Image detection method and device, electronic equipment and storage medium

Abstract

The disclosure provides an image detection method, relates to the technical field of artificial intelligence, in particular to the technical field of chips and the technical field of computer vision, and can be applied to an image processing scene or a video processing scene. A specific implementation scheme is as follows: generating an initial detection result tree according to an initial detection result of an input image, wherein the initial detection result tree comprises a plurality of first nodes; determining at least one first target node from the plurality of first nodes according to the plurality of first nodes and a plurality of second target nodes in a confidence threshold tree, wherein a first confidence value of the first target node is greater than or equal to a preset confidence threshold value of the second target node matched with the first target node; and obtaining a target detection result of the input image according to the at least one first target node. The disclosure also provides an image detection device, an electronic device and a storage medium.

Description

Technical Field

[0001] This disclosure relates to the field of artificial intelligence technology, and more particularly to the fields of chip technology and computer vision technology, and can be applied to image processing or video processing scenarios. More specifically, this disclosure provides an image detection method, apparatus, electronic device, and storage medium. Background Technology

[0002] With the development of artificial intelligence technology, object detection algorithms have been widely used in various video processing and image processing scenarios. In most scenarios, object detection tasks can include multi-level multi-object detection tasks. Multi-level multi-object detection tasks can include: first detecting a target at a certain level, and then detecting at least one target at the next lower level. Summary of the Invention

[0003] This disclosure provides an image detection method, apparatus, device, and storage medium.

[0004] According to one aspect of this disclosure, an image detection method is provided, the method comprising: generating an initial detection result tree based on an initial detection result of an input image, wherein the initial detection result tree includes a plurality of first nodes, and the relevant information of the first nodes includes first identification information and a first confidence value; determining at least one first target node from the plurality of first nodes based on the plurality of first nodes and a plurality of second target nodes in a confidence threshold tree, wherein the second target node is a second node in the plurality of second nodes of the confidence threshold tree whose second identification information matches the first identification information of a first node, and the first confidence value of the first target node is greater than or equal to a preset confidence threshold of the second target node matching the first target node; and obtaining a target detection result of the input image based on the at least one first target node.

[0005] According to another aspect of this disclosure, an image detection apparatus is provided, the apparatus comprising: a storage unit configured to store an input image; and a processor configured to: read the input image from the storage unit; generate an initial detection result tree based on an initial detection result of the input image, wherein the initial detection result tree includes a plurality of first nodes, and the relevant information of the first nodes includes first identification information and a first confidence value; determine at least one first target node from the plurality of first nodes based on the plurality of first nodes and a plurality of second target nodes in a confidence threshold tree, wherein the second target node is a second node in the plurality of second nodes of the confidence threshold tree whose second identification information matches the first identification information of a first node, and the first confidence value of the first target node is greater than or equal to a preset confidence threshold of the second target node matching the first target node; obtain a target detection result of the input image based on the at least one first target node; and write the target detection result into the storage unit.

[0006] According to another aspect of this disclosure, an electronic device is provided, including the image detection apparatus provided in this disclosure.

[0007] According to another aspect of this disclosure, an electronic device is provided, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform a method provided according to this disclosure.

[0008] According to another aspect of this disclosure, a non-transitory computer-readable storage medium is provided storing computer instructions for causing a computer to perform the methods provided according to this disclosure.

[0009] According to another aspect of this disclosure, a computer program product is provided, including a computer program that, when executed by a processor, implements the method provided according to this disclosure.

[0010] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0011] The accompanying drawings are provided to better understand this solution and do not constitute a limitation of this disclosure. Wherein:

[0012] Figure 1 This is an exemplary system architecture diagram of an image detection method and apparatus applicable according to an embodiment of the present disclosure;

[0013] Figure 2 This is a flowchart of an image detection method according to an embodiment of the present disclosure;

[0014] Figure 3A This is a schematic diagram of an initial detection result tree according to an embodiment of the present disclosure;

[0015] Figure 3B This is a schematic diagram of a confidence threshold tree according to an embodiment of the present disclosure;

[0016] Figure 3C This is a schematic diagram of the output detection result tree according to an embodiment of the present disclosure;

[0017] Figure 4 This is a schematic block diagram of an image detection apparatus according to an embodiment of the present disclosure;

[0018] Figure 5 This is a schematic diagram of an electronic device according to an embodiment of the present disclosure; and

[0019] Figure 6 This is a block diagram of an electronic device to which an image detection method can be applied, according to an embodiment of the present disclosure. Detailed Implementation

[0020] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding, and should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0021] In some embodiments, a multi-level, multi-target detection task may include: first detecting whether there are pedestrians on the road; and if pedestrians are detected, then detecting whether the pedestrian is holding a cigarette to determine whether the pedestrian is smoking.

[0022] To improve the accuracy of detection results, results can be filtered based on their confidence level. In some embodiments, detection results with confidence levels below a certain threshold can be deleted, or all next-level detection results associated with that result can be deleted. For example, if different detection results correspond to different confidence thresholds, the confidence level of each detection result can be compared with the corresponding confidence threshold layer by layer and target by target based on code related to conditional judgment. For example, if a detection result indicates that the object in the image is a pedestrian, the corresponding first confidence threshold is 0.5. If the confidence value of the detection result is less than the first confidence threshold, the detection result can be deleted. If the confidence value of the detection result is greater than the first confidence threshold and the label of the next-level detection result associated with that result is a cigarette, it can be determined whether the confidence value of the next-level detection result is greater than or equal to a second confidence threshold corresponding to cigarettes.

[0023] However, different detection results at different levels correspond to different confidence thresholds. If conditional judgment code is used to make judgments on a target-by-target basis, different code needs to be written for each scenario. This causes the amount of code to increase continuously with the number of detection result layers and the number of detection result categories, which in turn leads to a significant increase in the development cost of the algorithm and also leads to a continuous increase in the hardware resources required for the algorithm to run.

[0024] Figure 1 This is a schematic diagram of an exemplary system architecture for applying an image detection method and apparatus according to an embodiment of this disclosure. It should be noted that... Figure 1 The examples shown are merely examples of system architectures that can be applied to the embodiments of this disclosure, in order to help those skilled in the art understand the technical content of this disclosure, but do not mean that the embodiments of this disclosure cannot be used in other devices, systems, environments or scenarios.

[0025] like Figure 1 As shown, the system architecture 100 according to this embodiment may include terminal devices 101, 102, and 103, a network 104, and a server 105. The network 104 serves as a medium for providing a communication link between the terminal devices 101, 102, and 103 and the server 105. The network 104 may include various connection types, such as wired and / or wireless communication links, etc.

[0026] Users can use terminal devices 101, 102, and 103 to interact with server 105 via network 104 to receive or send messages, etc. Terminal devices 101, 102, and 103 can be various electronic devices with displays and web browsing capabilities, including but not limited to smartphones, tablets, laptops, and desktop computers.

[0027] Server 105 can be a server that provides various services, such as a backend management server that supports websites browsed by users using terminal devices 101, 102, and 103 (for example only). The backend management server can analyze and process data such as received user requests, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal devices.

[0028] It should be noted that the image detection method provided in this embodiment can generally be executed by server 105. Correspondingly, the image detection device provided in this embodiment can generally be located in server 105. The image detection method provided in this embodiment can also be executed by a server or server cluster that is different from server 105 and capable of communicating with terminal devices 101, 102, 103 and / or server 105. Correspondingly, the image detection device provided in this embodiment can also be located in a server or server cluster that is different from server 105 and capable of communicating with terminal devices 101, 102, 103 and / or server 105.

[0029] Figure 2 This is a flowchart of an image detection method according to an embodiment of the present disclosure.

[0030] like Figure 2 As shown, the method 200 may include operations S210 to S230.

[0031] In operation S210, an initial detection result tree is generated based on the initial detection results of the input image.

[0032] In this embodiment of the disclosure, the initial detection result tree includes multiple first nodes, and the relevant information of each first node includes first identification information and a first confidence value. Multiple first nodes can correspond to multiple detection results. The first identification information can correspond to the category of the detection result. For example, the first identification information of a first node can indicate that the category of the corresponding detection result is pedestrian. The first confidence value of this first node can be 0.9.

[0033] In operation S220, at least one first target node is determined from the multiple first nodes based on the multiple first nodes and the multiple second target nodes in the confidence threshold tree.

[0034] In this embodiment of the disclosure, the second target node is a second node in a confidence threshold tree whose second identification information matches the first identification information of a first node. The confidence threshold tree can be configured for the corresponding scenario. The confidence threshold tree can include multiple second nodes. These multiple second nodes can correspond to multiple preset categories. The relevant information of the second node can include second identification information and a preset confidence threshold. The second identification information can correspond to a preset category. For example, the second identification information of a second node can indicate that the preset category is pedestrian. The preset confidence threshold for this second node can be 0.8. Since the second identification information of this second node is consistent with the first identification information of the aforementioned first node, it can serve as a second target node.

[0035] In this embodiment of the disclosure, the first confidence value of the first target node is greater than or equal to a preset confidence threshold of the second target node that matches the first target node. For example, if the first confidence value (0.9) of the first node is greater than the preset confidence threshold (0.8) of the second target node, it can be used as a first target node.

[0036] In operation S230, the target detection result of the input image is obtained based on at least one first target node.

[0037] In this embodiment of the disclosure, at least one first target node can be used to adjust the initial detection result tree to obtain an output detection result tree, which serves as the target detection result.

[0038] Through the embodiments of this disclosure, based on the matching between tree structures, the initial detection results can be quickly adjusted, improving the efficiency of result screening. The confidence level of each node in the confidence threshold tree, the second identifier information, or the correlation between nodes can be adjusted to adapt to different scenarios. This can save a significant amount of work in writing conditional judgment code, improving the development efficiency of object detection algorithms.

[0039] As can be understood, the image detection algorithm of this disclosure has been described above, and the initial detection result tree and confidence threshold tree will be further explained below.

[0040] In some embodiments, various object detection algorithms can be used to detect objects in the input image to obtain initial detection results. These initial detection results can be converted into a tree-structured initial detection result tree. The initial detection result tree may include multiple first nodes. The following will combine... Figure 3A Further explanation is needed.

[0041] Figure 3A This is a schematic diagram of an initial detection result tree according to an embodiment of the present disclosure.

[0042] In this embodiment of the disclosure, the multiple first nodes may include multi-level first nodes. There can be at least one first node at the same level. Through this embodiment, the initial detection result tree including multi-level first nodes is more suitable for multi-level multi-target detection tasks. Based on multi-level first nodes, the first target node can be accurately selected, resulting in accurate target detection results.

[0043] For example, the initial detection result tree can also include a root node. The first node in a multi-level tree can be level I. For example... Figure 3A As shown, the initial detection result tree can include a root node R0 and three level-one nodes. There are three level-one nodes: R1, R2, and R3. There are three level-two nodes: R11, R21, and R22. There are two level-three nodes: R211 and R212.

[0044] In this embodiment of the disclosure, the root node of the initial detection result tree can be associated with a first-level node in the first-level first node group. For example, the relevant information of the root node R0 of the initial detection result tree includes association information. This association information can indicate at least one first-level node associated with the root node R0. Figure 3A As shown, the association information child of the root node R0: [R1, R2, R3] indicates that the root node R0 is associated with the first-level first node R1, the first-level first node R2, and the first-level first node R3, respectively.

[0045] In this embodiment of the disclosure, the relevant information of the first node may include first association information. The first association information may indicate the next-level first node associated with the first node, or it may indicate that the first node is not associated with a next-level first node. Among the first nodes of level I, the i-th level first node may be associated with at least one (i+1)-th level first node. I can be an integer greater than 1, and i can be an integer greater than or equal to 1 and less than I. For example... Figure 3AAs shown, the first association information child of the first node R1 at level 1: [R11] indicates that the first node R1 at level 1 is associated with the first node R11 at level 2. The first association information child of the first node R2 at level 1: [R21, R22] indicates that the first node R1 at level 1 is associated with the first nodes R21 and R22 at level 2 respectively. The first association information child of the first node R3 at level 1: [] indicates that the first node R3 at level 1 is not associated with the first node at the next level. The first association information child of the first node R11 at level 2: [] indicates that the first node R11 at level 2 is not associated with the first node at the next level. The first association information child of the first node R21 at level 2: [R211, R212] indicates that the first node R21 at level 2 is associated with the first nodes R211 and R212 at level 3 respectively. The first association information child of the first node R22 at level 2: [] indicates that the first node R22 at level 2 is not associated with the first node at the next level. The first association information child of the first node R211 at level 3: [] indicates that the first node R211 at level 3 is not associated with the first node at the next level. The first association information child of the first node R212 at level 3: [] indicates that the first node R212 at level 3 is not associated with the first node at the next level.

[0046] In this embodiment of the disclosure, the relevant information of the first node may further include first identification information and a first confidence value. The first identification information may indicate the category of an object in the input image. The first confidence value corresponding to the first identification information may be: the confidence that the object belongs to that category.

[0047] like Figure 3A As shown, the first identifier of the first node R1 at level 1, named "container," indicates that an object in the input image belongs to the category of a container. The first confidence value of the first node R1 at level 1 can be 0.5. The first node R11 at level 2 is associated with the first node R1 at level 1. The first identifier of the first node R11 at level 2, named "smoke," indicates that the container in the input image is emitting smoke. The first confidence value of the first node R11 at level 2 can be 0.8.

[0048] like Figure 3AAs shown, the first identifier of the first node R2 at level 1, named "human", indicates that an object in the input image belongs to the category of a pedestrian. The first confidence value of the first node R2 at level 1 can be 0.9. The first nodes R21 and R22 at level 2 are associated with the first node R2 at level 1. The first identifier of the first node R21 at level 2, named "overcoat", indicates that the pedestrian in the input image is wearing an overcoat. The first confidence value of the first node R21 at level 2 can be 0.4. The first identifier of the first node R22 at level 2, named "smoke", indicates that the pedestrian in the input image is smoking. The first confidence value of the first node R22 at level 2 can be 0.7. The first nodes R211 and R212 at level 3 are associated with the first node R21 at level 2. The first identifier of the first node R211 at level 3, named "black", indicates that the pedestrian in the input image is wearing a black overcoat. The first confidence value of the first node R211 at level 3 can be 0.1. The first identifier information of the first node R212 at level 3, name: "red-vest", can indicate that the pedestrian in the input image is wearing a red vest. The first confidence value of the first node R212 at level 3 can be 0.8.

[0049] like Figure 3A As shown, the first identifier information of the first node R3 at level 1, name: "obstacle", can indicate that an object in the input image belongs to the category of an obstacle. The first confidence value of the first node R3 at level 1 can be 0.4.

[0050] In this embodiment of the disclosure, the first node may correspond to a detection box. The relevant information of the first node may further include the detection box position information. For example... Figure 3AAs shown, the location information of the first detection box of the first node R1 in level 1, location: [a, b, c, d], indicates that the coordinates of the top left vertex of the detection box are (a, b), the height of the detection box is c, and the width of the detection box is d. The location information of the first detection box of the first node R2 in level 1, location: [e, f, g, h], indicates that the coordinates of the top left vertex of the detection box are (e, f), the height of the detection box is g, and the width of the detection box is h. The location information of the first detection box of the first node R3 in level 1, location: [k, l, m, n], indicates that the coordinates of the top left vertex of the detection box are (k, 1), the height of the detection box is m, and the width of the detection box is n. The location information of the first detection box of the first node R11 in level 2, location: [o, p, q, r], indicates that the coordinates of the top left vertex of the detection box are (o, p), the height of the detection box is q, and the width of the detection box is r. The location information of the first detection box of the first node R21 in level 2, [s,t,u,v], indicates that the coordinates of the top-left vertex of the detection box are (s,t), the height of the detection box is u, and the width of the detection box is v. The location information of the first detection box of the first node R22 in level 2, [w,x,y,z], indicates that the coordinates of the top-left vertex of the detection box are (w,x), the height of the detection box is y, and the width of the detection box is z.

[0051] Understandable, such as Figure 3A As shown, I = 3. This can also be understood as... Figure 3A The initial detection result tree shown is just an example; I can be any value greater than 1.

[0052] As can be understood, the initial detection result tree of this disclosure has been explained above, and the confidence threshold tree of this disclosure will be further explained below.

[0053] Figure 3B This is a schematic diagram of a confidence threshold tree according to an embodiment of the present disclosure.

[0054] In this embodiment of the disclosure, the confidence threshold tree may include multiple second nodes. These multiple second nodes may include multi-level second nodes. There may be at least one second node at the same level.

[0055] In this embodiment of the disclosure, the confidence threshold tree may further include a root node. The multi-level second nodes may be of level J. For example... Figure 3BAs shown, the confidence threshold tree can include a root node T0 and three levels of second nodes. There are three level 1 second nodes: T1, T2, and T3. There are five level 2 second nodes: T11, T12, T21, T22, and T23. There are three level 3 second nodes: T211, T212, and T213.

[0056] In this embodiment of the disclosure, the root node of the confidence threshold tree can be associated with a first-level second node in the J-level second nodes. The relevant information of the root node T0 of the initial detection result tree can include association information. This association information can indicate at least one first-level second node associated with the root node T0. For example... Figure 3B As shown, the association information child of the root node T0: [T1, T2, T3] indicates that the root node T0 is associated with the first-level second node T1, the first-level second node T2, and the first-level second node T3, respectively.

[0057] In this embodiment of the disclosure, a second node at level j can be associated with at least one second node at level j+1. J can be an integer greater than 1, and j can be an integer greater than or equal to 1 and less than J. The relevant information of the second node may include second association information. The second association information can indicate the next-level second node associated with the second node, or it can indicate that the second node is not associated with a next-level second node. For example... Figure 3BAs shown, the second association information child of level 1 second node T1: [T11, T12] indicates that level 1 second node T1 is associated with level 2 second nodes T11 and T12. The second association information child of level 1 second node T2: [T21, T22, T23] indicates that level 1 second node T2 is associated with level 2 second nodes T21, T22, and T23 respectively. The second association information child of level 1 second node T3: [] indicates that level 1 second node T3 is not associated with a next-level second node. The second association information child of level 2 second node T11: [] indicates that level 2 second node T11 is not associated with a next-level second node. The second association information child of level 2 second node T12: [] indicates that level 2 second node T12 is not associated with a next-level second node. The second association information child of level 2 second node T21: [T211, T212, T213] indicates that level 2 second node T21 is associated with level 3 second nodes T211, T212, and T213 respectively. The second association information child of level 2 second node T22: [] indicates that level 2 second node T22 is not associated with a next-level second node. The second association information child of level 2 second node T23: [] indicates that level 2 second node T23 is not associated with a next-level second node. The second association information child of level 3 second node T211: [] indicates that level 3 second node T211 is not associated with a next-level second node. The second association information child of level 3 second node T212: [] indicates that level 3 second node T212 is not associated with a next-level second node. The second association information child of the third-level second node T213: [] can indicate that the third-level second node T213 is not associated with the next level second node.

[0058] In this embodiment of the disclosure, the relevant information of the second node may further include second identification information and a preset confidence threshold. The second identification information may indicate a preset category of an object in the image related to the current scene. If the first confidence value of the first node matching the second identification information is greater than or equal to the corresponding preset confidence threshold, it can be determined that an object of the preset category exists in the image.

[0059] like Figure 3BAs shown, the second identifier information "name": "container" of the first-level second node T1 indicates that the preset category of an object in the image is a container. The preset confidence threshold of the first-level second node T1 can be 0.6. The second-level second node T11 is associated with the first-level second node T1. The second identifier information "name": "fire" of the second-level second node T11 indicates that there is a flame on the container in the image. The preset confidence threshold of the second-level second node T11 can be 0.2. The second-level second node T12 is associated with the first-level second node T1. The second identifier information "name": "smoke" of the second-level second node T12 indicates that the container in the image is emitting smoke. The preset confidence threshold of the second-level second node T12 can be 0.4.

[0060] like Figure 3B As shown, the second identifier information "name": "human" of the first-level second node T2 indicates that the preset category of an object in the image is a pedestrian. The preset reliability threshold of the first-level second node T2 can be 0.8. The second-level second nodes T21, T22, and T23 are associated with the first-level second node T2. The second identifier information "name": "overcoat" of the second-level second node T21 indicates that the pedestrian in the input image is wearing a coat. The preset reliability threshold of the second-level second node T21 can be 0.4. The second identifier information "name": "smoke" of the second-level second node T22 indicates that the pedestrian in the image is smoking. The preset reliability threshold of the second-level second node T22 can be 0.3. The second identifier information "name": "helinet" of the second-level second node T23 indicates that the pedestrian in the image is wearing a helmet. The preset reliability threshold of the second-level second node T23 can be 0.5. Level 3 second nodes T211, T212, and T213 are associated with Level 2 second node T21, respectively. The second identifier of Level 3 second node T211, named "black," indicates that the pedestrian in the image is wearing a black coat. The preset reliability threshold for Level 3 second node T211 can be 0.5. The second identifier of Level 3 second node T212, named "red-vest," indicates that the pedestrian in the image is wearing a red vest. The preset reliability threshold for Level 3 second node T212 can be 0.7. The second identifier of Level 3 second node T213, named "blue," indicates that the pedestrian in the image is wearing a blue coat. The preset reliability threshold for Level 3 second node T213 can be 0.6.

[0061] like Figure 3BAs shown, the second identifier information name: "obstacle" of the first-level second node T3 can indicate that the preset category of an object in the image is an obstacle. The preset confidence threshold of the first-level second node T3 can be 0.3.

[0062] It's understandable that J can be 3. It's also understandable that... Figure 3B The confidence threshold tree shown is just an example; J can be any value greater than 1.

[0063] As can be understood, the confidence threshold tree of this disclosure has been described above, and the image detection method of this disclosure will be further described below.

[0064] In embodiments of this disclosure, determining at least one target node from a plurality of first nodes may include: when there are multiple first nodes at the current level, sequentially determining whether each of the multiple first nodes at the current level is a first target node. For example, when there are multiple first nodes at the i-th level, it is possible to sequentially determine whether each of the multiple i-th level first nodes is a first target node. Figure 3A As shown, there are 3 first-level nodes. We can sequentially determine whether the first-level nodes R1, R2, and R3 are the first target nodes.

[0065] In this embodiment of the disclosure, determining whether a plurality of first nodes at the current level are to be used as first target nodes may include: for a first current node among a plurality of first nodes at the current level, identifying a second node whose second identification information matches the first identification information of the first current node as a second target node matching the first current node. In response to determining that a first confidence value of the first current node is less than a preset confidence threshold for the second target node matching the first current node, deleting the first current node and any subsequent first nodes associated with the first current node.

[0066] For example, for a current first node of level i among multiple first nodes of level i, the second node of level j whose second identification information matches the first identification information of the current first node of level i is taken as the second target node that matches the current first node of level i. In response to determining that the first confidence value of the current first node of level i is less than the preset confidence threshold of the second target node that matches the current first node of level i, the current first node of level i and the subsequent first nodes associated with the current first node of level i are deleted.

[0067] like Figure 3A and Figure 3BAs shown, the first node R1 of level 1 can be used as the first current node of the current level. The first identifier information name: "container" of the first node R1 of level 1 matches the second identifier information name: "container" of the second node T1 of level 1. The second node T1 of level 1 can be used as the second target node that matches the first node R1 of level 1. The first confidence value of the first node R1 of level 1 is 0.5, which is less than the preset confidence threshold (0.6) of the second node T1 of level 1. Therefore, the first node R1 of level 1 and the first node R11 of level 2 can be deleted. Next, it can be determined whether other first nodes of the same level as the first node R1 of level 1 can be used as the first target node. Through the embodiments of this disclosure, first nodes with a first confidence value less than the preset confidence threshold can be quickly deleted, which can save storage resources and also help to obtain accurate target detection results.

[0068] In this embodiment, determining whether a plurality of first nodes at the current level are designated as first target nodes may include: for a first current node among the plurality of first nodes at the current level, identifying a second node whose second identification information matches the first identification information of the first current node as a second target node matching the first current node. In response to determining that a first confidence value of the first current node is greater than or equal to a preset confidence threshold for the second target node matching the first current node, the first current node is designated as a first target node. Through this embodiment, first nodes with first confidence values ​​greater than or equal to a preset confidence threshold can be quickly determined, which helps to obtain more accurate target detection results.

[0069] For example, in response to determining that the first confidence value of the first current node of level i is greater than or equal to the preset confidence threshold of the second target node that matches the first current node of level i, the first current node of level i is taken as a first target node.

[0070] like Figure 3A and Figure 3B As shown, after the first node R1 of level 1, the first node R2 of level 1 can be taken as the first current node of the current level. The first identifier information name: "human" of the first node R2 of level 1 matches the second identifier information name: "human" of the second node T2 of level 1. The second node T2 of level 1 can be taken as the second target node that matches the first node R2 of level 1. The first confidence value of the first node R2 of level 1 is 0.9, which is greater than the preset confidence threshold (0.8) of the second node T2 of level 1. Therefore, the first node R2 of level 1 can be taken as a first target node.

[0071] In this embodiment of the disclosure, determining at least one first target node from a plurality of first nodes may include: in response to determining that the first current node of the current level is a first target node, determining whether any subsequent first node associated with the first current node of the current level is a first target node. Any subsequent first node may include: at least one next-level first node associated with the first current node. At least one first-level first node associated with the first current node via the next-level first node. Through this embodiment of the disclosure, after determining that a node is a first target node, determining whether the first node associated with it is also a first target node, combined with the detection results of the tree structure, can reduce the configuration difficulty of the identification information, ensure filtering accuracy, and further improve filtering efficiency. For example, the first identification information of the second-level first node R11 and the second-level first node R22 can both be name: "smoke", but they can indicate different information without reducing filtering accuracy.

[0072] like Figure 3A and Figure 3B As shown, if the first-level node R2 is determined to be a first target node, it can be determined whether the second-level nodes R21 and R22, which are associated with the first-level node R2, can also be first target nodes. Furthermore, via the second-level node, the third-level nodes R211 and R212 are associated with the first-level node R2. Therefore, it can also be determined whether the third-level nodes R211 and R212 can be first target nodes.

[0073] In this embodiment of the disclosure, determining whether any first node in a subsequent level associated with the first current node of the current level is a first target node includes: in response to determining that the first current node of the i-th level among a plurality of first nodes of the i-th level is a first target node, determining the (i+1)-th level first current node from at least one first node of the i-th level associated with the first current node of the i-th level. The first identification information of the first current node of the i-th level matches a second node of the j-th level. For example... Figure 3A and Figure 3BAs shown, when the first node R2 of the first level is a first target node, the first current node of the second level can be determined from the first nodes R21 and R22 of the second level. For example, the first node R21 of the second level can be used as the first current node of the second level. Furthermore, as mentioned above, the first node R2 of the first level is matched with the second node T2 of the first level. The second node T2 of the first level is associated with the second nodes T21, T22, and T23 of the second level. The second target node matching the first node R21 of the second level can be determined from the second nodes T21 to T23 of the second level. Through the embodiments of this disclosure, for the second node associated with the first target node, the second target node of the next level can be determined from the second nodes associated with the matched second target node of the next level, which can further improve the screening efficiency and accuracy, and help improve the detection accuracy.

[0074] In this embodiment of the disclosure, determining whether any first node in a subsequent level associated with the first current node of the current level is a first target node may include: identifying the (j+1)th level second node whose second identification information matches the first identification information of the (i+1)th level first current node as the second target node matching the (i+1)th level first current node. The (j+1)th level second node is associated with the (j)th level second node. Figure 3A and Figure 3B As shown, the second identifier information name: "overcoat" of the second node T21 of the second level matches the first identifier information name: "overcoat" of the first node R21 of the second level, and can be used as the second target node that matches the first node R21 of the second level.

[0075] In this embodiment of the disclosure, determining whether any first node in a subsequent level associated with the first current node of the current level is a first target node may include: in response to determining that a first confidence value of the first current node of the (i+1)th level is greater than or equal to a preset confidence threshold of a second target node that matches the first current node of the (i+1)th level, the first current node of the (i+1)th level is designated as a first target node. Figure 3A and Figure 3B As shown, the first confidence value of the first node R21 in the second level is 0.4, which is equal to the preset confidence threshold (0.4) of the second node T21 in the second level. Therefore, the first node R21 in the second level can be used as a first target node.

[0076] Having determined that the first node of level 2, R21, can be used as the first target node, it can be determined whether the first nodes of level 3, R211 and R212 associated with R21, can also be used as first target nodes. For example, R211 can be used as the first current node of level 3. Furthermore, as mentioned above, the second node of level 2, T21, is associated with the second nodes of level 3, T211, T212, and T213. A second target node matching the first node of level 3, R211, can be determined from the second nodes of level 3, T211 to T213.

[0077] like Figure 3A and Figure 3B As shown, the second identifier information "name: "black" of the second node T211 at level 3 matches the first identifier information "name: "black" of the first node R211 at level 3. Therefore, the second node T211 at level 3 can be used as the second target node that matches the first node R211 at level 3. The first confidence value of the first node R211 at level 3 is 0.1, which is less than the preset confidence threshold (0.5) for the second node T211 at level 3. Therefore, the first node R211 at level 3 can be deleted. Next, the first node R212 at level 3 can be used as the first current node at level 3.

[0078] like Figure 3A and Figure 3B As shown, the second identifier information "name: "red-vest" of the second node T212 at level 3 matches the first identifier information "name: "red-vest" of the first node R212 at level 3. Therefore, the second node T212 at level 3 can be considered as the second target node that matches the first node R212 at level 3. The first confidence value of the first node R212 at level 3 is 0.8, which is greater than the preset confidence threshold (0.7) for the second node T212 at level 3. Therefore, the first node R212 at level 3 can be considered as a first target node.

[0079] In this embodiment of the disclosure, determining at least one first target node from a plurality of first nodes further includes: in response to whether all first nodes associated with the current-level first current node have been determined to be first target nodes, determining a first subsequent node from a plurality of first nodes at the same level as the first current node. For example, in response to whether the first node associated with the i-th level first current node has been determined to be a first target node, determining the i-th level first subsequent node of the i-th level first current node from a plurality of i-th level first nodes. Figure 3AAs shown, the first node of level 3 is not associated with any subsequent first node. Therefore, to determine whether any of the first nodes associated with the first node R21 of level 2 can be considered the first target node, the first subsequent node can be determined from the first nodes at the same level as R21. For example, the first node R22 of level 2 can be considered the first subsequent node of level 2. It can then be determined whether the first node R22 of level 2 is the first target node. Furthermore, a second target node matching the first node R22 of level 2 can be determined from the second nodes T21 to T23 of level 2.

[0080] The second identifier information "name: "smoke" of the second node T22 at level 2 matches the first identifier information "name: "smoke" of the first node R22 at level 2, and can be used as the second target node matching the first node R22 at level 2. The first confidence value of the first node R22 at level 2 is 0.7, which is greater than the preset confidence threshold (0.3) for the second node T22 at level 2. Therefore, the first node R22 at level 2 can be used as a first target node.

[0081] like Figure 3A As shown, the first node R22 of level 2 is not associated with any subsequent first node. Therefore, having determined whether all first nodes associated with the first node R2 of level 1 are the first target node, the first subsequent node can be determined from the first nodes at the same level as the first node R2 of level 1. For example, the first node R3 of level 1 can be considered the first subsequent node of level 1. It can then be determined whether the first node R3 of level 1 is the first target node. Furthermore, a second target node matching the first node R3 of level 1 can be determined from the second nodes T1 to T3 of level 1.

[0082] The second identifier information "name: "obstacle" of the second node T3 in level 1 matches the first identifier information "name: "obstacle" of the first node R3 in level 1, and can be used as the second target node matching the first node R3 in level 1. The first confidence value of the first node R3 in level 1 is 0.4, which is greater than the preset confidence threshold (0.3) for the second node T3 in level 1. Therefore, the first node R3 in level 1 can be used as a first target node.

[0083] Therefore, it has been determined whether all first nodes of the initial detection result tree can be used as first target nodes. Next, the target detection result of the input image can be determined based on at least one first target node.

[0084] Figure 3C This is a schematic diagram of the output detection result tree according to an embodiment of the present disclosure.

[0085] As mentioned above, the first nodes of level 1 (R2 and R3), level 2 (R21 and R22), and level 3 (R211) can be used as the first target nodes. Therefore, it is possible to... Figure 3A The initial detection result tree shown is pruned to remove all nodes except the first target node, resulting in the following: Figure 3C The output detection result tree shown is used as the target detection result.

[0086] It is understood that the image detection method of this disclosure has been described above, and the image detection device of this disclosure will be described below.

[0087] Figure 4 This is a schematic block diagram of an image detection apparatus according to an embodiment of the present disclosure.

[0088] like Figure 4 As shown, the image detection device 400 may include a storage unit 410 and a processor 420.

[0089] Storage unit 410 is configured to store input images.

[0090] Processor 420 is configured to: read an input image from a storage unit; generate an initial detection result tree based on the initial detection results of the input image; for example, the initial detection result tree includes multiple first nodes, and the relevant information of the first nodes includes first identification information and a first confidence value; determine at least one first target node from the multiple first nodes based on the multiple first nodes and multiple second target nodes in a confidence threshold tree; for example, a second target node is a second node in the confidence threshold tree whose second identification information matches the first identification information of a first node, and the first confidence value of the first target node is greater than or equal to a preset confidence threshold of the second target node that matches the first target node; obtain the target detection result of the input image based on at least one first target node; and write the target detection result to the storage unit.

[0091] For example, the processor can execute method 200 described above.

[0092] In some embodiments, the plurality of first nodes includes multi-level first nodes, and there is at least one first node at the same level.

[0093] In some embodiments, the processor is further configured to: when there are multiple first nodes in the current level, sequentially determine whether the multiple first nodes in the current level are first target nodes.

[0094] In some embodiments, the processor is further configured to: for a first current node among a plurality of first nodes in the current level, identify a second node whose second identification information matches the first identification information of the first current node as a second target node matching the first current node. In response to determining that a first confidence value of the first current node is greater than or equal to a preset confidence threshold for the second target node matching the first current node, the first current node is identified as a first target node.

[0095] In some embodiments, the processor is further configured to: in response to determining that a first current node of the current level is a first target node, determine whether any first node in a subsequent level associated with the first current node of the current level is a first target node.

[0096] In some embodiments, any subsequent first node includes: at least one next-level first node associated with the first current node; at least one first-level first node associated with the first current node via a next-level first node.

[0097] In some embodiments, the processor is further configured to: determine a first subsequent node from a plurality of first nodes at the same level as the first current node in response to whether all first nodes associated with the first current node at the current level are first target nodes.

[0098] In some embodiments, the initial detection result tree includes a root node and a first node of level I. The root node of the initial detection result tree is associated with the first node of level I in the first node of level I. The first node of level I in the first node of level I is associated with at least one first node of level i+1. I is an integer greater than 1, and i is an integer greater than or equal to 1 and less than I.

[0099] In some embodiments, the confidence threshold tree includes a root node and a J-level second node. The root node of the confidence threshold tree is associated with the first-level second node in the J-level second node, and the j-th level second node in the J-level second node is associated with at least one (j+1)-th level second node. J is an integer greater than 1, and j is an integer greater than or equal to 1 and less than J.

[0100] In some embodiments, the processor is further configured to: in response to determining that a first current node of level i among a plurality of first nodes of level i is a first target node, determine a first current node of level i+1 from at least one first node of level i+1 associated with the first current node of level i. For example, a first identification information of the first current node of level i matches a second node of level j. The second node of level j+1 whose second identification information matches the first identification information of the first current node of level i+1 is designated as a second target node matching the first current node of level i+1. For example, the second node of level j+1 is associated with another second node of level j. In response to determining that a first confidence value of the first current node of level i+1 is greater than or equal to a preset confidence threshold of the second target node matching the first current node of level i+1, designate the first current node of level i+1 as a first target node.

[0101] In some embodiments, there are multiple first nodes of level i, and the processor is further configured to: in response to determining whether a first node associated with the first current node of level i is a first target node, determine the first subsequent node of level i of the first current node of level i from the multiple first nodes of level i.

[0102] It is understood that the apparatus of this disclosure has been described above, and an electronic device of this disclosure will be described below.

[0103] Figure 5 This is a schematic diagram of an electronic device according to an embodiment of the present disclosure.

[0104] like Figure 5 As shown, device 50 may include image detection device 500. Image detection device 500 may be, for example, the image detection device 400 described above.

[0105] It is understood that the above description describes one type of electronic device of this disclosure, and the following description will describe another electronic device, storage medium, and program product of this disclosure.

[0106] The collection, storage, use, processing, transmission, provision, and disclosure of user personal information involved in the technical solution disclosed herein comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0107] According to embodiments of this disclosure, this disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0108] Figure 6A schematic block diagram of an example electronic device 600 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0109] like Figure 6 As shown, device 600 includes a computing unit 601, which can perform various appropriate actions and processes based on a computer program stored in read-only memory (ROM) 602 or a computer program loaded from storage unit 608 into random access memory (RAM) 603. RAM 603 may also store various programs and data required for the operation of device 600. The computing unit 601, ROM 602, and RAM 603 are interconnected via bus 604. Input / output (I / O) interface 605 is also connected to bus 604.

[0110] Multiple components in device 600 are connected to I / O interface 605, including: input unit 606, such as keyboard, mouse, etc.; output unit 607, such as various types of monitors, speakers, etc.; storage unit 608, such as disk, optical disk, etc.; and communication unit 609, such as network card, modem, wireless transceiver, etc. Communication unit 609 allows device 600 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0111] The computing unit 601 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 601 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 601 performs the various methods and processes described above, such as image detection methods. For example, in some embodiments, the image detection method may be implemented as a computer software program tangibly contained in a machine-readable medium, such as storage unit 608. In some embodiments, part or all of the computer program may be loaded and / or installed on device 600 via ROM 602 and / or communication unit 609. When the computer program is loaded into RAM 603 and executed by the computing unit 601, one or more steps of the image detection method described above may be performed. Alternatively, in other embodiments, the computing unit 601 may be configured to perform the image detection method by any other suitable means (e.g., by means of firmware).

[0112] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0113] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0114] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0115] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) monitor or an LCD (liquid crystal display)) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0116] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with embodiments of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0117] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other.

[0118] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this disclosure can be achieved, and this is not limited herein.

[0119] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. An image detection method, comprising: Based on the initial detection results of the input image, an initial detection result tree is generated, wherein the initial detection result tree includes multiple first nodes, and the relevant information of the first node includes first identification information and a first confidence value; Based on a plurality of first nodes and a plurality of second target nodes in a confidence threshold tree, at least one first target node is determined from the plurality of first nodes, wherein the second target node is a second node in the plurality of second nodes of the confidence threshold tree whose second identification information matches the first identification information of a first node, and the first confidence value of the first target node is greater than or equal to a preset confidence threshold of the second target node that matches the first target node; and The target detection result of the input image is obtained based on at least one of the first target nodes.

2. The method according to claim 1, wherein, The plurality of first nodes includes multi-level first nodes, and there is at least one first node at the same level.

3. The method according to claim 2, wherein, Determining at least one first target node from a plurality of first nodes includes: If there are multiple first nodes in the current level, determine in turn whether each of the multiple first nodes in the current level is the first target node.

4. The method according to claim 3, wherein, The step of sequentially determining whether multiple first nodes at the current level are the first target node includes: For the first current node among multiple first nodes at the current level, the second node whose second identification information matches the first identification information of the first current node is taken as the second target node that matches the first current node; In response to determining that the first confidence value of the first current node is greater than or equal to the preset confidence threshold of the second target node that matches the first current node, the first current node is used as the first target node.

5. The method according to claim 2, wherein, Determining at least one first target node from a plurality of first nodes includes: In response to determining that the first current node of the current level is a first target node, determine whether any first node in the subsequent level associated with the first current node of the current level is the first target node.

6. The method according to claim 5, wherein, The first node in the subsequent level includes: At least one next-level first node associated with the first current node; and At least one first node associated with the first current node via the next-level first node.

7. The method according to claim 2, wherein, The step of determining at least one first target node from a plurality of first nodes further includes: In response to whether all first nodes associated with the first current node at the current level are the first target node, the first subsequent node is determined from among multiple first nodes at the same level as the first current node.

8. The method according to claim 5, wherein, The initial detection result tree includes a root node and a first node of level I. The root node of the initial detection result tree is associated with the first node of level I. The first node of level I is associated with at least one first node of level i+1. I is an integer greater than 1, and i is an integer greater than or equal to 1 and less than I. The confidence threshold tree includes a root node and J-level second nodes. The root node of the confidence threshold tree is associated with the first-level second node in the J-level second nodes. The j-th level second node in the J-level second nodes is associated with at least one (j+1)-th level second node. J is an integer greater than 1, and j is an integer greater than or equal to 1 and less than J.

9. The method according to claim 8, wherein, Determining whether any first node in a subsequent level associated with the first current node of the current level is the first target node includes: In response to determining that the first current node of the i-th level among a plurality of the first nodes of the i-th level is a first target node, a first current node of the (i+1)-th level is determined from at least one first node of the i-th level associated with the first current node of the i-th level, wherein the first identification information of the first current node of the i-th level matches a second node of the j-th level. The second node at level j+1 that matches the second identification information with the first identification information of the first current node at level i+1 is taken as the second target node that matches the first current node at level i+1, wherein the second node at level j+1 is associated with the second node at level j. In response to determining that the first confidence value of the first current node of level i+1 is greater than or equal to the preset confidence threshold of the second target node that matches the first current node of level i+1, the first current node of level i+1 is used as a first target node.

10. The method according to claim 8, wherein, There are multiple first nodes of the i-th level. Determining whether any first node in a subsequent level associated with the first current node of the current level is the first target node includes: In response to whether the first node associated with the first current node of the i-th level is the first target node, the first subsequent node of the i-th level of the first current node of the i-th level is determined from a plurality of first nodes of the i-th level.

11. An image detection apparatus, comprising: Storage unit, configured to store input images; The processor is configured as follows: Read the input image from the storage unit; Based on the initial detection results of the input image, an initial detection result tree is generated, wherein the initial detection result tree includes multiple first nodes, and the relevant information of the first node includes first identification information and a first confidence value; Based on multiple first nodes and multiple second target nodes in a confidence threshold tree, at least one first target node is determined from multiple first nodes, wherein the second target node is a second node in the multiple second nodes of the confidence threshold tree whose second identification information matches the first identification information of a first node, and the first confidence value of the first target node is greater than or equal to a preset confidence threshold of the second target node that matches the first target node; Based on at least one of the first target nodes, the target detection result of the input image is obtained; and The target detection result is written into the storage unit.

12. The apparatus according to claim 11, wherein, The plurality of first nodes includes multi-level first nodes, and there is at least one first node at the same level.

13. The apparatus according to claim 12, wherein, The processor is also configured to: If there are multiple first nodes in the current level, determine in turn whether each of the multiple first nodes in the current level is the first target node.

14. The apparatus according to claim 13, wherein, The processor is also configured to: For the first current node among multiple first nodes at the current level, the second node whose second identification information matches the first identification information of the first current node is taken as the second target node that matches the first current node; In response to determining that the first confidence value of the first current node is greater than or equal to the preset confidence threshold of the second target node that matches the first current node, the first current node is used as the first target node.

15. The apparatus according to claim 12, wherein, The processor is also configured to: In response to determining that the first current node of the current level is a first target node, determine whether any first node in the subsequent level associated with the first current node of the current level is the first target node.

16. The apparatus according to claim 15, wherein, The first node in the subsequent level includes: At least one next-level first node associated with the first current node; and At least one first node associated with the first current node via the next-level first node.

17. The apparatus according to claim 12, wherein, The processor is also configured to: In response to whether all first nodes associated with the first current node at the current level are the first target node, the first subsequent node is determined from among multiple first nodes at the same level as the first current node.

18. The apparatus according to claim 15, wherein, The initial detection result tree includes a root node and a first node of level I. The root node of the initial detection result tree is associated with the first node of level I. The first node of level I is associated with at least one first node of level i+1. I is an integer greater than 1, and i is an integer greater than or equal to 1 and less than I. The confidence threshold tree includes a root node and J-level second nodes. The root node of the confidence threshold tree is associated with the first-level second node in the J-level second nodes. The j-th level second node in the J-level second nodes is associated with at least one (j+1)-th level second node. J is an integer greater than 1, and j is an integer greater than or equal to 1 and less than J.

19. The apparatus according to claim 18, wherein, The processor is also configured to: In response to determining that the first current node of the i-th level among a plurality of the first nodes of the i-th level is a first target node, a first current node of the (i+1)-th level is determined from at least one first node of the i-th level associated with the first current node of the i-th level, wherein the first identification information of the first current node of the i-th level matches a second node of the j-th level. The second node at level j+1 that matches the second identification information with the first identification information of the first current node at level i+1 is taken as the second target node that matches the first current node at level i+1, wherein the second node at level j+1 is associated with the second node at level j. In response to determining that the first confidence value of the first current node of level i+1 is greater than or equal to the preset confidence threshold of the second target node that matches the first current node of level i+1, the first current node of level i+1 is used as a first target node.

20. The apparatus according to claim 18, wherein, There are multiple first nodes of the i-th level. The processor is also configured to: In response to whether the first node associated with the first current node of the i-th level is the first target node, the first subsequent node of the i-th level of the first current node of the i-th level is determined from a plurality of first nodes of the i-th level.

21. An electronic device comprising the means as described in any one of claims 11 to 20.

22. An electronic device, comprising: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1 to 10.

23. A non-transitory computer-readable storage medium storing computer instructions, wherein, The computer instructions are used to cause the computer to perform the method according to any one of claims 1 to 10.

24. A computer program product comprising a computer program that, when executed by a processor, implements the method according to any one of claims 1 to 10.

Images