CN114155555A - Human behavior artificial intelligence judgment system and method - Google Patents
Human behavior artificial intelligence judgment system and method Download PDFInfo
- Publication number
- CN114155555A CN114155555A CN202111463690.0A CN202111463690A CN114155555A CN 114155555 A CN114155555 A CN 114155555A CN 202111463690 A CN202111463690 A CN 202111463690A CN 114155555 A CN114155555 A CN 114155555A
- Authority
- CN
- China
- Prior art keywords
- human body
- behavior
- equipment
- human
- binary
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to a human behavior artificial intelligence judgment system and method, and belongs to the technical field of artificial intelligence. The system comprises: the behavior judgment device is arranged in a control chamber equipped with a showroom and used for taking a preselected number of human body image blocks of the same human body object in each content scene image respectively corresponding to each acquisition time within a preset time length as a plurality of input data of a trained depth belief network model and executing the trained model to obtain the human body behavior type of the human body object within the preset time length; and the remote alarm device is used for executing an alarm action to the big data server when the received human behavior type is that equipment is stolen or damaged. By the method and the system, the artificial intelligence judgment can be performed on the current behavior type of each human body object of the equipment showroom by adopting the deep belief network, so that the intelligent level of equipment showroom management is improved, and the labor cost of equipment showroom management is effectively reduced.
Description
Technical Field
The invention relates to the field of artificial intelligence, in particular to a human behavior artificial intelligence judgment system and a human behavior artificial intelligence judgment method.
Background
The definition of artificial intelligence can be divided into two parts, namely "artificial" and "intelligence". "artificial" is considered to be what human energy and production are, or whether the intelligence of a human is high enough to create artificial intelligence, but in general, "artificial system" is an artificial system in the general sense. While in terms of "intelligence", this relates to other issues such as CONSCIOUSNESS (conscieusses), SELF (SELF), thinking (MIND), including UNCONSCIOUS thinking (unconsciousness _ MIND), etc., the only intelligence that a person knows is that of the person himself, which is a generally agreed view. However, people have limited understanding of their own intelligence, and have limited knowledge of the necessary elements that make up a person's intelligence, so it is difficult to define what is "artificial" manufactured "intelligence. The study of artificial intelligence therefore often involves the study of the intelligence itself of a person. Other intelligence related to animals or other man-made systems is also commonly recognized as a research topic related to artificial intelligence.
Artificial intelligence has gained increasing attention in the computer field. And the method is applied to robots, economic and political decisions, control systems and simulation systems. Artificial intelligence is a branch of computer science and has been called one of the three most advanced technologies (space technology, energy technology, artificial intelligence) in the world since the seventies of the twentieth century. It is also considered to be one of the three most advanced technologies (genetic engineering, nanoscience, artificial intelligence) in the twenty-first century. This is because it has been rapidly developed in the last three decades, has been widely used in many disciplines, and has achieved great results, and artificial intelligence has gradually become an independent branch, both theoretically and practically, becoming a system.
However, there are still many problems to be explored and overcome between the theoretical recognition of artificial intelligence and the maturity of the application of artificial intelligence. For example, when it is desired to perform intelligent determination of human behavior in a equipment display room based on an artificial intelligence model, for example, various human behaviors including viewing tags, stealing equipment, destroying equipment, strolling back and forth, and replaying equipment, on the one hand, it is difficult to select an appropriate artificial intelligence model suitable for human behavior determination, and on the other hand, lack of an effective training mechanism results in failure to perform information mapping from data collected on site to human behavior categories.
Disclosure of Invention
In order to solve the problems, the invention provides a human body behavior artificial intelligence judgment system and a human body behavior artificial intelligence judgment method, wherein a deep belief network model of a targeted structure is selected to establish a mapping relation between a human body image and human body behaviors in an equipment display room, and on the basis, various human body behaviors in the equipment display room and a preselected number of human body contour pictures corresponding to each human body behavior are adopted as training data of the deep belief network model, so that the intelligent judgment of the human body behaviors from a time-sharing digital image of the same human body object in the equipment display room to the human body object is completed.
Compared with the prior art, the invention at least needs to have the following two key points:
(1) adopting a depth belief network model of a targeted structure to establish a mapping relation between a human body image and a human body behavior so as to finish intelligent judgment of the human body behavior from the time-sharing digital image of the same human body object in the equipment display room to the human body object;
(2) and training the deep belief network model by adopting a customized training mechanism to obtain a high-precision intelligent-level reliable behavior judgment mechanism, wherein each human body behavior and a preselected number of human body contour pictures corresponding to each human body behavior are adopted to perform one training on the deep belief network model so as to finish each training corresponding to each human body behavior, and the trained deep belief network model is obtained.
According to a first aspect of the present invention, there is provided a human behavior artificial intelligence determination system, the system comprising:
the data extraction equipment is arranged in a control room of the equipment display room and is used for acquiring various human body behaviors in the equipment display room and a preselected number of human body contour pictures corresponding to each human body behavior, the resolution of each human body contour picture is equal, and the various human body behaviors comprise watching a label, stealing equipment, destroying equipment, strolling back and forth and replaying equipment;
the parameter adjusting device is arranged in the control room, is connected with the data extraction device and is used for constructing a deep belief network model with initial model parameters, and adopts each human body behavior obtained by the data extraction device and a preselected number of human body contour pictures corresponding to each human body behavior to perform one training on the deep belief network model so as to finish each training corresponding to each human body behavior and obtain the trained deep belief network model;
the content acquisition equipment is arranged in the equipment showroom and used for executing image acquisition action on an internal scene of the equipment showroom at the current acquisition time so as to obtain a corresponding content scene image;
the object segmentation equipment is connected with the content acquisition equipment and used for identifying each human body image block occupied by each human body object in the content scene image based on preset human body imaging characteristics;
the behavior judgment device is arranged in the control room, connected with the object segmentation device and used for taking the preselected number of human body image blocks of the same human body object in each content scene image respectively corresponding to each acquisition time within a preset time length as a plurality of input data of the trained deep belief network model, executing the trained deep belief network model to take single output data of the deep belief network model as the human body behavior type of the human body object within the preset time length, and enabling the preset time length to be equal to the product of the preselected number and the time interval between two adjacent acquisition times uniformly acquired by the content acquisition device;
the remote alarm equipment is electrically connected with the behavior judgment equipment, is connected with a big data server responsible for maintaining the equipment showroom, and is used for compressing and transmitting each content scene image corresponding to each acquisition time within a preset time length to the big data server when the human behavior type output by the behavior judgment equipment is equipment theft or equipment damage;
wherein, in the parameter adjustment device, the deep belief network model has a plurality of input data, a single output data, and a plurality of hidden layers;
wherein, in the deep belief network model, the total number of the input data is equal to the preselected number, and the total number of the hidden layers is proportional to the total number of the various human behaviors;
when each training is executed, a preselected number of human body contour pictures corresponding to related human body behaviors are used as a preselected number of input data of the deep belief network model, and the related human body behaviors are used as single output data of the deep belief network model.
According to a second aspect of the invention, a human behavior artificial intelligence judgment method is provided, and the method comprises the step of using the human behavior artificial intelligence judgment platform to adopt a depth belief network model with a targeted structure to finish intelligent judgment of human behavior from a time-sharing digital image of the same human object in a display room to the human object.
According to a third aspect of the present invention, there is provided a computer-readable storage medium having stored thereon a computer program which, when executed, implements the steps of the human behavior artificial intelligence determination method as described above.
Drawings
Embodiments of the invention will now be described with reference to the accompanying drawings, in which:
fig. 1 is a technical flowchart of a human behavior artificial intelligence determination system and method according to the present invention.
Fig. 2 is a schematic structural diagram of a human behavior artificial intelligence determination system according to embodiment 1 of the present invention.
Fig. 3 is a schematic structural diagram of a human behavior artificial intelligence determination system according to embodiment 2 of the present invention.
Fig. 4 is a schematic structural diagram of a human behavior artificial intelligence determination system according to embodiment 3 of the present invention.
Fig. 5 is a schematic structural diagram of a human behavior artificial intelligence determination system according to embodiment 4 of the present invention.
Fig. 6 is a schematic diagram of a computer-readable storage medium shown in embodiment 6 of the present invention.
Detailed Description
Artificial intelligence is the subject of research on making computer to simulate some human thinking process and intelligent behavior (such as learning, reasoning, thinking, planning, etc.), and mainly includes the principle of computer to implement intelligence and the manufacture of computer similar to human brain intelligence to make computer implement higher-level application.
Artificial intelligence will relate to computer science, psychology, philosophy and linguistics. The artificial intelligence is in the technical application level of thinking science and is an application branch of the science. From the thinking point of view, artificial intelligence is not limited to logic thinking, the breakthrough development of the artificial intelligence can be promoted only by considering image thinking and inspiration thinking, mathematics is generally considered as the basic science of various disciplines, the mathematics also enters the fields of language and thinking, the artificial intelligence disciplines also need to borrow mathematical tools, the mathematics not only plays a role in the ranges of standard logic, fuzzy mathematics and the like, the mathematics enters the artificial intelligence disciplines, and the mathematics can be promoted and developed more quickly.
For example, heavy scientific and engineering calculations are born by the human brain, and nowadays, computers can not only complete the calculations, but also can be faster and more accurate than the human brain, so that contemporary people no longer regard the calculations as "complex tasks that need human intelligence to complete", it can be seen that the definition of complex work changes with the development of the times and the progress of technology, and the specific goals of science, artificial intelligence, also naturally develop with the changes of the times. He is on the one hand constantly gaining new progress and on the other hand turning to more meaningful, more difficult goals.
In particular to application of artificial intelligence to an equipment showroom, an equipment maintainer hopes to intelligently analyze the current human behavior type of each human object in the equipment showroom in an unattended scene, wherein various human behaviors comprise watching labels, stealing equipment, destroying equipment, strolling and replaying equipment and the like, so as to execute corresponding alarm operation and information recording operation when determining that the equipment is stolen or the equipment is destroyed, and therefore the management cost of the equipment showroom is reduced. However, the prior art lacks a corresponding artificial intelligence judgment model.
In order to overcome the defects, the invention builds a human behavior artificial intelligence judgment system and a human behavior artificial intelligence judgment method, a deep belief network model is established through comparison of the artificial intelligence judgment models and is used for judging the current human behavior type of each human body object in the equipment display room in an unattended scene, and corresponding alarm operation and information recording operation are executed when the condition that the equipment is stolen or the equipment is damaged is determined, so that the maintenance of the equipment display room order is formed, wherein the important point is that the training data of the deep belief network model is stored and used, and the accuracy and the reliability of the human behavior type judgment are ensured.
As shown in fig. 1, a technical flowchart of the human behavior artificial intelligence determination system and method according to the present invention is shown.
Firstly, a large amount of training data for the artificial intelligence judgment model, namely known human behaviors including viewing labels, stealing equipment, destroying equipment, strolling back and forth, replaying equipment and the like, and a preselected number of human outline pictures corresponding to each known human behavior are stored
Secondly, constructing an artificial intelligence judgment model based on a deep belief network, wherein the artificial intelligence judgment model is provided with a plurality of input data, single output data and a plurality of hidden layers, as shown in fig. 1, the artificial intelligence judgment model is provided with p input data and single output data, the data in the hidden layers are represented by p, the total number of the input data is equal to the preselected number, the total number of the hidden layers is in direct proportion to the total number of various human behaviors to realize the matching of judgment complexity, the preselected number of human body contour pictures corresponding to the related human behaviors are used as the preselected number of input data of the model, and the related human behaviors are used as the single output data of the model;
thirdly, finishing multiple times of training of the model by using the reserved training data for judging the model by artificial intelligence so as to ensure the precision of the model;
finally, a preselected number of human images of each human subject at the scene are acquired, as shown in FIG. 1, using acquisition elements distributed at different locations of the display room, such as RGB sensing devices with wireless communication capabilities, image acquisition is carried out on different areas of the equipment showroom, the acquired different images are spliced to obtain images to be processed for subsequent human body object analysis, alternatively, panoramic image acquisition is performed on the equipment showroom using a single panoramic acquisition component and the acquired panoramic image is used for subsequent human subject analysis, and then, inputting the preselected number of human body images into the trained model to finish the intelligent judgment of the current behavior of each human body object of the display room under the unmanned management scene, and when the judgment result is the abnormal behavior of stealing or destroying the equipment, sending corresponding alarm information to the big data server through the remote alarm equipment so as to execute corresponding alarm operation.
The method has the key points that on one hand, a deep belief network capable of completing information mapping from time-sharing digital images of the same human body object in the equipment showroom to the human body behavior category of the human body object is selected for building an artificial intelligence judgment model, on the other hand, enough and effective training data are stored to realize the targeted training of the built artificial intelligence judgment model, so that the precision of the trained model is ensured, an electronic monitoring mechanism is adopted to replace an artificial monitoring mechanism to detect and alarm the abnormal human body behavior in the equipment showroom in an unmanned management scene, and the labor cost is reduced for the operation of the equipment showroom.
Hereinafter, the human behavior artificial intelligence determination system and method of the present invention will be described in detail by way of example.
Example 1
Fig. 2 is a schematic structural diagram of a human behavior artificial intelligence determination system according to embodiment 1 of the present invention.
As shown in fig. 2, the human behavior artificial intelligence determination system includes the following components:
the data extraction equipment is arranged in a control room of the equipment display room and is used for acquiring various human body behaviors in the equipment display room and a preselected number of human body contour pictures corresponding to each human body behavior, the resolution of each human body contour picture is equal, and the various human body behaviors comprise watching a label, stealing equipment, destroying equipment, strolling back and forth and replaying equipment;
the parameter adjusting device is arranged in the control room, is connected with the data extraction device and is used for constructing a deep belief network model with initial model parameters, and adopts each human body behavior obtained by the data extraction device and a preselected number of human body contour pictures corresponding to each human body behavior to perform one training on the deep belief network model so as to finish each training corresponding to each human body behavior and obtain the trained deep belief network model;
the content acquisition equipment is arranged in the equipment showroom and used for executing image acquisition action on an internal scene of the equipment showroom at the current acquisition time so as to obtain a corresponding content scene image;
the object segmentation equipment is connected with the content acquisition equipment and used for identifying each human body image block occupied by each human body object in the content scene image based on preset human body imaging characteristics;
the behavior judgment device is arranged in the control room, connected with the object segmentation device and used for taking the preselected number of human body image blocks of the same human body object in each content scene image respectively corresponding to each acquisition time within a preset time length as a plurality of input data of the trained deep belief network model, executing the trained deep belief network model to take single output data of the deep belief network model as the human body behavior type of the human body object within the preset time length, and enabling the preset time length to be equal to the product of the preselected number and the time interval between two adjacent acquisition times uniformly acquired by the content acquisition device;
the remote alarm equipment is electrically connected with the behavior judgment equipment, is connected with a big data server responsible for maintaining the equipment showroom, and is used for compressing and transmitting each content scene image corresponding to each acquisition time within a preset time length to the big data server when the human behavior type output by the behavior judgment equipment is equipment theft or equipment damage;
wherein, in the parameter adjustment device, the deep belief network model has a plurality of input data, a single output data, and a plurality of hidden layers;
wherein, in the deep belief network model, the total number of the input data is equal to the preselected number, and the total number of the hidden layers is proportional to the total number of the various human behaviors;
when each training is executed, a preselected number of human body contour pictures corresponding to related human body behaviors are used as a preselected number of input data of the deep belief network model, and the related human body behaviors are used as single output data of the deep belief network model.
Example 2
Fig. 3 is a schematic structural diagram of a human behavior artificial intelligence determination system according to embodiment 2 of the present invention.
As shown in fig. 3, compared with embodiment 1 of the present invention, the human behavior artificial intelligence determination system further includes:
and the feature storage device is connected with the object segmentation device and is used for storing preset human body imaging features, and the preset human body imaging features are preset human body color imaging features and/or preset human body appearance imaging features.
Example 3
Fig. 4 is a schematic structural diagram of a human behavior artificial intelligence determination system according to embodiment 3 of the present invention.
As shown in fig. 4, compared with embodiment 1 of the present invention, the human behavior artificial intelligence determination system further includes:
and the big data server is used for simultaneously maintaining each equipment display room in a certain urban area, and each remote alarm device corresponding to each equipment display room which is in charge of maintenance is connected through a wireless network.
Example 4
Fig. 5 is a schematic structural diagram of a human behavior artificial intelligence determination system according to embodiment 4 of the present invention.
As shown in fig. 5, compared with embodiment 1 of the present invention, the human behavior artificial intelligence determination system further includes:
and the image upgrading equipment is arranged between the content acquisition equipment and the object segmentation equipment and is used for performing Gaussian white noise filtering processing and impulse noise filtering processing on the received content scene image to obtain a directional de-noised image, and replacing the corresponding content scene image with the directional de-noised image and inputting the directional de-noised image into the object segmentation equipment.
In any embodiment of the foregoing embodiments, optionally, in the human behavior artificial intelligence determination system:
performing a training on the deep belief network model once by using each human behavior acquired by the data extraction device and a preselected number of human profile pictures corresponding to each human behavior to complete each training corresponding to each human behavior, wherein acquiring the trained deep belief network model comprises: and correspondingly adjusting the model parameters of the obtained deep belief network model in a matching way with the training every time the training is finished.
In any embodiment of the foregoing embodiments, optionally, in the human behavior artificial intelligence determination system:
when each training is executed, the method takes the human body outline pictures with the preselected quantity corresponding to the related human body behaviors as the input data with the preselected quantity of the deep belief network model, and the step of taking the related human body behaviors as the single output data of the deep belief network model comprises the following steps: each input data is a binary code of a relevant human body outline picture, and the single output data is a binary code of a relevant human body behavior;
wherein, every input data is the binary code of relevant human body profile picture, and single output data is the binary code of relevant human body action and includes: the binary code of the relevant human body contour picture is a binary value obtained by sequentially connecting binary values after binary representation of pixel values of pixel points of the human body contour picture according to the order of left-right first and top-bottom first, and the binary code of the relevant human body behavior is a binary value corresponding to a character string of the relevant human body behavior.
In the preselected number of human body contour pictures corresponding to each human body behavior, each frame of human body contour picture only comprises a single human body contour.
In any embodiment of the foregoing embodiments, optionally, in the human behavior artificial intelligence determination system:
identifying each human body image block occupied by each human body object in the content scene image based on preset human body imaging characteristics comprises: when the preset human body imaging characteristic is a preset human body color imaging characteristic, taking a pixel point of the content scene image, which has a value of a color channel matched with the preset human body color imaging characteristic, as a single pixel point forming a human body image block;
wherein identifying each human body image block occupied by each human body object in the content scene image based on preset human body imaging characteristics further comprises: when the preset human body imaging feature is a preset human body appearance imaging feature, taking an image block with an edge shape in the content scene image matched with an edge shape of a picture corresponding to the preset human body appearance imaging feature as a single human body image block;
the binary code of the relevant human body contour picture is a binary value obtained by sequentially connecting binary values after binary representation of each pixel value of each pixel point of the human body contour picture according to the order of left, right, top and bottom, and the binary code of the relevant human body behavior is a binary value corresponding to a character string of the relevant human body behavior and comprises the following steps: binary values obtained by binary representation of the pixel value of each pixel point of the human body contour picture are obtained by sequentially connecting the binary values obtained after binary representation of each color channel of the pixel point;
wherein, the binary number value obtained by binary representation of the pixel value of each pixel point of the human body contour picture is the binary number value obtained by sequentially connecting the binary number values obtained after binary representation of each color channel of the pixel point respectively, and comprises: when each color channel is an RGB color channel, the binary values obtained by sequentially connecting the binary values obtained after binary representation of each color channel of the pixel point are the binary values obtained by sequentially connecting the binary values obtained after binary representation of each color channel of the pixel point, the R color channel, the G color channel, and the B color channel, according to the connection sequence of RGB.
Example 5
In this embodiment, the invention builds a human behavior artificial intelligence judgment method, which comprises the step of using the human behavior artificial intelligence judgment platform to adopt a deep belief network model with a targeted structure to finish the intelligent judgment of human behavior from a time-sharing digital image of the same human object in a display room to the human object.
Example 6
Fig. 6 is a schematic diagram of a computer-readable storage medium shown in embodiment 6 of the present invention. As shown in fig. 6, a computer-readable storage medium 60, having non-transitory computer-readable instructions 61 stored thereon, in accordance with an embodiment of the present disclosure. When the non-transitory computer readable instructions 61 are executed by a processor, all or part of the steps of the human behavior artificial intelligence determination method of embodiment 6 of the present invention are performed.
In addition, Deep Belief Networks (DBN) are proposed by Geoffrey Hinton. It is a generative model, and by training the weights between its neurons, one can let the whole neural network generate training data according to the maximum probability. One can use the DBN not only to identify features, classify data, but also to generate data with him.
DBNs are composed of a plurality of layers of neurons, which are further classified into dominant neurons and recessive neurons (hereinafter, referred to as dominant neurons and recessive neurons). The explicit element is used for accepting input, and the implicit element is used for extracting features. Hence, the hidden elements also have individual names, called feature detectors (features detectors). The connections between the top two layers are undirected and constitute an associative memory (associative memory). There are upward and downward directed connections between other lower layers. The bottom most layer represents data vectors (data vectors), each neuron representing a dimension of the data vector.
The constituent elements of the DBN are Restricted Boltzmann Machines (RBM). The process of training the DBN is performed layer by layer. In each layer, a hidden layer is deduced by using a data vector, and the hidden layer is regarded as a data vector of the next layer (higher layer). As previously mentioned, RBM is a component of DBN. In fact, each RBM can be used individually as a clusterer. The RBM has only two layers of neurons, one layer is called a visual layer and consists of visual units (visual units) for inputting training data. The other layer is called Hidden layer (Hidden layer) and, correspondingly, consists of Hidden elements (Hidden units) which serve as feature detectors (features detectors).
It should be noted that the neurons within both the explicit and hidden layers are not interconnected, and only the neurons between the layers have symmetric connecting lines. This has the advantage that, given the values of all primitives, it is irrelevant what value each hidden primitive takes. That is, also, when a hidden layer is given, the values of all the primitives are not related to each other: with this important property, one does not need to compute one value for each neuron, but compute the neurons in the whole layer in parallel.
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
Through the above description of the embodiments, those skilled in the art will clearly understand that the method of the above embodiments can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware, but in many cases, the former is a better implementation manner. Based on such understanding, the technical solutions of the present invention may be embodied in the form of a software product, which is stored in a storage medium (such as ROM/RAM, magnetic disk, optical disk) and includes instructions for enabling a terminal (such as a mobile phone, a computer, a server, an air conditioner, or a network device) to execute the method according to the embodiments of the present invention.
While the present invention has been described with reference to the embodiments shown in the drawings, the present invention is not limited to the embodiments, which are illustrative and not restrictive, and it will be apparent to those skilled in the art that various changes and modifications can be made therein without departing from the spirit and scope of the invention as defined in the appended claims.
Claims (10)
1. An artificial intelligence judgment system for human behavior, the system comprising:
the data extraction equipment is arranged in a control room of the equipment display room and is used for acquiring various human body behaviors in the equipment display room and a preselected number of human body contour pictures corresponding to each human body behavior, the resolution of each human body contour picture is equal, and the various human body behaviors comprise watching a label, stealing equipment, destroying equipment, strolling back and forth and replaying equipment;
the parameter adjusting device is arranged in the control room, is connected with the data extraction device and is used for constructing a deep belief network model with initial model parameters, and adopts each human body behavior obtained by the data extraction device and a preselected number of human body contour pictures corresponding to each human body behavior to perform one training on the deep belief network model so as to finish each training corresponding to each human body behavior and obtain the trained deep belief network model;
the content acquisition equipment is arranged in the equipment showroom and used for executing image acquisition action on an internal scene of the equipment showroom at the current acquisition time so as to obtain a corresponding content scene image;
the object segmentation equipment is connected with the content acquisition equipment and used for identifying each human body image block occupied by each human body object in the content scene image based on preset human body imaging characteristics;
the behavior judgment device is arranged in the control room, connected with the object segmentation device and used for taking the preselected number of human body image blocks of the same human body object in each content scene image respectively corresponding to each acquisition time within a preset time length as a plurality of input data of the trained deep belief network model, executing the trained deep belief network model to take single output data of the deep belief network model as the human body behavior type of the human body object within the preset time length, and enabling the preset time length to be equal to the product of the preselected number and the time interval between two adjacent acquisition times uniformly acquired by the content acquisition device;
the remote alarm equipment is electrically connected with the behavior judgment equipment, is connected with a big data server responsible for maintaining the equipment showroom, and is used for compressing and transmitting each content scene image corresponding to each acquisition time within a preset time length to the big data server when the human behavior type output by the behavior judgment equipment is equipment theft or equipment damage;
wherein, in the parameter adjustment device, the deep belief network model has a plurality of input data, a single output data, and a plurality of hidden layers;
wherein, in the deep belief network model, the total number of the input data is equal to the preselected number, and the total number of the hidden layers is proportional to the total number of the various human behaviors;
when each training is executed, a preselected number of human body contour pictures corresponding to related human body behaviors are used as a preselected number of input data of the deep belief network model, and the related human body behaviors are used as single output data of the deep belief network model.
2. The human behavior artificial intelligence judgment system of claim 1, wherein the system further comprises:
and the feature storage device is connected with the object segmentation device and is used for storing preset human body imaging features, and the preset human body imaging features are preset human body color imaging features and/or preset human body appearance imaging features.
3. The human behavior artificial intelligence judgment system of claim 1, wherein the system further comprises:
and the big data server is used for simultaneously maintaining each equipment display room in a certain urban area, and each remote alarm device corresponding to each equipment display room which is in charge of maintenance is connected through a wireless network.
4. The human behavior artificial intelligence judgment system of claim 1, wherein the system further comprises:
and the image upgrading equipment is arranged between the content acquisition equipment and the object segmentation equipment and is used for performing Gaussian white noise filtering processing and impulse noise filtering processing on the received content scene image to obtain a directional de-noised image, and replacing the corresponding content scene image with the directional de-noised image and inputting the directional de-noised image into the object segmentation equipment.
5. The human behavior artificial intelligence judgment system according to any one of claims 1 to 4, characterized in that:
performing a training on the deep belief network model once by using each human behavior acquired by the data extraction device and a preselected number of human profile pictures corresponding to each human behavior to complete each training corresponding to each human behavior, wherein acquiring the trained deep belief network model comprises: and correspondingly adjusting the model parameters of the obtained deep belief network model in a matching way with the training every time the training is finished.
6. The human behavior artificial intelligence judgment system according to any one of claims 1 to 4, characterized in that:
when each training is executed, the method takes the human body outline pictures with the preselected quantity corresponding to the related human body behaviors as the input data with the preselected quantity of the deep belief network model, and the step of taking the related human body behaviors as the single output data of the deep belief network model comprises the following steps: each input data is a binary code of a relevant human body outline picture, and the single output data is a binary code of a relevant human body behavior;
wherein, every input data is the binary code of relevant human body profile picture, and single output data is the binary code of relevant human body action and includes: the binary code of the relevant human body contour picture is a binary value obtained by sequentially connecting binary values after binary representation of pixel values of pixel points of the human body contour picture according to the order of left and right, top and bottom, and the binary code of the relevant human body behavior is a binary value corresponding to a character string of the relevant human body behavior, wherein in a preselected number of human body contour pictures corresponding to each human body behavior, each frame of human body contour picture only comprises a single human body contour.
7. The human behavior artificial intelligence judgment system according to any one of claims 1 to 4, characterized in that:
identifying each human body image block occupied by each human body object in the content scene image based on preset human body imaging characteristics comprises: when the preset human body imaging characteristic is a preset human body color imaging characteristic, taking a pixel point of the content scene image, which has a value of a color channel matched with the preset human body color imaging characteristic, as a single pixel point forming a human body image block;
wherein identifying each human body image block occupied by each human body object in the content scene image based on preset human body imaging characteristics further comprises: and when the preset human body imaging feature is the preset human body shape imaging feature, taking the image block of which the edge shape in the content scene image is matched with the edge shape of the picture corresponding to the preset human body shape imaging feature as a single human body image block.
8. The human behavior artificial intelligence judgment system of claim 7, characterized in that:
the binary code of the relevant human body contour picture is a binary value obtained by sequentially connecting binary values after binary representation of pixel values of pixel points of the human body contour picture according to the order of left and right, and top and bottom, and the binary code of the relevant human body behavior is a binary value corresponding to a character string of the relevant human body behavior and comprises the following steps: binary values obtained by binary representation of the pixel value of each pixel point of the human body contour picture are obtained by sequentially connecting the binary values obtained after binary representation of each color channel of the pixel point;
wherein, the binary number value obtained by binary representation of the pixel value of each pixel point of the human body contour picture is the binary number value obtained by sequentially connecting the binary number values obtained after binary representation of each color channel of the pixel point respectively, and comprises: when each color channel is an RGB color channel, the binary values obtained by sequentially connecting the binary values obtained after binary representation of each color channel of the pixel point are the binary values obtained by sequentially connecting the binary values obtained after binary representation of each color channel of the pixel point, the R color channel, the G color channel, and the B color channel, according to the connection sequence of RGB.
9. A human behavior artificial intelligence judgment method, comprising using the human behavior artificial intelligence judgment platform according to any one of claims 1 to 8 to perform intelligent judgment of human behavior from time-sharing digital images of the same human object to the human object in a display room by adopting a deep belief network model with a targeted structure.
10. A computer-readable storage medium, having stored thereon a computer program which, when executed, performs the steps of the method of claim 9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111463690.0A CN114155555B (en) | 2021-12-02 | 2021-12-02 | Human behavior artificial intelligence judgment system and method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111463690.0A CN114155555B (en) | 2021-12-02 | 2021-12-02 | Human behavior artificial intelligence judgment system and method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114155555A true CN114155555A (en) | 2022-03-08 |
| CN114155555B CN114155555B (en) | 2022-06-10 |
Family
ID=80456072
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111463690.0A Active CN114155555B (en) | 2021-12-02 | 2021-12-02 | Human behavior artificial intelligence judgment system and method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114155555B (en) |
Citations (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102831442A (en) * | 2011-06-13 | 2012-12-19 | 索尼公司 | Abnormal behavior detection method and equipment and method and equipment for generating abnormal behavior detection equipment |
| CN104200197A (en) * | 2014-08-18 | 2014-12-10 | 北京邮电大学 | Three-dimensional human body behavior recognition method and device |
| CN104850846A (en) * | 2015-06-02 | 2015-08-19 | 深圳大学 | Human behavior recognition method and human behavior recognition system based on depth neural network |
| CN107403200A (en) * | 2017-08-10 | 2017-11-28 | 北京亚鸿世纪科技发展有限公司 | Improve the multiple imperfect picture sorting technique of image segmentation algorithm combination deep learning |
| CN108399367A (en) * | 2018-01-31 | 2018-08-14 | 深圳市阿西莫夫科技有限公司 | Hand motion recognition method, apparatus, computer equipment and readable storage medium storing program for executing |
| CN108537145A (en) * | 2018-03-21 | 2018-09-14 | 东北电力大学 | Human bodys' response method based on space-time skeleton character and depth belief network |
| CN109740446A (en) * | 2018-12-14 | 2019-05-10 | 深圳壹账通智能科技有限公司 | Classroom students ' behavior analysis method and device |
| CN109961037A (en) * | 2019-03-20 | 2019-07-02 | 中共中央办公厅电子科技学院(北京电子科技学院) | A kind of examination hall video monitoring abnormal behavior recognition methods |
| CN110765964A (en) * | 2019-10-30 | 2020-02-07 | 常熟理工学院 | Method for detecting abnormal behaviors in elevator car based on computer vision |
| CN111325281A (en) * | 2020-03-05 | 2020-06-23 | 新希望六和股份有限公司 | Deep learning network training method and device, computer equipment and storage medium |
| CN111738218A (en) * | 2020-07-27 | 2020-10-02 | 成都睿沿科技有限公司 | Human body abnormal behavior recognition system and method |
| CN111832386A (en) * | 2020-05-22 | 2020-10-27 | 大连锐动科技有限公司 | Method and device for estimating human body posture and computer readable medium |
| CN111860460A (en) * | 2020-08-05 | 2020-10-30 | 江苏新安电器股份有限公司 | Application method of improved LSTM model in human behavior recognition |
| CN112132009A (en) * | 2020-09-22 | 2020-12-25 | 湖南惟楚有才教育科技有限公司 | Classroom behavior analysis method and system and electronic equipment |
| CN112183438A (en) * | 2020-10-13 | 2021-01-05 | 深圳龙岗智能视听研究院 | Image identification method for illegal behaviors based on small sample learning neural network |
| CN112277957A (en) * | 2020-10-27 | 2021-01-29 | 广州汽车集团股份有限公司 | Early warning method and system for driver distraction correction and storage medium |
| CN112311952A (en) * | 2020-08-25 | 2021-02-02 | 北京沃东天骏信息技术有限公司 | Image processing method, system and device |
| CN113610878A (en) * | 2021-07-15 | 2021-11-05 | 南京信息工程大学 | Structural information hiding method based on automatic contour generation |
| CN113688676A (en) * | 2021-07-20 | 2021-11-23 | 杭州兼果网络科技有限公司 | Artificial intelligence-based suspicious behavior identification method and cloud server |
-
2021
- 2021-12-02 CN CN202111463690.0A patent/CN114155555B/en active Active
Patent Citations (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102831442A (en) * | 2011-06-13 | 2012-12-19 | 索尼公司 | Abnormal behavior detection method and equipment and method and equipment for generating abnormal behavior detection equipment |
| CN104200197A (en) * | 2014-08-18 | 2014-12-10 | 北京邮电大学 | Three-dimensional human body behavior recognition method and device |
| CN104850846A (en) * | 2015-06-02 | 2015-08-19 | 深圳大学 | Human behavior recognition method and human behavior recognition system based on depth neural network |
| CN107403200A (en) * | 2017-08-10 | 2017-11-28 | 北京亚鸿世纪科技发展有限公司 | Improve the multiple imperfect picture sorting technique of image segmentation algorithm combination deep learning |
| CN108399367A (en) * | 2018-01-31 | 2018-08-14 | 深圳市阿西莫夫科技有限公司 | Hand motion recognition method, apparatus, computer equipment and readable storage medium storing program for executing |
| CN108537145A (en) * | 2018-03-21 | 2018-09-14 | 东北电力大学 | Human bodys' response method based on space-time skeleton character and depth belief network |
| CN109740446A (en) * | 2018-12-14 | 2019-05-10 | 深圳壹账通智能科技有限公司 | Classroom students ' behavior analysis method and device |
| CN109961037A (en) * | 2019-03-20 | 2019-07-02 | 中共中央办公厅电子科技学院(北京电子科技学院) | A kind of examination hall video monitoring abnormal behavior recognition methods |
| CN110765964A (en) * | 2019-10-30 | 2020-02-07 | 常熟理工学院 | Method for detecting abnormal behaviors in elevator car based on computer vision |
| CN111325281A (en) * | 2020-03-05 | 2020-06-23 | 新希望六和股份有限公司 | Deep learning network training method and device, computer equipment and storage medium |
| CN111832386A (en) * | 2020-05-22 | 2020-10-27 | 大连锐动科技有限公司 | Method and device for estimating human body posture and computer readable medium |
| CN111738218A (en) * | 2020-07-27 | 2020-10-02 | 成都睿沿科技有限公司 | Human body abnormal behavior recognition system and method |
| CN111860460A (en) * | 2020-08-05 | 2020-10-30 | 江苏新安电器股份有限公司 | Application method of improved LSTM model in human behavior recognition |
| CN112311952A (en) * | 2020-08-25 | 2021-02-02 | 北京沃东天骏信息技术有限公司 | Image processing method, system and device |
| CN112132009A (en) * | 2020-09-22 | 2020-12-25 | 湖南惟楚有才教育科技有限公司 | Classroom behavior analysis method and system and electronic equipment |
| CN112183438A (en) * | 2020-10-13 | 2021-01-05 | 深圳龙岗智能视听研究院 | Image identification method for illegal behaviors based on small sample learning neural network |
| CN112277957A (en) * | 2020-10-27 | 2021-01-29 | 广州汽车集团股份有限公司 | Early warning method and system for driver distraction correction and storage medium |
| CN113610878A (en) * | 2021-07-15 | 2021-11-05 | 南京信息工程大学 | Structural information hiding method based on automatic contour generation |
| CN113688676A (en) * | 2021-07-20 | 2021-11-23 | 杭州兼果网络科技有限公司 | Artificial intelligence-based suspicious behavior identification method and cloud server |
Non-Patent Citations (1)
| Title |
|---|
| 耿驰: "基于深度学习的人体动作识别", 《中国优秀博硕士学位论文全文数据库(硕士) 信息科技辑》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114155555B (en) | 2022-06-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111709409B (en) | Face living body detection method, device, equipment and medium | |
| CN110728209B (en) | Gesture recognition method and device, electronic equipment and storage medium | |
| CN107463949B (en) | Video action classification processing method and device | |
| CN110263681B (en) | Facial expression recognition method and device, storage medium and electronic device | |
| CN112183577A (en) | Training method of semi-supervised learning model, image processing method and equipment | |
| JP6517681B2 (en) | Image pattern learning apparatus, method and program | |
| CN111368943B (en) | Method and device for identifying object in image, storage medium and electronic device | |
| CN110555481A (en) | Portrait style identification method and device and computer readable storage medium | |
| CN110197116B (en) | Human behavior recognition method, device and computer readable storage medium | |
| CN112560810A (en) | Micro-expression recognition method based on multi-scale space-time characteristic neural network | |
| CN110909680A (en) | Facial expression recognition method and device, electronic equipment and storage medium | |
| CN110084293A (en) | A kind of determination method and apparatus in complete bright pattern house | |
| CN111046213B (en) | Knowledge base construction method based on image recognition | |
| CN116977674A (en) | Image matching method, related device, storage medium and program product | |
| CN109740527B (en) | Image processing method in video frame | |
| CN114463837A (en) | Human behavior recognition method and system based on self-adaptive space-time convolution network | |
| CN115620122A (en) | Training method of neural network model, image re-recognition method and related equipment | |
| CN113673308A (en) | Object identification method, device and electronic system | |
| CN111612090B (en) | Image emotion classification method based on content color cross correlation | |
| CN110348395B (en) | Skeleton behavior identification method based on space-time relationship | |
| CN114155555B (en) | Human behavior artificial intelligence judgment system and method | |
| CN112016592A (en) | Domain adaptive semantic segmentation method and device based on cross domain category perception | |
| CN113283306B (en) | Rodent identification analysis method based on deep learning and migration learning | |
| CN115578624A (en) | Agricultural disease and pest model construction method, detection method and device | |
| CN114445691A (en) | Model training method and device, electronic equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information |
Address after: Room 327, third floor, No. 2, Stadium South Road, Shijingshan District, Beijing 100043 Applicant after: Beijing Zhongke Zhiyi Technology Co.,Ltd. Address before: Room 330, third floor, No. 2, Stadium South Road, Shijingshan District, Beijing 100043 Applicant before: Beijing Zhongke Zhiyi Technology Co.,Ltd. |
|
| CB02 | Change of applicant information | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CP01 | Change in the name or title of a patent holder |
Address after: Room 327, third floor, No. 2, Stadium South Road, Shijingshan District, Beijing 100043 Patentee after: Beijing Zhongke Zhiyi Technology Co.,Ltd. Address before: Room 327, third floor, No. 2, Stadium South Road, Shijingshan District, Beijing 100043 Patentee before: Beijing Zhongke Zhiyi Technology Co.,Ltd. |
|
| CP01 | Change in the name or title of a patent holder |