured to determine the nth target pose information based on the nth intermediate pose detection box and the nth head detection box corresponding to the nth head position information by the following operation:
[0022] If the overlap between the nth head detection box and the nth intermediate pose detection box is greater than or equal to the second preset threshold, the nth head object corresponding to the nth head detection box and the nth pose object corresponding to the nth intermediate pose detection box are matched to obtain the nth matching information.
[0023] If the nth matching information indicates that the nth head object and the nth pose object match, then the nth target pose information indicates that a preset pose exists in the nth target video frame.
[0024] According to embodiments of this disclosure, the image processing apparatus further includes:
[0025] A raw video acquisition circuit is configured to acquire raw video, wherein the raw video comprises P raw video frames, where P is an integer greater than N; and
[0026] The target video frame determination circuit is configured to perform frame extraction processing on P original video frames to obtain N target video frames.
[0027] According to embodiments of this disclosure, the image processing apparatus further includes:
[0028] The circuit for determining the area to be played is configured to determine the nth area to be played from the nth target video frame when it is determined that the nth target attitude information indicates that a preset attitude exists in the nth target video frame;
[0029] The target playback area determination circuit is configured to determine the nth target playback area based on the nth to-be-played area and the (n-1)th target playback area; and
[0030] The playback circuit is configured to play the nth playback area.
[0031] According to embodiments of this disclosure, the target playback area determination circuit is further configured to determine the nth target playback area based on the nth to be played area and the (n-1)th target playback area by the following operation:
[0032] Based on the location information of the nth region to be played and the location information of the (n-1)th target playback region, determine the overlap between the nth region to be played and the (n-1)th target playback region; and
[0033] If the overlap between the nth region to be played and the (n-1)th target playback region is less than or equal to the third preset threshold, the nth region to be played is determined as the nth target playback region.
[0034] A second aspect of this disclosure provides an image processing method, comprising:
[0035] Obtain the target video, which consists of N target video frames, where N is an integer greater than or equal to 2;
[0036] Head detection is performed on the nth target video frame to obtain the nth head pose information, where the nth head pose information represents the first probability that a preset pose exists in the nth target video frame, n is an integer greater than 1 and less than or equal to N, and m is an integer greater than or equal to 1 and less than n.
[0037] The nth target video frame is subjected to pose detection to obtain the nth intermediate pose information, wherein the nth intermediate pose information represents the second probability that a preset pose exists in the nth target video frame;
[0038] Based on the nth head pose information and the nth intermediate pose information, the nth target pose information is determined, where the nth pose information represents the probability that a preset pose exists in the nth target video frame.
[0039] A third aspect of this disclosure provides a processor for executing a computer program to implement the image processing method described above.
[0040] A fourth aspect of this disclosure provides an electronic device comprising: one or more processors; and a memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors perform the image processing method described above.
[0041] A fifth aspect of this disclosure also provides a computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, cause the processor to perform the image processing method described above.
[0042] A sixth aspect of this disclosure also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described image processing method.
[0043] According to the image processing apparatus, method, processor, electronic device, storage medium, and program product provided in this disclosure, since head detection and behavior detection of the nth target video frame are independent detection operations, they can be executed in parallel, thus improving the computational efficiency of pose detection. Furthermore, since the pose information of the nth target is determined based on the nth head pose information and the nth intermediate pose information, the information used to determine the pose information of the nth target is richer, thus improving the accuracy of pose detection. Moreover, due to the improved computational efficiency, it can meet the needs of deployment in edge devices. Attached Figure Description
[0044] The foregoing contents, as well as other objects, features, and advantages of this disclosure, will become clearer from the following description of embodiments with reference to the accompanying drawings, in which:
[0045] Figure 1A schematic diagram of the structure of an image processing apparatus according to an embodiment of the present disclosure is shown.
[0046] Figure 2A This illustration schematically depicts head detection in the m-th target video frame according to an embodiment of the present disclosure.
[0047] Figure 2B This schematic diagram illustrates the head detection of the nth target video frame according to an embodiment of the present disclosure;
[0048] Figure 3 This schematic diagram illustrates the pose detection of the nth target video frame according to an embodiment of the present disclosure;
[0049] Figure 4 A schematic diagram of the structure of an image processing apparatus according to another embodiment of the present disclosure is shown.
[0050] Figure 5 A schematic diagram illustrating the nth region to be played and the (n-1)th target playback region according to an embodiment of the present disclosure;
[0051] Figure 6 A flowchart illustrating an image processing method according to an embodiment of the present disclosure is shown schematically.
[0052] Figure 7 A block diagram schematically illustrates a processor suitable for implementing an image processing method according to an embodiment of the present disclosure;
[0053] Figure 8 A block diagram schematically illustrates an electronic device suitable for implementing an image processing method according to an embodiment of the present disclosure. Detailed Implementation
[0054] To make the objectives, technical solutions, and advantages of the embodiments of this disclosure clearer, the technical solutions of the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this disclosure. Based on the described embodiments of this disclosure, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this disclosure. In the following description, some specific embodiments are for descriptive purposes only and should not be construed as limiting this disclosure in any way, but are merely examples of embodiments of this disclosure. Conventional structures or constructions will be omitted where they may cause confusion in understanding this disclosure. It should be noted that the shapes and dimensions of the components in the figures do not reflect actual size and proportion, but are only schematic representations of the contents of the embodiments of this disclosure.
[0055] Unless otherwise defined, the technical or scientific terms used in the embodiments of this disclosure shall have the ordinary meaning as understood by those skilled in the art. The terms "first," "second," and similar words used in the embodiments of this disclosure do not indicate any order, quantity, or importance, but are merely used to distinguish different components.
[0056] In the technical solutions disclosed herein, the collection, storage, use, processing, transmission, provision, disclosure, and application of data (including but not limited to user personal information) comply with the provisions of relevant laws and regulations, necessary confidentiality measures have been taken, and they do not violate public order and good morals.
[0057] In closed environments such as smart classrooms and intelligent recording studios, it is necessary to detect the posture of objects within the closed environment, which may present the following challenges.
[0058] On the one hand, high computing efficiency is required to meet the needs of devices deployed at the edge.
[0059] On the other hand, the uniformity of the objects' clothing and the high density of objects make detection difficult, making it hard to accurately detect the objects' posture.
[0060] In view of this, embodiments of the present disclosure provide an image processing apparatus, method, processor, electronic device, storage medium, and program product. According to embodiments of the present disclosure, since head detection and behavior detection of the nth target video frame are independent detection operations and can be executed in parallel, the computational efficiency of pose detection is improved. Furthermore, since the pose information of the nth target is determined based on the nth head pose information and the nth intermediate pose information, the information used to determine the pose information of the nth target is richer, thus improving the accuracy of pose detection. Moreover, due to the improved computational efficiency, it can meet the needs of deployment in edge devices.
[0061] Figure 1 A schematic diagram of the structure of an image processing apparatus according to an embodiment of the present disclosure is shown.
[0062] like Figure 1 As shown, the image processing device 100 includes a target video acquisition circuit 110, a head posture information determination circuit 120, an intermediate posture information determination circuit 130, and a target posture information determination circuit 140.
[0063] The target video acquisition circuit 110 is configured to acquire the target video.
[0064] The target video consists of N target video frames, where N is an integer greater than or equal to 2.
[0065] According to embodiments of this disclosure, the target video can be acquired using an image acquisition device. The image acquisition device may include a camera, a mobile phone with a shooting function, a webcam, etc.
[0066] According to embodiments of this disclosure, the target video may include moving images of an object. For example, in scenarios such as smart classrooms or intelligent recording and broadcasting, moving images of an object captured by an image acquisition device are used as the target video.
[0067] The head pose information determination circuit 120 is configured to perform head detection on the nth target video frame to obtain the nth head pose information.
[0068] According to embodiments of this disclosure, n can be an integer greater than 1 and less than or equal to N. m can be an integer greater than or equal to 1 and less than n. N and M can be configured according to actual business needs and are not limited herein.
[0069] According to embodiments of this disclosure, the preset posture may include one of a standing posture, a sitting posture, and a held posture.
[0070] According to embodiments of this disclosure, the nth head pose information can characterize a first probability that a preset pose exists in the nth target video frame. For example, if the first probability is greater than or equal to a first preset probability threshold, the nth head pose information can characterize that a preset pose exists in the nth target video frame. If the first probability is less than the first preset probability threshold, the nth head pose information can characterize that a preset pose does not exist in the nth target video frame. The first preset probability threshold can be configured according to actual business needs and is not limited herein. For example, the first preset probability threshold can be 0.8.
[0071] According to embodiments of this disclosure, the nth target video frame can be input into a pre-trained head detection model for head detection, and the nth head position information can be output. Alternatively, the head contour of the object can be determined based on the pixel values in the nth target video frame, and the nth head position information can be obtained based on the head contour.
[0072] According to embodiments of this disclosure, the head detection model may be obtained by training a Convolutional Neural Network (CNN) model or a Deep Convolutional Neural Network (DCNN) model, and it is understood that this disclosure is not limited thereto.
[0073] According to embodiments of this disclosure, the pose change of an object in the nth target video frame is determined based on the nth head position information and at least one mth head position information, thereby determining the nth head pose information.
[0074] The intermediate attitude information determination circuit 130 is configured to perform attitude detection on the nth target video frame to obtain the nth intermediate attitude information.
[0075] According to embodiments of this disclosure, the nth intermediate pose information can characterize a second probability that a preset pose exists in the nth target video frame. For example, if the second probability is greater than or equal to a second preset probability threshold, the nth intermediate pose information can characterize that a preset pose exists in the nth target video frame. If the second probability is less than the second preset probability threshold, the nth intermediate pose information can characterize that a preset pose does not exist in the nth target video frame. The second preset probability threshold can be configured according to actual business needs and is not limited herein. The second preset probability threshold can be the same as or different from the first preset probability threshold. For example, the second preset probability threshold can be 0.8. The second preset probability threshold can also be 0.85.
[0076] According to embodiments of this disclosure, the nth target video frame can be input into a pre-trained pose detection model for pose detection, and the nth intermediate pose information can be output. Alternatively, the overall contour of the object can be determined based on the pixel values in the nth target video frame, and the nth intermediate pose information can be obtained based on the overall contour.
[0077] According to embodiments of this disclosure, the pose detection model may be obtained by training a CNN model or a DCNN model, and it is understood that this disclosure is not limited thereto.
[0078] The target attitude information determination circuit 140 is configured to determine the nth target attitude information based on the nth head attitude information and the nth intermediate attitude information.
[0079] According to embodiments of this disclosure, the nth target pose information can characterize the probability that a preset pose exists in the nth target video frame.
[0080] According to embodiments of this disclosure, if both the nth head pose information and the nth intermediate pose information indicate the presence of a preset pose in the nth target video frame, then the nth target pose information is determined to indicate the presence of a preset pose in the nth target video frame. If at least one of the nth head pose information and the nth intermediate pose information indicates the absence of a preset pose in the nth target video frame, then the nth target information is determined to indicate the absence of a preset pose in the nth target video frame.
[0081] Alternatively, if both the nth head pose information and the nth intermediate pose information indicate that a preset pose exists in the nth target video frame, the nth target pose information is determined based on the nth head position information.
[0082] According to embodiments of this disclosure, since head detection and behavior detection of the nth target video frame are independent detection operations, they can be executed in parallel, thus improving the computational efficiency of pose detection. Furthermore, since the pose information of the nth target is determined based on the nth head pose information and the nth intermediate pose information, the information used to determine the pose information of the nth target is richer, thus improving the accuracy of pose detection. Moreover, due to the improved computational efficiency, it can meet the requirements for deployment in edge devices.
[0083] According to embodiments of this disclosure, in order to perform head detection on the nth target video frame and obtain the nth head pose information, the head pose information determination circuit 120 can also be configured to perform head detection on the nth target video frame, obtain the nth head position information, determine the nth position change information based on the nth head position information and at least one mth head position information, and determine the nth head pose information based on the nth position change information.
[0084] According to embodiments of this disclosure, the nth header position information may include the nth position information of at least one nth object in the nth target video frame. The mth header position information may include the mth position information of at least one mth object in the mth target video frame. The nth object and the mth object may be the same or different. At least one mth header position information may include one or more mth header position information. The mth target video frame may be a target video frame preceding the nth target video frame.
[0085] For example, n can be 10, and the nth head position information is the 10th head position information of the 10th target video frame. m can be any integer from 1 to 9, and at least one mth head position information can include the 9th head position information of the 9th target video frame, or it can include the 7th head position information of the 7th target video frame and the 5th head position information of the 5th target video frame.
[0086] According to embodiments of this disclosure, head detection can be performed on the nth target video frame to obtain the nth head detection box, and the position information of the nth head can be determined based on the nth head detection box. Head detection can be performed on the mth target video frame to obtain the mth head detection box, and the position information of the mth head can be determined based on the mth head detection box.
[0087] According to embodiments of this disclosure, the nth position change information can characterize the position change information of the nth head object from the mth target video frame to the nth target video frame.
[0088] According to embodiments of this disclosure, the position information of the nth head and the position change information between each of the at least one mth head position information can be determined to obtain at least one position change information corresponding to the nth target video frame. The nth position change information is determined based on the at least one position change information corresponding to the nth target video frame. For example, the average change information of the at least one position change information corresponding to the nth target video frame can be determined, and the average change information corresponding to the nth target video frame is determined as the nth position change information. Alternatively, the maximum position change information can be determined from the at least one position change information corresponding to the nth target video frame, and the maximum position change information corresponding to the nth target video frame is determined as the nth position change information. Alternatively, the minimum position change information can be determined from the at least one position change information corresponding to the nth target video frame, and the minimum position change information corresponding to the nth target video frame is determined as the nth position change information.
[0089] Figure 2A The illustration shows a head detection diagram of the m-th target video frame according to an embodiment of the present disclosure.
[0090] like Figure 2A As shown, in the m-th target video frame, head detection can be performed on the m-th target video frame to obtain at least one m-th head detection box. At least one m-th head detection box may include m-th head detection box 201_m, m-th head detection box 202_m, and m-th head detection box 203_m. m-th head detection box 201_m may correspond to m-th object 211_m. m-th head detection box 202_m may correspond to m-th object 212_m. m-th head detection box 203_m may correspond to m-th object 213_m.
[0091] According to embodiments of this disclosure, the position information of the m-th head can be determined based on the position information of the m-th head detection box in the m-th target video frame.
[0092] Figure 2B A schematic diagram illustrating head detection of the nth target video frame according to an embodiment of the present disclosure is shown.
[0093] like Figure 2B As shown, in the nth target video frame, head detection can be performed on the nth target video frame to obtain at least one nth head detection box. The at least one nth head detection box may include nth head detection box 201_n, nth head detection box 202_n, and nth head detection box 203_n. nth head detection box 201_n may correspond to nth object 211_n. nth head detection box 202_n may correspond to nth object 212_n. nth head detection box 203_n may correspond to nth object 213_n.
[0094] According to embodiments of this disclosure, the position information of the nth head can be determined based on the position information of the nth head detection box in the nth target video frame.
[0095] According to embodiments of this disclosure, the nth object in the nth target video frame and the mth object in the mth target video frame can be matched to determine the mth object in the mth target video frame that matches the nth object. Based on the position information of the nth object in the nth target video frame and the position information of the mth object in the mth target video frame that matches the nth object, the nth position change information of the nth object in the nth target video frame is determined.
[0096] For example, Figure 2A The m-th object 211_m and Figure 2B If the nth object 211_n is matched, the nth position change information of the nth object 211_n in the nth target video frame can be determined based on the nth head detection box 201_n and the mth head detection box 201_m.
[0097] Figure 2A The m-th object 212_m and Figure 2B The nth object 212_n is matched with the nth head detection box 202_n and the mth head detection box 202_m. The nth position change information of the nth object 212_n in the nth target video frame can be determined based on the nth head detection box 202_n and the mth head detection box 202_m.
[0098] Figure 2A The m-th object 213_m and Figure 2B If the nth object 213_n is matched, the nth position change information of the nth object 213_n in the nth target video frame can be determined based on the nth head detection box 203_n and the mth head detection box 203_m.
[0099] Depend on Figure 2A and Figure 2B It can be seen that since the position information of the nth head detection box 201_n in the nth target video frame is consistent with the position information of the mth head detection box 201_m in the mth target video frame, it can be determined that the nth position change information of the nth object 211_n in the nth target video frame is zero. Therefore, it can be determined that the nth head pose information of the nth object 211_n in the nth target video frame indicates that the nth object 211_n in the nth target video frame does not have a preset pose.
[0100] Since the position information of the nth head detection box 202_n in the nth target video frame is inconsistent with the position information of the mth head detection box 202_m in the mth target video frame, it can be determined that the nth position change information of the nth object 212_n in the nth target video frame is not zero. Therefore, it can be determined that the nth head pose information of the nth object 212_n in the nth target video frame indicates that the nth object 212_n has a preset pose in the nth target video frame.
[0101] Since the position information of the nth head detection box 203_n in the nth target video frame is consistent with the position information of the mth head detection box 203_m in the mth target video frame, it can be determined that the nth position change information of the nth object 213_n in the nth target video frame is zero. Therefore, it can be determined that the nth head pose information of the nth object 213_n in the nth target video frame indicates that the nth object 213_n does not have a preset pose in the nth target video frame.
[0102] According to embodiments of this disclosure, the changes in the nth head object can be determined from the nth position change information, thereby enabling a more accurate determination of the nth head pose information.
[0103] According to an embodiment of this disclosure, in order to determine the nth head pose information based on the nth position change information, the head pose information determination circuit 120 can also be configured to determine that the nth head pose information represents the existence of a preset pose in the nth target video frame when the nth position change value is determined to be greater than or equal to a first preset threshold.
[0104] According to embodiments of this disclosure, the nth position change information may include the nth position change value. The nth position change value can characterize the magnitude of the position change of the nth object corresponding to the nth position change information.
[0105] According to embodiments of this disclosure, the nth position change value can be obtained by subtracting the height of the mth object in the mth head position information from the height of the nth object in the nth head position information. Therefore, when the nth position change value is positive, it indicates that the height of the nth object in the nth target video frame has increased, meaning the nth object is standing. When the nth position change value is zero, it indicates that the height of the nth object in the nth target video frame remains unchanged, meaning the position of the nth object remains unchanged. When the nth position change value is negative, it indicates that the height of the nth object in the nth target video frame has decreased, meaning the nth object is sitting.
[0106] According to embodiments of this disclosure, a first preset threshold can be set, and the change value at the nth position can be compared with the first preset threshold. The first preset threshold can be the height difference between when the object is sitting and when it is standing.
[0107] According to the embodiments of this disclosure, when it is determined that the change value of the nth position is greater than or equal to the first preset threshold, that is, the position change of the nth object is large, it can indicate that the nth object has stood up, that is, the nth object is in a standing posture, and the nth head posture information represents that there is a preset posture in the nth target video frame.
[0108] According to embodiments of this disclosure, if the change value at the nth position is determined to be less than a first preset threshold, meaning the position change of the nth object is small, it can be indicated that the nth object remains unchanged, that is, the pose of the nth object remains unchanged. In this case, the nth head pose information indicates that there is no preset pose in the nth target video frame.
[0109] According to embodiments of this disclosure, if the change value at the nth position is determined to be greater than or equal to a first preset threshold, a first probability can be determined to be greater than or equal to a first preset probability threshold. If the first probability is greater than or equal to the first preset probability threshold, the nth head pose information can characterize the presence of a preset pose in the nth target video frame. If the change value at the nth position is determined to be less than the first preset threshold, a first probability can be determined to be less than the first preset probability threshold. If the first probability is less than the first preset probability threshold, the nth head pose information can characterize the absence of a preset pose in the nth target video frame.
[0110] According to embodiments of this disclosure, by setting a first preset threshold, the nth head pose information can be divided, which can quickly determine the nth head pose information that represents the preset pose in the nth target video frame, and then perform subsequent processing based on the nth head pose information, thereby improving the efficiency of data processing.
[0111] According to an embodiment of this disclosure, in order to perform attitude detection on the nth target video frame and obtain the nth intermediate attitude information through the following operations, the intermediate attitude information determination circuit 130 can also be configured to perform attitude detection on the nth target video frame, obtain at least one nth intermediate attitude detection box, and determine the nth intermediate attitude information based on the at least one nth intermediate attitude detection box.
[0112] Figure 3 The illustration shows a schematic diagram of pose detection for the nth target video frame according to an embodiment of the present disclosure.
[0113] like Figure 3 As shown, in the nth target video frame, pose detection can be performed on the nth target video frame to obtain at least one nth intermediate pose detection box. The at least one nth intermediate pose detection box includes nth intermediate pose detection box 301_n, nth intermediate pose detection box 302_n and nth intermediate pose detection box 303_n.
[0114] According to embodiments of this disclosure, the nth intermediate pose information can be determined based on the nth pose object in the nth intermediate pose detection frame.
[0115] For example, in Figure 3 In the nth intermediate pose detection box 301_n, the nth pose object is a sitting pose. Based on the nth intermediate pose detection box 301_n, the nth intermediate pose information indicates that a sitting pose exists in the nth target video frame. Similarly, the nth pose object in the nth intermediate pose detection box 302_n is a standing pose. Based on the nth intermediate pose detection box 302_n, the nth intermediate pose information indicates that a standing pose exists in the nth target video frame. Likewise, the nth pose object in the nth intermediate pose detection box 303_n is a standing pose. Based on the nth intermediate pose detection box 303_n, the nth intermediate pose information indicates that a standing pose exists in the nth target video frame.
[0116] According to an embodiment of this disclosure, in order to determine the nth intermediate pose information based on at least one nth intermediate pose detection box, the target pose information determination circuit 140 can also be configured to determine the nth target pose information based on the nth intermediate pose detection box and the nth head detection box corresponding to the nth head position information when it is determined that the nth head pose information indicates that a preset pose exists in the nth target video frame and the nth intermediate pose information indicates that a preset pose exists in the nth target video frame.
[0117] According to embodiments of this disclosure, when it is determined that the nth head pose information indicates the existence of a preset pose in the nth target video frame and the nth intermediate pose information indicates the existence of a preset pose in the nth target video frame, the positions of the preset pose in the nth target video frame of the nth intermediate pose information and the preset pose in the nth target video frame of the nth head pose information can be determined by the nth intermediate pose detection box and the nth head detection box, thereby determining whether the preset pose exists in the nth target video frame of the nth target pose information.
[0118] For example, in Figure 2B In the context of the nth head detection box 202_n, the nth head pose information represents the presence of a preset pose in the nth target video frame. Figure 3 In the nth target video frame, a preset pose exists, corresponding to the nth intermediate pose detection box 302_n and the nth intermediate pose detection box 303_n. Therefore, the nth target pose information can be determined based on the nth head detection box 202_n, the nth intermediate pose detection box 302_n, and the nth intermediate pose detection box 303_n to indicate that a preset pose exists in the nth target video frame.
[0119] According to an embodiment of this disclosure, in order to determine the nth target pose information based on the nth intermediate pose detection box and the nth head detection box corresponding to the nth head position information, the target pose information determination circuit 140 can also be configured to determine that the nth target pose information represents the existence of a preset pose in the nth target video frame when the overlap between the nth head detection box and the nth intermediate pose detection box is greater than or equal to a second preset threshold.
[0120] According to the embodiments of this disclosure, the overlap between the nth head detection box and the nth intermediate pose detection box is greater than or equal to the second preset threshold, that is, the overlap between the region where the nth head detection box is located and the region where the nth intermediate pose detection box is located is high. Therefore, it can be determined that the nth target pose information represents the existence of a preset pose in the nth target video frame.
[0121] According to embodiments of this disclosure, by determining the overlap between the nth head detection box and the nth intermediate pose detection box, the preset pose of the nth target video frame in the nth head pose information and the preset pose in the nth intermediate pose information can be in the same area, further improving the accuracy of pose detection.
[0122] According to an embodiment of this disclosure, in order to determine the nth target pose information based on the nth intermediate pose detection box and the nth head detection box corresponding to the nth head position information, the target pose information determination circuit 140 can be further configured to, when it is determined that the overlap between the nth head detection box and the nth intermediate pose detection box is greater than or equal to a second preset threshold, match the nth head object corresponding to the nth head detection box and the nth pose object corresponding to the nth intermediate pose detection box to obtain nth matching information. When it is determined that the nth matching information indicates that the nth head object and the nth pose object match, it is determined that the nth target pose information indicates that a preset pose exists in the nth target video frame.
[0123] According to embodiments of this disclosure, since head detection and pose detection are performed separately, there may be a situation where the nth head object corresponding to the nth head detection box is not the same as the nth pose object corresponding to the nth intermediate pose detection box. By matching the nth head object corresponding to the nth head detection box with the nth pose object corresponding to the nth intermediate pose detection box, the accuracy of pose detection can be improved, ensuring that the nth head object corresponding to the nth head detection box and the nth pose object corresponding to the nth intermediate pose detection box are the same object.
[0124] According to the embodiments of this disclosure, matching the nth head object corresponding to the nth head detection box and the nth pose object corresponding to the nth intermediate pose detection box can ensure that the nth head object and the nth pose object are the same object, thereby further improving the accuracy of the existence of a preset pose in the nth target pose information.
[0125] Figure 4 A schematic diagram of the structure of an image processing apparatus according to another embodiment of the present disclosure is shown.
[0126] like Figure 4 As shown, the image processing device includes a raw video acquisition circuit 410, a target video frame determination circuit 420, a target video acquisition circuit 430, a head posture information determination circuit 440, an intermediate posture information determination circuit 450, a target posture information determination circuit 460, a playback area determination circuit 470, a target playback area determination circuit 480, and a playback circuit 490.
[0127] The raw video acquisition circuit 410 is configured to acquire raw video. The raw video consists of P raw video frames, where P is an integer greater than N.
[0128] The target video frame determination circuit 420 is configured to perform frame extraction processing on P original video frames to obtain N target video frames.
[0129] According to embodiments of this disclosure, the original video can be a video directly acquired from an image acquisition device, and the P original video frames are arranged according to the acquisition order.
[0130] According to embodiments of this disclosure, P original video frames are subjected to frame extraction processing to obtain N target video frames. This can be done by extracting one original video frame from every other original video frame in the P original video frames, or by extracting one original video frame from every two original video frames. No limitation is made here.
[0131] According to embodiments of this disclosure, frame extraction is performed on the original video frames to obtain N target video frames, which reduces the amount of video frames processed. Furthermore, since object activities are relatively simple in scenarios such as smart classrooms or smart broadcasting, frame extraction does not affect the accuracy of the processing results.
[0132] According to embodiments of this disclosure, the target video acquisition circuit 430 can correspond to the operation performed by the target video acquisition circuit 110, the head posture information determination circuit 440 can correspond to the operation performed by the head posture information determination circuit 120, the intermediate posture information determination circuit 450 can correspond to the operation performed by the intermediate posture information determination circuit 130, and the target posture information determination circuit 460 can correspond to the operation performed by the target posture information determination circuit 140, which will not be described in detail here.
[0133] The playback area determination circuit 470 is configured to determine the nth playback area from the nth target video frame when the nth target attitude information indicates that a preset attitude exists in the nth target video frame.
[0134] The target playback area determination circuit 480 is configured to determine the nth target playback area based on the nth to be played area and the (n-1)th target playback area.
[0135] The playback circuit 490 is configured to play the nth playback area.
[0136] According to embodiments of this disclosure, when it is determined that the nth target pose information represents a preset pose in the nth target video frame, the nth playback area can be determined based on the nth head object representing a standing pose in the nth target video frame.
[0137] According to embodiments of this disclosure, the nth playback area can be centered on the nth head detection box corresponding to the nth head object in a standing posture, extending upwards by the height of one head detection box, downwards by the height of two head detection boxes, and left and right by the width of two head detection boxes each. This allows the nth head object to be located in the upper middle position of the nth playback area.
[0138] According to embodiments of this disclosure, the (n-1)th target playback area can be the playback area in the previous target video frame. Since the target video frame is obtained by extracting frames from the original video frame, the original video frame that has not been extracted can be played using the target playback area in the previous target video frame adjacent to the original video frame.
[0139] According to embodiments of this disclosure, when the (n-1)th target playback area and the nth target playback area are different, the playback area can be smoothed. Smoothing can be achieved by playing the nth playback area by slowly transitioning from the (n-1)th target playback area to the nth target playback area, or by moving to the nth target playback area at equal intervals, resulting in a smooth and non-abrupt change in the image.
[0140] According to embodiments of this disclosure, by determining the nth target playback area and playing the nth target playback area, it is possible to switch the playback of the nth playback area with a preset posture in closed scenarios such as smart classrooms and intelligent recording and broadcasting, which helps to evaluate the quality of classroom teaching and student learning.
[0141] According to an embodiment of this disclosure, in order to determine the nth target playback area based on the nth to-be-played area and the (n-1)th target playback area, the target playback area determination circuit 480 can also be configured to determine the overlap between the nth to-be-played area and the (n-1)th target playback area based on the position information of the nth to-be-played area and the position information of the (n-1)th target playback area. If the overlap between the nth to-be-played area and the (n-1)th target playback area is determined to be less than or equal to a third preset threshold, the nth to-be-played area is determined as the nth target playback area.
[0142] According to embodiments of this disclosure, the position information of the nth playback area can be determined by taking the lower left corner of the nth target video frame as the origin coordinate and the coordinates of the upper left corner of the nth playback area. The position information of the (n-1)th target playback area is similar and will not be described again here.
[0143] According to embodiments of this disclosure, the overlap between the nth playback area and the (n-1)th target playback area can characterize the distance between objects at the center of the screen in the two playback areas.
[0144] According to the embodiments of this disclosure, a third preset threshold can be set. When the overlap between the nth playback area and the (n-1)th target playback area is greater than the third preset threshold, that is, the objects in the center of the screen in the two playback areas are close to each other, the objects in the center of the screen in the nth playback area can be in the (n-1)th target playback area. Therefore, the (n-1)th target playback area can continue to be used as the nth target playback area, and the screen does not need to be switched.
[0145] According to the embodiments of this disclosure, when the overlap between the nth playback area and the (n-1)th target playback area is less than or equal to a third preset threshold, that is, the objects at the center of the screen in the two playback areas are far apart, and the object at the center of the screen in the nth playback area is not in the (n-1)th target playback area, or is located at a relatively edge position in the (n-1)th target playback area, the nth playback area can be determined as the nth target playback area.
[0146] Figure 5 A schematic diagram illustrating the nth region to be played and the (n-1)th target playback region according to an embodiment of the present disclosure is shown.
[0147] like Figure 5 As shown, the overlap between the (n-1)th target playback area 501_n-1 and the nth playback area 501_n can be less than the third preset threshold. Since the object in the (n-1)th target playback area 501_n-1 is located at a relatively edge position within the (n-1)th target playback area, the nth playback area 501_n can be determined as the nth target playback area.
[0148] Figure 6A flowchart illustrating an image processing method according to an embodiment of the present disclosure is shown schematically.
[0149] like Figure 6 As shown, the image processing method includes operations S610 to S640.
[0150] During operation of S610, the target video is acquired. The target video consists of N target video frames, where N is an integer greater than or equal to 2.
[0151] According to an embodiment of this disclosure, operation S610 is performed by the target video acquisition circuit 110, which corresponds to the operation performed by the target video acquisition circuit 110 described above, and will not be repeated here for the sake of simplicity.
[0152] In operation S620, head detection is performed on the nth target video frame to obtain the nth head pose information. The nth head pose information represents the first probability that a preset pose exists in the nth target video frame, where n is an integer greater than 1 and less than or equal to N, and m is an integer greater than or equal to 1 and less than n.
[0153] According to an embodiment of this disclosure, operation S620 is performed by the head posture information determination circuit 120, which corresponds to the operation performed by the head posture information determination circuit 120 described above, and will not be repeated here for the sake of simplicity.
[0154] In operation S630, pose detection is performed on the nth target video frame to obtain the nth intermediate pose information. The nth intermediate pose information represents the second probability that a preset pose exists in the nth target video frame.
[0155] According to an embodiment of this disclosure, operation S630 is performed by intermediate attitude information determination circuit 130, which corresponds to the operation performed by the intermediate attitude information determination circuit 130 described above, and will not be repeated here for the sake of simplicity.
[0156] In operation S640, the nth target pose information is determined based on the nth head pose information and the nth intermediate pose information. The nth pose information represents the probability that a preset pose exists in the nth target video frame.
[0157] According to an embodiment of this disclosure, operation S640 is executed by the target attitude information determination circuit 140, which corresponds to the operation executed by the target attitude information determination circuit 140 described above, and will not be repeated here for the sake of simplicity.
[0158] According to embodiments of this disclosure, any plurality of modules in the target video acquisition circuit 110, head pose information determination circuit 120, intermediate pose information determination circuit 130, and target pose information determination circuit 140 can be combined into one module, or any one of these modules can be split into multiple modules. Alternatively, at least part of the functionality of one or more of these modules can be combined with at least part of the functionality of other modules and implemented in one module. According to embodiments of this disclosure, at least one of the target video acquisition circuit 110, head pose information determination circuit 120, intermediate pose information determination circuit 130, and target pose information determination circuit 140 can be at least partially implemented as hardware circuitry, such as a field-programmable gate array (FPGA), a programmable logic array (PLA), a system-on-a-chip, a system-on-a-substrate, a system-on-package, an application-specific integrated circuit (ASIC), or implemented in hardware or firmware by any other reasonable means of integrating or packaging the circuitry, or implemented in any one of software, hardware, and firmware methods, or in a suitable combination of any of these methods. Alternatively, at least one of the target video acquisition circuit 110, head pose information determination circuit 120, intermediate pose information determination circuit 130, and target pose information determination circuit 140 can be at least partially implemented as a computer program module, which can perform corresponding functions when the computer program module is run.
[0159] Figure 7 A block diagram schematically illustrates a processor suitable for implementing an image processing method according to an embodiment of the present disclosure.
[0160] like Figure 7 As shown, the processor 700 is used to execute computer programs to implement the image processing method of the present disclosure embodiments. The processor 700 may include a central processing unit (CPU), a graphics processing unit (GPU), or a neural network processing unit (NPU).
[0161] Figure 8 A block diagram schematically illustrates an electronic device suitable for implementing an image processing method according to an embodiment of the present disclosure.
[0162] like Figure 8As shown, an electronic device 800 according to an embodiment of this disclosure includes a processor 801, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 802 or a program loaded from a storage portion 808 into a random access memory (RAM) 803. The processor 801 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or an associated chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 801 may also include onboard memory for caching purposes. The processor 801 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of this disclosure.
[0163] RAM 803 stores various programs and data required for the operation of electronic device 800. Processor 801, ROM 802, and RAM 803 are interconnected via bus 804. Processor 801 performs various operations of the method flow according to embodiments of the present disclosure by executing programs in ROM 802 and / or RAM 803. It should be noted that the programs may also be stored in one or more memories other than ROM 802 and RAM 803. Processor 801 may also perform various operations of the method flow according to embodiments of the present disclosure by executing programs stored in said one or more memories.
[0164] According to embodiments of this disclosure, the electronic device 800 may further include an input / output (I / O) interface 805, which is also connected to a bus 804. The electronic device 800 may also include one or more of the following components connected to the input / output (I / O) interface 805: an input section 806 including a keyboard, mouse, etc.; an output section 807 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 808 including a hard disk, etc.; and a communication section 809 including a network interface card such as a LAN card, modem, etc. The communication section 809 performs communication processing via a network such as the Internet. A drive 810 is also connected to the input / output (I / O) interface 805 as needed. A removable medium 811, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on the drive 810 as needed so that computer programs read from it can be installed into the storage section 808 as needed.
[0165] This disclosure also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments; or it may exist independently and not assembled into the device / apparatus / system. The computer-readable storage medium carries one or more programs, which, when executed, implement the image processing method according to the embodiments of this disclosure.
[0166] According to embodiments of this disclosure, the computer-readable storage medium may be a non-volatile computer-readable storage medium, such as including, but not limited to: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this disclosure, the computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. For example, according to embodiments of this disclosure, the computer-readable storage medium may include ROM 802 and / or RAM 803 and / or one or more memories other than ROM 802 and RAM 803 described above.
[0167] Embodiments of this disclosure also include a computer program product comprising a computer program containing program code for performing the methods shown in the flowchart. When the computer program product is run on a computer system, the program code is used to enable the computer system to implement the image processing methods provided in the embodiments of this disclosure.
[0168] When the computer program is executed by the processor 801, it performs the functions defined in the system / apparatus of this disclosure embodiments. According to embodiments of this disclosure, the systems, apparatuses, modules, units, etc., described above can be implemented by computer program modules.
[0169] In one embodiment, the computer program may rely on a tangible storage medium such as an optical storage device or a magnetic storage device. In another embodiment, the computer program may also be transmitted and distributed in the form of signals over a network medium, and may be downloaded and installed via the communication section 809, and / or installed from a removable medium 811. The program code contained in the computer program can be transmitted using any suitable network medium, including but not limited to: wireless, wired, etc., or any suitable combination thereof.
[0170] In such an embodiment, the computer program can be downloaded and installed from a network via communication section 809, and / or installed from removable medium 811. When the computer program is executed by processor 801, it performs the functions defined in the system of this disclosure embodiment. According to embodiments of this disclosure, the systems, devices, apparatuses, modules, units, etc., described above can be implemented by computer program modules.
[0171] According to embodiments of this disclosure, program code for executing the computer programs provided in embodiments of this disclosure can be written in any combination of one or more programming languages. Specifically, these computational programs can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages include, but are not limited to, languages such as Java, C++, Python, "C", or similar programming languages. The program code can execute entirely on the user's computing device, partially on the user's device, partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).
[0172] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.
[0173] Those skilled in the art will understand that the features described in the various embodiments and / or claims of this disclosure can be combined or combined in various ways, even if such combinations or combinations are not explicitly described in this disclosure. In particular, the features described in the various embodiments and / or claims of this disclosure can be combined or combined in various ways without departing from the spirit and teachings of this disclosure. All such combinations and / or combinations fall within the scope of this disclosure.
[0174] The embodiments of this disclosure have been described above. However, these embodiments are for illustrative purposes only and are not intended to limit the scope of this disclosure. Although various embodiments have been described above, this does not mean that the measures in the various embodiments cannot be used advantageously in combination. The scope of this disclosure is defined by the appended claims and their equivalents. Various substitutions and modifications can be made by those skilled in the art without departing from the scope of this disclosure, and all such substitutions and modifications should fall within the scope of this disclosure.
Claims
1. An image processing apparatus, comprising: A target video acquisition circuit is configured to acquire a target video, wherein the target video includes N target video frames, where N is an integer greater than or equal to 2; The head pose information determination circuit is configured to perform head detection on the nth target video frame to obtain the nth head pose information, wherein the nth head pose information represents the first probability that a preset pose exists in the nth target video frame, wherein the preset pose includes one of standing pose, sitting pose and holding pose, n is an integer greater than 1 and less than or equal to N, and m is an integer greater than or equal to 1 and less than n. An intermediate pose information determination circuit is configured to perform pose detection on the nth target video frame to obtain nth intermediate pose information, wherein the nth intermediate pose information represents a second probability that the preset pose exists in the nth target video frame; and The target pose information determination circuit is configured to determine the nth target pose information based on the nth head pose information and the nth intermediate pose information, wherein the nth target pose information represents the probability that a preset pose exists in the nth target video frame.
2. The apparatus according to claim 1, wherein, The head pose information determination circuit is further configured to perform head detection on the nth target video frame to obtain the nth head pose information through the following operations: Head detection is performed on the nth target video frame to obtain the nth head position information; based on the nth head position information and at least one mth head position information, the nth position change information is determined, and based on the nth position change information, the nth head pose information is determined.
3. The apparatus according to claim 2, wherein, The information about the change at the nth position includes the change value at the nth position; The head pose information determination circuit is further configured to determine the nth head pose information based on the nth position change information through the following operations: If the change value at the nth position is determined to be greater than or equal to the first preset threshold, the nth head pose information is determined to indicate that the preset pose exists in the nth target video frame.
4. The apparatus according to any one of claims 1 to 3, wherein, The intermediate pose information determination circuit is further configured to perform pose detection on the nth target video frame to obtain the nth intermediate pose information through the following operations: The nth target video frame is subjected to pose detection to obtain at least one nth intermediate pose detection box, and the nth intermediate pose information is determined based on the at least one nth intermediate pose detection box.
5. The apparatus according to claim 4, wherein, The target pose information determination circuit is further configured to determine the nth intermediate pose information based on the at least one nth intermediate pose detection box by the following operation: If the nth head pose information indicates that the preset pose exists in the nth target video frame and the nth intermediate pose information indicates that the preset pose exists in the nth target video frame, the nth target pose information is determined based on the nth intermediate pose detection box and the nth head detection box corresponding to the nth head position information.
6. The apparatus according to claim 5, wherein, The target pose information determination circuit is further configured to determine the nth target pose information based on the nth intermediate pose detection box and the nth head detection box corresponding to the nth head position information through the following operations: If the overlap between the nth head detection box and the nth intermediate pose detection box is greater than or equal to the second preset threshold, the nth target pose information is determined to indicate that the preset pose exists in the nth target video frame.
7. The apparatus according to claim 5, wherein, The target pose information determination circuit is further configured to determine the nth target pose information based on the nth intermediate pose detection box and the nth head detection box corresponding to the nth head position information through the following operations: If the overlap between the nth head detection box and the nth intermediate pose detection box is greater than or equal to the second preset threshold, the nth head object corresponding to the nth head detection box and the nth pose object corresponding to the nth intermediate pose detection box are matched to obtain the nth matching information. as well as If it is determined that the nth matching information represents that the nth head object and the nth pose object are matched, then it is determined that the nth target pose information represents that the preset pose exists in the nth target video frame.
8. The apparatus according to any one of claims 1 to 3, further comprising: A raw video acquisition circuit is configured to acquire raw video, wherein the raw video comprises P raw video frames, where P is an integer greater than N; and The target video frame determination circuit is configured to perform frame extraction processing on the P original video frames to obtain N target video frames.
9. The apparatus according to any one of claims 1 to 3, further comprising: The circuit for determining the area to be played is configured to determine the nth area to be played from the nth target video frame when it is determined that the nth target pose information indicates that the preset pose exists in the nth target video frame; The target playback area determination circuit is configured to determine the nth target playback area based on the nth playback area to be played and the (n-1)th target playback area; as well as The playback circuit is configured to play the nth playback area.
10. The apparatus according to claim 9, wherein, The target playback area determination circuit is further configured to determine the nth target playback area based on the nth to-be-played area and the (n-1)th target playback area through the following operations: Based on the location information of the nth region to be played and the location information of the (n-1)th target playback region, the overlap between the nth region to be played and the (n-1)th target playback region is determined; as well as If the overlap between the nth region to be played and the (n-1)th target playback region is less than or equal to a third preset threshold, the nth region to be played is determined as the nth target playback region.
11. An image processing method, comprising: Acquire the target video, wherein the target video comprises N target video frames, where N is an integer greater than or equal to 2; Head detection is performed on the nth target video frame to obtain the nth head pose information. The nth head pose information represents the first probability that a preset pose exists in the nth target video frame. The preset pose includes one of standing pose, sitting pose, and holding pose. n is an integer greater than 1 and less than or equal to N, and m is an integer greater than or equal to 1 and less than n. Pose detection is performed on the nth target video frame to obtain the nth intermediate pose information, wherein the nth intermediate pose information represents the second probability that the preset pose exists in the nth target video frame; and Based on the nth head pose information and the nth intermediate pose information, the nth target pose information is determined, wherein the nth pose information represents the probability that the preset pose exists in the nth target video frame.
12. A processor for executing a computer program to implement the method of claim 11.
13. An electronic device, comprising: One or more processors; Memory, used to store one or more programs. Wherein, when the one or more programs are executed by the one or more processors, the one or more processors perform the method according to claim 11.
14. A computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, cause the processor to perform the method according to claim 11.
15. A computer program product comprising a computer program that, when executed by a processor, implements the method according to claim 11.