Animal behavior analysis method, device and electronic equipment

Abstract

The present invention discloses an animal behavior analysis method, which includes: obtaining the action information of the experimental animal before and after the experiment, and arranging the action information in chronological order to form an action sequence; performing action analysis on the action sequence from the perspectives of action frequency, action conversion probability, action timing distribution and disorder degree to obtain action analysis results; and comparing the action analysis results before and after the experiment to obtain the corresponding behavior analysis results of the experimental animal. The present invention analyzes multiple indicators of animal behavior by using multi-angle sequence analysis technology, and can obtain the sequential relationship of two or more actions of the experimental animal appearing in sequence in time within a specified time range, the probability of mutual conversion between different behavior actions, and the frequency distribution of different actions at different times. This multi-angle animal behavior analysis method obtains richer behavior information, makes the analysis more comprehensive and more accurate, and thus meets more animal behavior analysis needs.

Description

Technical Field

[0001] The present invention relates to the technical field of animal behavior, and in particular to an animal behavior analysis method, device and electronic equipment. Background Art

[0002] When it comes to the pathogenesis research of psychiatric diseases such as anxiety, depression and Parkinson's disease or the evaluation of drug efficacy, it is generally necessary to involve behavioral evaluation of disease animal models. The complex behaviors of mammals, including rodents, are composed of simple action modules combined in a certain time sequence. Taking laboratory mice as an example, the experiments currently widely used in the observation of rodent behavior include open field experiments, elevated maze experiments and forced swimming experiments. However, the animal behavioral data obtained based on these experiments are often linear and quantitative, such as the length of animal tracks, the number of grids crossed by animals, and the length of time they stay in the central area in open field experiments; the number of times animals enter open arms and closed arms in elevated maze experiments, the length of time they stay in open arms and closed arms, and the length of time animals stay still in forced swimming experiments. The information obtained from these linear quantitative data is relatively thin, and the evaluation indicators are single, resulting in incomplete analysis and low accuracy, which is difficult to meet the needs of animal behavior analysis. Summary of the invention

[0003] In view of this, the embodiments of the present invention provide an animal behavior analysis method, device and electronic device to solve the problem that the traditional animal behavior analysis indicator single analysis is incomplete, low in accuracy, and difficult to meet the needs of animal behavior analysis.

[0004] In order to achieve the above object, the present invention provides the following technical solutions:

[0005] The embodiment of the present invention provides an animal behavior analysis method, comprising:

[0006] Acquiring video information of the experimental animal before and after the experiment, processing the video information to obtain action information, and arranging the action information in chronological order to form an action sequence;

[0007] Performing motion analysis on the motion sequence from the perspectives of motion frequency, motion transition probability, motion timing distribution and disorder degree to obtain motion analysis results;

[0008] The action analysis results before and after the experiment are compared to obtain the corresponding behavior analysis results of the experimental animals.

[0009] Optionally, the preprocessing the video information to obtain the action information includes:

[0010] Performing frame segmentation on the video information to obtain segmented action images;

[0011] Dividing the action image into action sets to obtain a plurality of action image sets, each action image set including a plurality of continuous action images;

[0012] Performing image fusion on the action images in the action image set to obtain a fused action image set;

[0013] Extracting features from the fused action image set using a preset algorithm to obtain action feature information;

[0014] Substitute the action feature information into a preset classification model to obtain action information.

[0015] Optionally, performing motion analysis on the motion sequence from the perspectives of motion frequency, motion conversion probability and motion timing distribution to obtain the motion analysis result includes:

[0016] Performing frequent item set detection on the action sequence to obtain the degree of association between different actions, wherein the frequent item set is a set of action association combinations consisting of different numbers of actions based on the degree of association between the actions;

[0017] Performing statistics on changes in conversion probabilities of the action sequence based on a preset transfer matrix algorithm;

[0018] Counting the order and frequency changes of different actions in the action sequence to obtain the temporal distribution of action frequency;

[0019] By analyzing the continuous actions in the action sequence, the disorder degree of the actions is obtained.

[0020] Optionally, the detecting of frequent itemsets on the action sequence to obtain the degree of association between different actions includes:

[0021] Matching the actions in the action sequence based on each action association combination in the frequent item set to obtain a matching result of each action association combination;

[0022] The matching results are counted to obtain the frequencies of occurrence of different action association combinations, and the association degrees between different actions are determined based on the frequencies of occurrence of different action association combinations.

[0023] Optionally, the animal behavior analysis method further comprises:

[0024] The actions are associated and classified by establishing multiple hierarchies, where different hierarchies have different orders, and each hierarchy includes a number of different behavior modules, and the behavior modules contain a number of actions corresponding to the order of the corresponding hierarchy;

[0025] Based on the action sequence, the actions are classified and combined according to the hierarchy to form different behavior modules;

[0026] The frequent item set is formed by integrating the behavior modules.

[0027] Optionally, performing statistics on changes in conversion probabilities of the action sequence based on a preset transfer matrix algorithm includes:

[0028] Extracting actions appearing in the action sequence, and grouping the actions in pairs as a basis group for mutual conversion of actions;

[0029] Comparing each of the basis groups with the action sequence respectively to obtain the probability of occurrence of each of the basis groups;

[0030] The probability of occurrence of each basis set is statistically analyzed to obtain the action transition probability between different actions.

[0031] Optionally, the counting of the order and frequency changes of different actions in the action sequence to obtain the temporal distribution of action frequencies includes:

[0032] Extracting actions appearing in the action sequence;

[0033] Count the order and frequency changes of different actions, and get the probability of different actions appearing;

[0034] The probabilities are arranged in time series to obtain a time series distribution of the action frequencies.

[0035] The embodiment of the present invention further provides an animal behavior analysis device, comprising:

[0036] Acquisition module: used to acquire the action information of the experimental animals before and after the experiment, and arrange the action information in chronological order to form an action sequence;

[0037] Processing module: used to process the action sequence from the perspectives of action frequency, action conversion probability and action timing distribution to obtain processing results;

[0038] Analysis module: used to analyze the processing results before and after the experiment to obtain the final results.

[0039] An embodiment of the present invention further provides an electronic device, including:

[0040] A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the animal behavior analysis method provided by the embodiment of the present invention by executing the computer instructions.

[0041] An embodiment of the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable the computer to execute the animal behavior analysis method provided by the embodiment of the present invention.

[0042] The technical solution of the present invention has the following advantages:

[0043] The present invention provides an animal behavior analysis method, which obtains the action information of the experimental animal before and after the experiment, and arranges the action information in chronological order to form an action sequence; performs action analysis on the action sequence from the perspectives of action frequency, action conversion probability, action timing distribution and disorder degree to obtain action analysis results; compares the action analysis results before and after the experiment to obtain the corresponding behavior analysis results of the experimental animal. The present invention analyzes multiple indicators of animal behavior by using multiple angle sequence analysis techniques respectively, and can obtain the sequential relationship of two or more actions of the experimental animal appearing in sequence in time within a specified time range, the probability of mutual conversion between different behavior actions, and the frequency distribution of different actions at different times. This multi-angle animal behavior analysis method can obtain richer behavioral information, making the analysis more comprehensive and more accurate, thereby meeting more animal behavior analysis needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0045] Figure 1 is a flow chart of an animal behavior analysis method in an embodiment of the present invention;

[0046] Figure 2 A flowchart of preprocessing video information to obtain action information according to an embodiment of the present invention;

[0047] Figure 3 A flowchart of performing action analysis on an action sequence from the perspectives of action frequency, action conversion probability and action timing distribution to obtain an action analysis result according to an embodiment of the present invention;

[0048] Figure 4 is a flow chart of detecting frequent itemsets for an action sequence according to an embodiment of the present invention;

[0049] Figure 5 is a flow chart of forming frequent itemsets according to an embodiment of the present invention;

[0050] Figure 6 Schematic diagram of frequent item sets before and after modeling of a depression modeling group and a normal control group according to an embodiment of the present invention;

[0051] Figure 7 A schematic diagram of frequent item sets constituting the first week of modeling for a depression modeling group and a normal control group according to an embodiment of the present invention;

[0052] Figure 8 A schematic diagram of frequent item sets constituting the second week of modeling for a depression modeling group and a normal control group according to an embodiment of the present invention;

[0053] Fig. 9 A schematic diagram of frequent item sets constituting the third week of modeling for the depression modeling group and the normal control group according to an embodiment of the present invention;

[0054] Fig.10 A statistical graph of the frequency of frequent itemsets of different lengths according to an embodiment of the present invention;

[0055] Fig.11 A statistical graph of the frequency of frequent itemsets containing exploration actions according to an embodiment of the present invention;

[0056] Fig.12 A statistical graph of the frequency of frequent itemsets containing combing actions according to an embodiment of the present invention;

[0057] Fig.13 A statistical graph of the frequency of frequent itemsets containing stationary actions according to an embodiment of the present invention;

[0058] Fig.14 A flowchart of performing statistics on changes in conversion probabilities of action sequences based on a preset transfer matrix algorithm according to an embodiment of the present invention;

[0059] Fig.15 Schematic diagram of the action transfer probability of mice in the depression modeling group and the normal control group at different periods according to an embodiment of the present invention;

[0060] Fig.16 is a schematic diagram of action conversion probability according to an embodiment of the present invention;

[0061] Fig.17 A flowchart for obtaining the disorder degree of actions by counting the order and frequency changes of different actions in an action sequence according to an embodiment of the present invention;

[0062] Fig.18 is a schematic structural diagram of an animal behavior analysis device in an embodiment of the present invention;

[0063] Fig.19 Schematic diagram of the structure of an electronic device in an embodiment of the present invention. DETAILED DESCRIPTION

[0064] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0065] According to an embodiment of the present invention, an embodiment of an animal behavior analysis method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0066] In this embodiment, a method for analyzing animal behavior is provided. Figure 1 As shown, the animal behavior analysis method comprises the following steps:

[0067] Step S1: Obtain video information of the experimental animal before and after the experiment, process the video information to obtain action information, and arrange the action information in chronological order to form an action sequence. Specifically, the action information includes the actions of the experimental animal within a certain time range. Sorting the actions by time is more conducive to analyzing the relevance of the experimental animal's actions.

[0068] Step S2: Performing motion analysis on the motion sequence from the perspectives of motion frequency, motion conversion probability, motion timing distribution and disorder degree to obtain motion analysis results. Specifically, by analyzing from the perspectives of motion frequency, motion conversion probability and motion timing distribution, we can obtain the sequential relationship of two or more actions of the experimental animal in a specified time range, the probability of mutual conversion between different behavioral actions and the frequency distribution of different actions at different times. By analyzing from multiple perspectives, we can obtain richer behavioral information.

[0069] Step S3: Compare the action analysis results before and after the experiment to obtain the corresponding behavior analysis results of the experimental animals. Specifically, the experiment can be divided into the behavior comparison of the same experimental animal at different times, or it can be conducted in the form of an experimental group and a control group. Comparisons in different ways are convenient for drawing more intuitive conclusions.

[0070] Through the above steps S1 to S3, the animal behavior analysis method provided by the embodiment of the present invention can analyze multiple indicators of animal behavior from multiple angles of sequence analysis technology, and can obtain the temporal sequence relationship of two or more actions of experimental animals within a specified time range, the probability of mutual conversion between different behavioral actions, and the frequency distribution of different actions at different times. This multi-angle animal behavior analysis method can obtain richer behavioral information, making the analysis more comprehensive and accurate, thereby meeting more animal behavior analysis needs.

[0071] Specifically, in one embodiment, the above step S1, as Figure 2 As shown, the specific steps include:

[0072] Step S11: segmenting the video information into frames according to a preset frequency to obtain segmented motion images. Specifically, for example, segmenting the video into frames at a frequency of 24 frames per second to obtain sub-second motion images.

[0073] Step S12: dividing the action images into action sets to obtain multiple action image sets, each action image set including multiple continuous action images. Specifically, multiple continuous frames of images are divided into one action image set, for example, 16 continuous frames of images are taken as one action image set.

[0074] Step S13: fusing the action images in the action image set to obtain a fused action image set. Specifically, the image fusion process is to fuse the images every n frames by using an inter-frame difference method.

[0075] Step S14: extracting features from the fused action image set using a preset algorithm to obtain action feature information. Specifically, image feature extraction can use a suitable algorithm as needed, such as HOG image feature extraction algorithm, LBP feature extraction algorithm, Harris corner point detection, SIFT algorithm, etc.

[0076] Step S15: Substitute the action feature information into a preset classification model to obtain action information. Specifically, a batch of action images for training the model are obtained in advance through the above steps S11 to S13, and labels are defined for the action images to obtain a model training data set. A classification model is trained based on the model training data set through a deep learning method, and the action feature information is classified according to the classification model to obtain the action information.

[0077] Specifically, in one embodiment, the above step S2, such as Figure 3 As shown, the specific steps include:

[0078] Step S21: Detect frequent item sets on the action sequence to obtain the degree of association between different actions. Frequent item sets are sets of action association combinations consisting of different numbers of actions established based on the action sequence. Specifically, the sequential relationship of two or more actions appearing in sequence in time is analyzed by methods such as Apriori association analysis, GSP, Prefixspan, etc., and the detected action relationships are classified according to the different types of actions they contain. The final behavior analysis results are obtained by analyzing whether there are significant changes in different action relationships before and after the experiment.

[0079] Step S22: Perform statistics on the change of conversion probability of the action sequence based on the preset transfer matrix. Specifically, through, for example, lag sequence analysis, Markov transfer matrix, hidden Markov model, etc., study the probability of conversion between different actions, calculate the probability of another action occurring after one action occurs, and whether there is statistical significance.

[0080] Step S23: Count the order and frequency changes of different actions in the action sequence to obtain the temporal distribution of the action frequency. Specifically, by analyzing the temporal distribution of the action frequency, a more intuitive comparison result can be obtained.

[0081] Step S24: Analyze the continuous actions in the action sequence to obtain the disorder degree of the action. Extract the continuous actions in the action sequence and substitute them into the preset algorithm to obtain an entropy value. The larger the entropy value, the higher the disorder degree, and the more inconsistent the actions made by multiple experimental animals at the same time; conversely, the smaller the entropy value, the lower the disorder degree, and the higher the degree to which multiple experimental animals make the same action at the same time. The preset algorithm is:

[0082]

[0083] where h(p 1 ,p 2 ,...,p n ) is an entropy sequence of the same length as the action sequence, n is the length of the sequence, i is any i-th action in the sequence, and is the probability of the action occurring when the i-th action is reached. Here, log is the natural logarithm with e as the base. For example, if the actions in an action sequence are the same, the entropy value h(p 1 ,p 2 ,...,p n )=h(0,0,...,0), the actions are consistent and the entropy value is zero.

[0084] Specifically, by comparing the consistency of the movements of different experimental animals in the experimental group and the control group before and after the experiment, the degree of disorder of the movements is determined, and the impact of the experiment on the animals in the experimental group is obtained through the degree of disorder.

[0085] Specifically, in one embodiment, in the above step S21, frequent item sets are detected for the action sequence to obtain the degree of association between different actions, such as Figure 4 As shown, the specific steps include:

[0086] Step S211: Match the actions in the action sequence based on each action association combination in the frequent item set to obtain the matching results of each action association combination. Specifically, the action association combination is divided into multiple levels, each level corresponds to a different action association combination, for example, the second-order action association combination is associated with two actions, and the third-order action association combination is associated with three actions; through matching, the number of times each action association combination appears in the action sequence can be obtained.

[0087] Step S212: Count the matching results to obtain the frequency of occurrence of different action association combinations, and determine the degree of association between different actions based on the frequency of occurrence of different action association combinations. Specifically, the action association combination may show differences in action type and order before and after the experiment. These differences can be easily detected by the program according to certain rules, which is helpful for analyzing the emotions and actions of experimental animals before and after the experiment; taking the depression modeling of mice as an example, after the depression modeling, the mice have fewer high-order action association combinations in the action sequence, and the action association combinations of the same order also show differences in action type.

[0088] Specifically, in one embodiment, the above step S2, such as Figure 5 As shown, the following steps are also included:

[0089] Step S201: Associatively classify actions by establishing multiple hierarchies, each of which contains a number of different behavior modules, and the behavior module contains a number of actions corresponding to the order of the corresponding hierarchy. Specifically, for example, each first-order behavior module includes two actions (A, B), and each second-order behavior module includes three actions ((A, B), C).

[0090] Step S202: Based on the action sequence, the actions are classified and combined according to the hierarchy to form different behavior modules. Specifically, for example: first find two actions A and B that are closely connected as a first-order action module (A, B), and then traverse the remaining action modules. If there is a closely connected action module C, a second-order action module ((A, B), C) is formed. Next, as above, higher-order behavior modules are searched in turn. Then the second-order, third-order, fourth-order and other order behavior modules are counted.

[0091] Step S203: forming frequent item sets by integrating the behavior modules. Figure 6-9As shown in Figure 1, frequent itemsets are two-level or even multi-level structures composed of several actions. The detected frequent itemsets are classified according to the different types of actions they contain.

[0092] Specifically, in one embodiment, Figure 10-13 As shown, the frequent item sets mined by the algorithm can be further analyzed from the following perspectives:

[0093] Compare the number of types of frequent itemsets before and after the experiment, and whether there are changes in the number of frequent itemsets;

[0094] Count the lengths of the frequent itemsets before and after the experiment that need to be compared and the number of occurrences of the frequent itemsets corresponding to each length;

[0095] Count the number of layers of frequent itemsets in the two sets of data to be compared and the number of occurrences of frequent itemsets corresponding to each layer. The level represents the complexity of a frequent itemset.

[0096] Count the number of occurrences of frequent itemsets containing different actions; a frequent itemset is composed of several identical or different actions. In two or more groups of data, one group of frequent itemsets contains a frequent itemset consisting of a certain type of action, while the other group of frequent itemsets does not contain this action, or both groups of data contain frequent itemsets with the same action, but the frequency of occurrence of this action is significantly different.

[0097] Specifically, in one embodiment, the above step S22, such as Fig.14 As shown, the specific steps include:

[0098] Step S221: extract the actions appearing in the action sequence, and group each action into two groups as a base group for action conversion. Specifically, since the purpose is to analyze the probability of conversion between two actions, the actions appearing in the action sequence are extracted in advance, and then the two actions are grouped together for classification, which is convenient for subsequent comparison.

[0099] Step S222: Compare each basis set with the action sequence to obtain the probability of occurrence of each basis set. Fig.15 As shown in the figure, taking the experimental mouse depression modeling as an example, the figure shows the action transfer probability diagram of the depression modeling group and the normal control group mice at different periods. In order to make the change of transfer probability more obvious, the probability less than 0.3 is set to zero. It can be clearly seen that in the second and third weeks of the depression modeling group, the probability of converting other actions to circling actions is relatively high, while this phenomenon did not occur in the normal control group.

[0100] Step S223: Statistic the probability of occurrence of each basis group to obtain the action transition probability between different actions. Specifically, if the probability of transition from behavior A to behavior B is relatively high, it does not mean that the frequency of occurrence of behavior A to behavior B is high. The study of the frequency of continuous occurrence of two or more behaviors is achieved by detecting complex item sets; the high probability of transition from behavior A to behavior B means that if behavior A occurs, the probability of behavior B occurring next is relatively high. Fig.16 As shown, the action conversion probability of different actions can be reflected in the form of an action conversion probability diagram. Different actions are represented by circles, and the diameter of the circle is proportional to the frequency of each action. The curve between each circle represents the probability of conversion between different actions. The darker the color and the thicker the width of the curve, the closer the connection between the two actions.

[0101] Specifically, in one embodiment, the above step S23, as Fig.17 As shown, the specific steps include:

[0102] S231: Extract actions appearing in the action sequence. Specifically, an action sequence with a length of n corresponds to n time points, and n time points correspond to K actions, and the entire action sequence is traversed.

[0103] S232: Count the order and frequency changes of different actions, and obtain the probability of different actions. Specifically, by calculating the probability of K actions corresponding to n time points, Kn probability values ​​can be obtained.

[0104] S233: Arrange the probabilities in time series to obtain a time series distribution of the action frequencies.

[0105] Specifically, the K probability values ​​corresponding to each time point are arranged in descending order, and the probability with the largest value is taken as the most likely probability of the action at that point; the n probability values ​​selected are arranged in time series, a histogram is drawn, and different colors are used to represent different actions; in order to facilitate the comparison of the same actions, similar actions can be separated and the differences in the timing of occurrence of similar actions can be compared.

[0106] Specifically, in one embodiment, the above-mentioned animal behavior analysis method also establishes a standard behavior library, which regards eight consecutive frames of images as a continuous action, that is, each action lasts for 0.44 seconds, reaching the sub-second level, so as to obtain richer and more detailed behavioral information. It can be compared to human language, where different action combinations form information with different meanings (secondary or even multi-level behaviors). Here, twenty sub-second actions of mice are listed as an example:

[0107] 1. Combing Action

[0108] Grooming usually starts with rubbing the front paws against the face, then moves along the body to the end of the tail. Grooming is done with the front paws, while licking is done with the mouth, and the two actions are often performed alternately and simultaneously.

[0109] 1.1 Head grooming: Use both front paws to touch the head in circles at the same time. The large range can touch the ears, while the small range is only around the nose and eyes. In short, there is a tendency for both front paws to move towards the tip of the nose.

[0110] 1.2 Licking the front paws: This action is often accompanied by the head-grooming action. Both front paws are placed in front of the mouth and licking. Both paws may move or may not, but the displacement is very small.

[0111] 1.3 Licking the hind paw

[0112] 1.4 Licking the belly

[0113] 1.5 Licking the ribs

[0114] 1.6 Licking the tail

[0115] 1.7 Licking thighs

[0116] 2. Still Action

[0117] 2.1 Prone: Front and rear paws on the ground, motionless.

[0118] 2.2 Squat stand: being still in a standing position.

[0119] 3. Explore the action

[0120] 3.1 Smell the ground: Smell the ground while walking forward.

[0121] 3.2 Walking along the wall: Keep your body close to the acrylic tube wall and move forward slowly.

[0122] 3.3 Non-wall walking: The body does not stick to the acrylic tube wall, and moves forward in a straight line at a slow speed.

[0123] 3.4 Running along the wall: Keep your body close to the acrylic tube wall and move forward at a faster speed.

[0124] 3.5 Non-wall running: The body does not stick to the acrylic tube wall, and moves forward in a straight line at a fast speed.

[0125] 3.6 Circle: Rotate in place

[0126] 3.7 Turn left: Turn left while moving forward

[0127] 3.8 Turn right: Turn right while moving forward

[0128] 3.9 Back

[0129] 3.10 Looking around: The hind paws are on the ground, and the front paws are suspended in the air. The head often looks back and forth quickly, but sometimes there may be no violent movement of the head.

[0130] 3.11 Climbing: The front paws lie on the acrylic tube wall, the hind paws touch the ground, and the body is in a stretched state.

[0131] In this embodiment, an animal behavior analysis device is also provided, which is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions that have been made will not be repeated. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0132] This embodiment provides an animal behavior analysis device, such as Fig.18 As shown, including:

[0133] The acquisition module 101 is used to obtain the motion information of the experimental animal before and after the experiment, and arrange the motion information in chronological order to form a motion sequence. For details, please refer to the relevant description of step S1 in the above method embodiment, which will not be repeated here.

[0134] The processing module 102 is used to perform motion analysis on the motion sequence from the perspectives of motion frequency, motion conversion probability and motion timing distribution to obtain motion analysis results. For details, please refer to the relevant description of step S2 in the above method embodiment, which will not be repeated here.

[0135] The analysis module 103 compares the action analysis results before and after the experiment to obtain the corresponding behavior analysis results of the experimental animals. For details, please refer to the relevant description of step S3 in the above method embodiment, which will not be repeated here.

[0136] Specifically, in one embodiment, if Fig.18 As shown, the processing module 102 includes:

[0137] The frequency detection module 1021 is used to detect frequent item sets for the action sequence to obtain the degree of association between different actions. The frequent item set is a set of action association combinations consisting of different numbers of actions established based on the action sequence. For details, please refer to the relevant description of step S21 in the above method embodiment, which will not be repeated here.

[0138] The action conversion module 1022 is used to perform statistics on the change of conversion probability of the action sequence based on a preset transfer matrix. For details, please refer to the relevant description of step S22 in the above method embodiment, which will not be repeated here.

[0139] The disorder statistics module 1023 is used to count the order and frequency changes of different actions in the action sequence to obtain the disorder degree of the actions. For details, please refer to the relevant description of step S23 in the above method embodiment, which will not be repeated here.

[0140] The animal behavior analysis device in this embodiment is presented in the form of a functional unit, where the unit refers to an ASIC circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above functions.

[0141] The further functional description of each of the above modules is the same as that of the above corresponding embodiments and will not be repeated here.

[0142] According to an embodiment of the present invention, there is also provided an electronic device, such as Fig.19 As shown, the electronic device may include a processor 901 and a memory 902, wherein the processor 901 and the memory 902 may be connected via a bus or other means. Fig.19 The example of connecting through bus is taken in the following.

[0143] The processor 901 may be a central processing unit (CPU). The processor 901 may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or a combination of the above chips.

[0144] The memory 902 is a non-transitory computer-readable storage medium that can be used to store non-transitory software programs, non-transitory computer executable programs and modules, such as program instructions / modules corresponding to the methods in the method embodiments of the present invention. The processor 901 executes various functional applications and data processing of the processor by running the non-transitory software programs, instructions and modules stored in the memory 902, that is, implementing the methods in the above method embodiments.

[0145] The memory 902 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required by at least one function; the data storage area may store data created by the processor 901, etc. In addition, the memory 902 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory 902 may optionally include a memory remotely arranged relative to the processor 901, and these remote memories may be connected to the processor 901 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0146] One or more modules are stored in the memory 902 , and when executed by the processor 901 , the method in the above method embodiment is executed.

[0147] The specific details of the above electronic device can be understood by referring to the corresponding descriptions and effects in the above method embodiments, and will not be repeated here.

[0148] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, the storage medium can be a disk, an optical disk, a read-only memory (ROM), a random access memory (RAM), a flash memory, a hard disk drive (HDD) or a solid-state drive (SSD), etc.; the storage medium can also include a combination of the above-mentioned types of memory.

[0149] Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention, and such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A method for analyzing animal behavior, characterized in that: include: Acquiring video information of the experimental animal before and after the experiment, processing the video information to obtain action information, and arranging the action information in chronological order to form an action sequence; The preprocessing of the video information to obtain the action information includes: performing frame segmentation on the video information to obtain segmented action images; dividing the action images into action sets to obtain multiple action image sets, each action image set including multiple continuous action images; performing image fusion on the action images in the action image sets to obtain a fused action image set; performing feature extraction on the fused action image set by a preset algorithm to obtain action feature information; substituting the action feature information into a preset classification model to obtain action information; The action sequence is analyzed from the perspectives of action frequency, action conversion probability, action time series distribution and disorder degree to obtain the action analysis result; including: detecting frequent item sets on the action sequence to obtain the degree of association between different actions, the frequent item sets are a set of action association combinations composed of different numbers of actions based on the degree of association between actions; statistically analyzing the change of conversion probability of the action sequence based on a preset transfer matrix algorithm; statistically analyzing the order and frequency changes of different actions in the action sequence to obtain the time series distribution of action frequency; and obtaining the degree of disorder of the action by analyzing continuous actions in the action sequence; The step of analyzing the continuous actions in the action sequence to obtain the disorder degree of the actions specifically includes: extracting the continuous actions in the action sequence and substituting them into a preset algorithm to obtain an entropy value; the entropy value is used to characterize the degree of disorder, the larger the entropy value, the higher the disorder degree, and the smaller the entropy value, the lower the disorder degree; The action analysis results before and after the experiment are compared to obtain the corresponding behavior analysis results of the experimental animals.

2. The animal behavior analysis method according to claim 1, characterized in that: The detecting of frequent itemsets on the action sequence to obtain the degree of association between different actions includes: Matching the actions in the action sequence based on each action association combination in the frequent item set to obtain a matching result of each action association combination; The matching results are counted to obtain the frequencies of occurrence of different action association combinations, and the association degrees between different actions are determined based on the frequencies of occurrence of different action association combinations.

3. The animal behavior analysis method according to claim 1, characterized in that: Also includes: The actions are associated and classified by establishing multiple hierarchies, where different hierarchies have different orders, and each hierarchy includes a number of different behavior modules, and the behavior modules contain a number of actions corresponding to the order of the corresponding hierarchy; Based on the action sequence, the actions are classified and combined according to the hierarchy to form different behavior modules; The frequent item set is formed by integrating the behavior modules.

4. The animal behavior analysis method according to claim 1, characterized in that: The performing statistics on changes in conversion probabilities of the action sequence based on a preset transfer matrix algorithm includes: Extracting actions appearing in the action sequence, and grouping the actions in pairs as a basis group for mutual conversion of actions; Comparing each of the basis groups with the action sequence respectively to obtain the probability of occurrence of each of the basis groups; The probability of occurrence of each basis set is statistically analyzed to obtain the action transition probability between different actions.

5. The animal behavior analysis method according to claim 1, characterized in that: The statistically analyzing the order and frequency changes of different actions in the action sequence to obtain the temporal distribution of action frequencies includes: Extracting actions appearing in the action sequence; Count the order and frequency of different actions, and get the probability of different actions; The probabilities are arranged in time series to obtain a time series distribution of the action frequencies.

6. An animal behavior analysis device, characterized in that: include: Acquisition module: acquiring video information of the experimental animal before and after the experiment, processing the video information to obtain action information, and arranging the action information in chronological order to form an action sequence; the preprocessing of the video information to obtain the action information includes: performing frame segmentation on the video information to obtain segmented action images; dividing the action images into action sets to obtain multiple action image sets, each action image set including multiple continuous action images; performing image fusion on the action images in the action image sets to obtain a fused action image set; performing feature extraction on the fused action image set by a preset algorithm to obtain action feature information; substituting the action feature information into a preset classification model to obtain action information; Processing module: Processing the action sequence from the perspectives of action frequency, action conversion probability and action time series distribution to obtain processing results; including: detecting frequent item sets on the action sequence to obtain the degree of association between different actions, the frequent item sets are sets of action association combinations composed of different numbers of actions based on the degree of association between actions; performing statistics on the change of conversion probability of the action sequence based on a preset transfer matrix algorithm; performing statistics on the order and frequency changes of different actions in the action sequence to obtain the time series distribution of action frequency; and obtaining the degree of disorder of actions by analyzing continuous actions in the action sequence; The step of analyzing the continuous actions in the action sequence to obtain the disorder degree of the actions specifically includes: extracting the continuous actions in the action sequence and substituting them into a preset algorithm to obtain an entropy value; the entropy value is used to characterize the degree of disorder, the larger the entropy value, the higher the disorder degree, and the smaller the entropy value, the lower the disorder degree; Analysis module: Analyze the processing results before and after the experiment to obtain the final result.

7. An electronic device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the animal behavior analysis method according to any one of claims 1 to 5 by executing the computer instructions.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable the computer to execute the animal behavior analysis method according to any one of claims 1 to 5.

Images