Video saving strategy determination method and device, equipment and storage medium

Abstract

The invention discloses a video saving strategy determination method and device, equipment and a storage medium. The method comprises the following steps: acquiring scene related parameters, and dynamically adjusting video storage parameters according to the scene related parameters; when the system detects that a transaction is generated, transaction related information of the current transaction is acquired, and a transaction condition is determined according to the scene related parameters, the transaction related information and a pre-constructed machine learning model; and when the transaction condition is an abnormal transaction, marking the collected video corresponding to the current transaction as an abnormal video, and locking and storing the abnormal video. By dynamically adjusting the video storage parameters, storage resources are allocated according to needs, and the effective video storage duration is prolonged. The transaction condition is judged in combination with multi-dimensional information and a model, the accuracy of abnormal transaction recognition is improved, and risk scenes which are difficult to capture by manual rules are covered. By locking and storing the abnormal video, the risk video can be ensured not to be covered, and the reliability of risk event tracing is improved.

Description

Technical Field

[0001] The present invention relates to the field of audio - video technology, and in particular to a method, device, equipment and storage medium for determining a video saving strategy. Background Art

[0002] Banks usually monitor teller workstations from multiple angles for recording, so as to provide evidence in case of risk events, protecting consumers' rights and interests, etc. With the refinement of regulatory requirements and the expansion of business scale, banks have put forward higher requirements for the storage efficiency, risk identification ability and retrieval convenience of monitoring videos.

[0003] Currently, the banking industry generally adopts a 24 - hour non - discriminatory high - quality video acquisition scheme. Whether the teller counter is in the business handling state or not, videos are saved at a fixed bit rate and frame rate, which leads to insufficient utilization of disk space. When identifying risks in the prior art, it mainly relies on manually - set rules based on experience, such as marking large - amount transaction codes, or screening suspected risk transactions through simple statistical models. Manual rules are difficult to cover complex risk scenarios and cannot dynamically adapt to business changes, while simple statistical models have insufficient ability to analyze the correlation of multi - dimensional features and are prone to missing potential risk transactions. Summary of the Invention

[0004] The present invention provides a method, device, equipment and storage medium for determining a video saving strategy. By dynamically adjusting video saving parameters, the purpose of saving storage resources is achieved, and machine learning is used to determine whether there are abnormal risks in relevant transactions, so as to lock and save suspected risk videos.

[0005] According to one aspect of the present invention, a method for determining a video saving strategy is provided. The method includes:

[0006] Obtain scene - related parameters, and dynamically adjust video saving parameters according to the scene - related parameters, where the scene - related parameters include system usage conditions and personnel detection conditions;

[0007] When the system detects a transaction, obtain transaction - related information of the current transaction, and determine the transaction situation according to the scene - related parameters, transaction - related information and a pre - constructed machine learning model;

[0008] When the transaction situation is an abnormal transaction, mark the captured video corresponding to the current transaction as an abnormal video, and lock and save the abnormal video.

[0009] Optionally, the video saving parameters are dynamically adjusted according to scene-related parameters, including: determining whether the system usage is the system in use. If so, determining the video saving parameters as a preset first bit rate and first frame rate; otherwise, when the personnel detection situation is that no person is detected, determining the video saving parameters as a preset second bit rate and second frame rate; when the personnel detection situation is that a person is detected, determining the video saving parameters as a preset third bit rate and first frame rate, where the first bit rate is less than the third bit rate and greater than the second bit rate, and the second frame rate is less than the first frame rate.

[0010] Optionally, the transaction situation is determined according to scene-related parameters, transaction-related information, and a pre-constructed machine learning model, including: when the system usage is that the system is not in use and the personnel detection situation is that a person is detected, determining the transaction situation as an abnormal transaction; obtaining a preset important code list, and determining the transaction situation according to the preset important code list and transaction-related information; predicting the transaction-related information through the pre-constructed machine learning model to determine the transaction situation.

[0011] Optionally, the transaction situation is determined according to a preset important code list and transaction-related information, including: extracting a transaction code and a transaction amount from the transaction-related information; when the transaction code is in the important code list, determining the transaction situation as an abnormal transaction; when the transaction amount is greater than a preset amount threshold, determining the transaction situation as an abnormal transaction.

[0012] Optionally, the transaction-related information is predicted through a pre-constructed machine learning model to determine the transaction situation, including: extracting transaction features from the transaction-related information; inputting the transaction features into the pre-constructed machine learning model and obtaining the output risk value; when the risk value is greater than or equal to a preset risk threshold, determining the transaction situation as an abnormal transaction.

[0013] Optionally, the method further includes: extracting a transaction code from the transaction-related information, generating a chapter break point according to the transaction code; adding the chapter break point to the collected video corresponding to the current transaction; when the transaction is completed, obtaining the transaction key information of the current transaction and inserting the transaction key information into the chapter break point to generate a retrieval information point, where the transaction key information includes a transaction serial number, a customer number, a teller number, a transaction code, a handling start time, and a handling end time.

[0014] According to another aspect of the present invention, a video saving policy determination device is provided, and the device includes:

[0015] A video saving parameter determination module, configured to obtain scene-related parameters and dynamically adjust video saving parameters according to the scene-related parameters, where the scene-related parameters include system usage and personnel detection situation;

[0016] A transaction situation determination module, configured to obtain transaction-related information of the current transaction when the system detects a transaction, and determine the transaction situation according to scenario-related parameters, transaction-related information, and a pre-constructed machine learning model;

[0017] An abnormal video locking module, configured to mark the captured video corresponding to the current transaction as an abnormal video and lock and save the abnormal video when the transaction situation is an abnormal transaction.

[0018] According to another aspect of the present invention, there is provided an electronic device, the electronic device includes:

[0019] At least one processor;

[0020] And a memory communicatively connected to the at least one processor;

[0021] Wherein, the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute a method for determining a video saving policy according to any embodiment of the present invention.

[0022] According to another aspect of the present invention, there is provided a computer-readable storage medium, the computer-readable storage medium stores computer instructions, and the computer instructions are used to implement a method for determining a video saving policy according to any embodiment of the present invention when executed by a processor.

[0023] According to another aspect of the present invention, there is provided a computer program product, the computer program product includes a computer program, and the computer program implements a method for determining a video saving policy according to any embodiment of the present invention when executed by a processor.

[0024] The technical solution of the embodiments of the present invention can dynamically adjust video saving parameters, so as to achieve on-demand allocation of video storage resources, reduce storage waste during unattended periods, and extend the effective video storage duration. By combining multi-dimensional information and a model to determine the transaction situation, the comprehensiveness and accuracy of abnormal transaction recognition can be improved, covering risk scenarios that are difficult to capture by manual rules. By locking and saving abnormal videos, it can ensure that key risk videos are not overwritten, and improve the reliability and efficiency of risk event traceability.

[0025] It should be understood that the content described in this part is not intended to identify the key or important features of the embodiments of the present invention, nor is it used to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. Description of the Drawings

[0026] To more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0027] Figure 1 is a flowchart of a method for determining a video storage strategy provided in Embodiment 1 of the present invention;

[0028] Figure 2 is a flowchart of another method for determining a video storage strategy provided in Embodiment 2 of the present invention;

[0029] Figure 3 is a schematic structural diagram of a device for determining a video storage strategy provided in Embodiment 3 of the present invention;

[0030] Figure 4 is a schematic structural diagram of an electronic device for implementing the method for determining a video storage strategy in the embodiments of the present invention. Specific Embodiments

[0031] In order to enable those skilled in the art to better understand the solutions of the present invention, the following clearly and completely describes the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0032] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that such used data can be interchanged under appropriate circumstances so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device comprising a series of steps or units does not necessarily have to be limited to those clearly listed steps or units, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0033] Embodiment 1

[0034] Figure 1The following is a flowchart of a method for determining a video saving policy provided by Embodiment 1 of the present invention. This embodiment is applicable to the transaction scenario of a bank teller counter. This method can be executed by a video saving policy determining device, which can be implemented in the form of hardware and / or software, and can be configured in a computer controller. As Figure 1 shown, the method includes:

[0035] S110. Obtain scene-related parameters, and dynamically adjust video saving parameters according to the scene-related parameters, where the scene-related parameters include the system usage situation and the personnel detection situation.

[0036] Among them, the scene-related parameters refer to the key indicators describing the current state of the bank counter, including the system usage situation and the personnel detection situation. The system usage situation refers to whether the teller is operating the bank counter terminal system. For example, when the teller uses the system for transactions such as transferring money or opening an account, the system usage situation is in use at this time, while when the teller leaves the work station or the system is idle, the system usage situation is not in use. The personnel detection situation refers to judging whether there is anyone in the counter area through image detection technology. For example, when there is no business being handled, the personnel detection result is that no one is detected, while when a customer is handling business in front of the counter, the personnel detection result is that someone is detected. The video saving parameters refer to the parameters controlling the video storage quality, including the bit rate and the number of frames.

[0037] Optionally, dynamically adjusting the video saving parameters according to the scene-related parameters includes: judging whether the system usage situation is using the system. If so, determining the video saving parameters as a preset first bit rate and a first number of frames; otherwise, when the personnel detection situation is that no one is detected, determining the video saving parameters as a preset second bit rate and a second number of frames; when the personnel detection situation is that someone is detected, determining the video saving parameters as a preset third bit rate and the first number of frames, where the first bit rate is less than the third bit rate and greater than the second bit rate, and the second number of frames is less than the first number of frames.

[0038] Among them, the bit rate refers to the data traffic used by the video file per unit time, and the number of frames refers to the number of pictures displayed per second of the video. The first bit rate and the first number of frames refer to the preset normal bit rate and normal number of frames, which are the clear video saving standards applicable to the normal business scenario. The second bit rate and the second number of frames refer to the preset low bit rate and low number of frames, which are the compression saving standards applicable to the situation of no one and the system being idle, and can reduce the storage space occupancy. The third bit rate refers to the preset high bit rate, which is the high-definition saving standard applicable to the situation of someone but the system not being used, such as when a customer is waiting in front of the counter but the teller is not operating.

[0039] Specifically, when the system is in use, important business operations may be in progress, and high-quality video recording is required for subsequent viewing and analysis. Therefore, when it is determined that the system is in use, the controller will dynamically adjust the video saving parameters to the preset first bitrate and first frame rate. When the system is not in use and no person is detected, the controller will dynamically adjust the video saving parameters to the preset second bitrate and second frame rate. At this time, the importance of the video content is relatively low, and reducing the bitrate and frame rate can reduce the storage space occupancy. For example, after the bank closes, there is no one in the counter area and the system is also idle. When the system is not in use but there are people in the area, there may be potential risks, and high-quality video recording is also required at this time. Therefore, when the system is not in use but a person is detected, the controller will dynamically adjust the video saving parameters to the preset third bitrate and first frame rate to ensure that the behavior of the customer and the surrounding environment can be clearly recorded and prevent accidents.

[0040] S120. When the system detects a transaction, obtain the transaction-related information of the current transaction, and determine the transaction situation according to the scenario-related parameters, transaction-related information, and the pre-constructed machine learning model.

[0041] Among them, the transaction-related information refers to the data directly associated with the specific transaction operation and is used to analyze the transaction risk. The pre-constructed machine learning model refers to a risk prediction model constructed using the Gradient Boosting Decision Tree (GBDT) algorithm and is used to identify abnormal transactions. The transaction situation includes normal transactions and abnormal transactions.

[0042] S130. When the transaction situation is an abnormal transaction, mark the captured video corresponding to the current transaction as an abnormal video and save the abnormal video with a lock.

[0043] Among them, an abnormal transaction refers to a transaction determined to be a high-risk transaction by the machine learning model, such as a large cash transaction, a transaction with frequent operation errors, and a transaction involving sensitive operations, etc. Saving with a lock means adding a lock mark to the video corresponding to the abnormal transaction. The video with the lock mark has a retention priority in the storage server and is avoided from being overwritten by subsequent normal videos. For example, when the monitoring server can only store videos for 6 months, the locked video can be retained for a longer time.

[0044] Optionally, the method further includes: extracting a transaction code from the transaction-related information, generating a chapter break point according to the transaction code; adding the chapter break point to the captured video corresponding to the current transaction; when the transaction is completed, obtaining the transaction key information of the current transaction, and inserting the transaction key information into the chapter break point to generate a retrieval information point, where the transaction key information includes the transaction serial number, customer number, teller number, transaction code, handling start time, and handling end time.

[0045] Among them, the transaction code is an identifier for standardizing the encoding of different types of transactions, used to distinguish various transaction services, and each transaction code corresponds to a specific transaction type. The chapter breakpoint refers to a specific marker point set in the video. The chapter breakpoint can segment the video according to transactions, facilitating subsequent rapid positioning and retrieval of video segments of specific transactions. Taking the bank surveillance video as an example, for each transaction that occurs, a chapter breakpoint is generated at the corresponding video position, dividing the video into small segments corresponding to the transactions. The captured video refers to the video obtained by real-time shooting of a specified area, such as a bank counter, through devices such as cameras, used to record the actual scenes during the transaction process, including the operations and communications of personnel, etc. The transaction key information is a set of information with important identification and recording value in the transaction, including the transaction serial number, customer number, teller number, transaction code, start time of handling, and end time of handling. The transaction serial number is a unique number automatically generated by the system, used to identify each transaction. The customer number is a unique number assigned by the bank system to each customer, facilitating the management and identification of customers. The teller number is the unique identification number of each bank teller, used to clarify the operator of the transaction. The retrieval information point refers to a marker point added to the video that contains the transaction key information. By means of the retrieval information point, when it is necessary to search for the target transaction video subsequently, the corresponding video segment can be quickly located according to the transaction key information, improving the efficiency of video retrieval.

[0046] In a specific implementation, when a teller handles an account opening business for a customer, the controller can extract the transaction code of this account opening business as "0000" from the transaction-related information, and then generate a chapter breakpoint at the corresponding position in the currently captured video, marking this video segment as "the account opening business corresponding to the transaction code 0000". After the transaction is completed, obtain the transaction key information of the current transaction, including the transaction serial number "202504280002", customer number "C00002", teller number "T002", transaction code "9999", start time of handling "2026 / 04 / 28 10:00:00", and end time of handling "2025 / 04 / 28 10:30:00", and then insert the transaction key information into the chapter breakpoint to form a retrieval information point. When it is necessary to search for the video of this account opening business, just input any transaction key information such as the transaction serial number, customer number, teller number, etc., and the video segment can be quickly located, improving the convenience of video retrieval.

[0047] The technical solution of the embodiment of the present invention can achieve the on-demand allocation of video storage resources by dynamically adjusting video saving parameters, reduce the storage waste during unattended periods, and extend the effective video storage duration. By combining multi-dimensional information and model determination of transaction situations, the comprehensiveness and accuracy of abnormal transaction recognition can be improved, covering risk scenarios that are difficult to capture by manual rules. By locking and saving abnormal videos, it can ensure that key risk videos are not overwritten, improving the reliability and efficiency of risk event traceability.

[0048] Embodiment 2

[0049] Figure 2 It is a flowchart of a method for determining a video saving strategy provided by Embodiment 2 of the present invention. In this embodiment, on the basis of Embodiment 1 above, the specific process of determining transaction situations according to scenario-related parameters, transaction-related information, and a pre-constructed machine learning model is added. Among them, the specific contents of steps S210 and S260 are substantially the same as those of steps S110 and S130 in Embodiment 1, so they will not be elaborated in this embodiment. As Figure 2 shown, the method includes:

[0050] S210. Obtain scenario-related parameters, and dynamically adjust video saving parameters according to the scenario-related parameters, where the scenario-related parameters include system usage and personnel detection.

[0051] Optionally, dynamically adjusting video saving parameters according to scenario-related parameters includes: determining whether the system usage is that the system is in use. If so, determining the video saving parameters as a preset first bitrate and first frame rate; otherwise, when the personnel detection is that no person is detected, determining the video saving parameters as a preset second bitrate and second frame rate; when the personnel detection is that a person is detected, determining the video saving parameters as a preset third bitrate and first frame rate, where the first bitrate is less than the third bitrate and greater than the second bitrate, and the second frame rate is less than the first frame rate.

[0052] S220. When the system detects that a transaction occurs, obtain the transaction-related information of the current transaction.

[0053] S230. When the system usage is that the system is not in use and the personnel detection is that a person is detected, determine the transaction situation as an abnormal transaction.

[0054] Among them, an abnormal transaction refers to a transaction that does not conform to the normal transaction mode or has potential risks. Abnormal transactions may be manifested as situations such as the transaction amount exceeding the normal range, the transaction behavior not conforming to the customer's historical habits, and involving sensitive operations. For example, if a customer's usual transaction amount is within a few thousand yuan and suddenly makes a large transfer of several million yuan, it will be determined as an abnormal transaction.

[0055] Specifically, when the system usage is that the system is not in use and the personnel detection situation is that a person is detected, there are some potential risks. For example, after the bank closes, the system has been shut down, but the surveillance camera detects human activities in the bank counter area. Therefore, at this time, the controller will determine that the transaction situation is an abnormal transaction.

[0056] S240. Obtain a preset list of important codes, and determine the transaction situation based on the preset list of important codes and transaction-related information.

[0057] Among them, the preset list of important codes refers to a list that is pre-configured and contains important transaction codes. Transactions corresponding to important transaction codes usually have relatively high risks or importance and require key attention. For example, in a bank system, codes for transactions involving large-scale fund transfers, account privilege changes, etc. will be included in the preset list of important codes.

[0058] Specifically, the controller will obtain the preset list of important codes and compare the transaction code in the transaction-related information with the preset list of important codes. If the transaction code is in the preset list of important codes, then this transaction will be determined as an abnormal transaction.

[0059] Optionally, determining the transaction situation based on the preset list of important codes and transaction-related information includes: extracting the transaction code and transaction amount from the transaction-related information; when the transaction code is in the list of important codes, determining that the transaction situation is an abnormal transaction; when the transaction amount is greater than the preset amount threshold, determining that the transaction situation is an abnormal transaction.

[0060] Among them, the preset amount threshold refers to a pre-set amount limit. When the transaction amount exceeds the limit, this transaction may be considered to have certain risks or require special handling. For example, if the bank sets the preset amount threshold at 500,000 yuan, then when the amount of a transfer transaction exceeds 500,000 yuan, it will trigger the corresponding risk warning mechanism.

[0061] In a specific implementation, when a customer handles a transfer business, the system will record the relevant information of this transaction and extract the transaction code representing the transfer business and the transfer amount from it. Then, the controller will compare the extracted transaction code with the preset list of important codes. If the transaction code is in the list of important codes, then it is determined that the transaction situation is an abnormal transaction. The controller will also compare the extracted transaction amount with the preset amount threshold. When the transaction amount is greater than the preset amount threshold, it is determined that the transaction situation is an abnormal transaction. By determining the transaction situation based on the preset list of important codes and the transaction amount threshold, abnormal transactions that may have risks can be effectively identified, thereby ensuring the security and compliance of transactions.

[0062] S250. Use a pre-constructed machine learning model to predict the transaction-related information to determine the transaction situation.

[0063] Among them, the pre - constructed machine - learning model refers to a model trained using historical data before conducting transaction situation analysis. In this embodiment, it is a model constructed using the GBDT algorithm. The machine - learning model takes transaction features extracted from transaction - related information as input, and after learning and training, outputs a predicted risk value indicating whether the transaction is an abnormal transaction. For example, a large amount of transaction data over a past period of time, including both normal and abnormal transaction data, is used to train the GBDT model so that it can determine whether a new transaction is abnormal based on the input transaction information.

[0064] Optionally, the pre - constructed machine - learning model is used to predict transaction - related information to determine the transaction situation, including: extracting transaction features from the transaction - related information; inputting the transaction features into the pre - constructed machine - learning model and obtaining the output risk value; when the risk value is greater than or equal to the preset risk threshold, determining that the transaction situation is an abnormal transaction.

[0065] Among them, the risk value refers to the quantitative evaluation result of the machine - learning model for transaction risk, usually a probability value in the range of 0 - 1, reflecting the likelihood that the transaction is abnormal. The preset risk threshold refers to the preset critical value for risk judgment, used to screen transactions that need to be monitored closely. For example, the preset risk threshold can be 0.6, and when the risk value ≥ 0.6, an abnormal mark is triggered.

[0066] Furthermore, the construction process of the machine - learning model includes data collection, data pre - processing, label definition, model training, and model optimization phases. Among them, data collection refers to collecting historical transaction data from multiple systems of the bank, including the counter system, risk control system, video monitoring system, etc. The data collected specifically includes: teller attributes: such as working years, consecutive days on duty, historical risk occurrence frequency, days since the last risk event, training participation frequency, etc.; transaction attributes: transaction code, business serial number, transaction amount, whether cash receipt and payment are involved, whether review and authorization are required, number of operation steps, transaction familiarity, etc.; environmental attributes: transaction time period, number of people queuing in the hall, ratio of the number of people queuing to the number of tellers, etc.

[0067] Among them, data pre - processing includes data cleaning, normalization, and one - hot encoding. Data cleaning refers to removing noise, duplicate values, and missing values from the data. Normalization refers to normalizing numerical features, such as teller working years, ratio of the number of people queuing to the number of tellers, etc., and scaling them to the 0 - 1 interval to eliminate the influence of different feature dimensions. One - hot encoding refers to, for categorical features, such as transaction time period, whether password input is involved, etc., converting them into binary vectors using one - hot encoding for easy model processing.

[0068] Among them, label definition refers to transactions marked as "risk" by the bank's post - supervision system. Regardless of the degree of risk, they are marked as positive samples (Y = 1). Specifically, it can include: actual risks have occurred, being notified due to potential risks discovered, the parties being punished such as having their performance deducted, and the parties being forced to participate in risk training. At the same time, normal transactions can be marked as negative samples (Y = 0). By means of label generalization, the scope of risk samples is expanded, which helps the model learn more risk features.

[0069] Among them, model training refers to dividing the pre - processed data into a training set and a test set, and using the training set to train the GBDT model. The model trains multiple decision trees through iterative training. Each tree is trained based on the residuals of the previous model, gradually reducing the overall error. Finally, the results of all trees are accumulated to obtain the final prediction value. Model training also includes an evaluation process, that is, using the test set to evaluate the trained model to measure the performance of the model. Model optimization refers to optimizing the model according to the evaluation results. For example, regularly updating the training data, such as importing new risk cases every month, to adapt to changes in the business model. In addition, the performance of the model can be further improved by adjusting the parameters of the model, such as the number of trees and the depth of the trees.

[0070] S260. When the transaction situation is an abnormal transaction, mark the acquisition video corresponding to the current transaction as an abnormal video, and lock and save the abnormal video.

[0071] Optionally, the method further includes: extracting a transaction code from the transaction - related information, generating a chapter break point according to the transaction code; adding the chapter break point to the acquisition video corresponding to the current transaction; when the transaction is completed, obtaining the transaction key information of the current transaction, and inserting the transaction key information into the chapter break point to generate a retrieval information point, where the transaction key information includes the transaction serial number, customer number, teller number, transaction code, handling start time, and handling end time.

[0072] The technical solution of the embodiment of the present invention can quickly identify potential anomalies through the scenario rule of a system that is not in use but detects people, making up for the missed detection of environmental anomalies by traditional rules. By matching important transaction codes and directly marking abnormal transactions, it can achieve the rapid response of the rule engine and the pre - risk interception. Using the GBDT algorithm to comprehensively analyze multi - dimensional data such as teller behavior and transaction characteristics to identify hidden risks, improving the identification ability of complex risk scenarios, and reducing the limitations of manual rules.

[0073] Embodiment Three

[0074] Figure 3 It is a structural schematic diagram of a video storage strategy determination device provided by Embodiment Three of the present invention. As Figure 3As shown in the figure, the device includes: a video saving parameter determination module 310, configured to obtain scene-related parameters and dynamically adjust video saving parameters according to the scene-related parameters, where the scene-related parameters include system usage and personnel detection situation;

[0075] A transaction situation determination module 320, configured to obtain transaction-related information of the current transaction when the system detects a transaction, and determine the transaction situation according to the scene-related parameters, the transaction-related information, and a pre-constructed machine learning model;

[0076] An abnormal video locking module 330, configured to mark the captured video corresponding to the current transaction as an abnormal video and save the abnormal video in a locked state when the transaction situation is an abnormal transaction.

[0077] Optionally, the video saving parameter determination module 310 is specifically configured to: determine whether the system usage is that the system is in use, and if so, determine the video saving parameters as a preset first bit rate and first frame rate; otherwise, when the personnel detection situation is that no person is detected, determine the video saving parameters as a preset second bit rate and second frame rate; when the personnel detection situation is that a person is detected, determine the video saving parameters as a preset third bit rate and first frame rate, where the first bit rate is less than the third bit rate and greater than the second bit rate, and the second frame rate is less than the first frame rate.

[0078] Optionally, the transaction situation determination module 320 is specifically configured to: determine that the transaction situation is an abnormal transaction when the system usage is that the system is not in use and the personnel detection situation is that a person is detected; obtain a preset list of important codes, and determine the transaction situation according to the preset list of important codes and the transaction-related information; perform a prediction on the transaction-related information through a pre-constructed machine learning model to determine the transaction situation.

[0079] Optionally, the transaction situation determination module 320 specifically includes: a preset list verification unit, configured to: extract a transaction code and a transaction amount from the transaction-related information; determine that the transaction situation is an abnormal transaction when the transaction code is in the list of important codes; determine that the transaction situation is an abnormal transaction when the transaction amount is greater than a preset amount threshold.

[0080] Optionally, the transaction situation determination module 320 specifically includes: a machine learning prediction unit, configured to: extract transaction features from the transaction-related information; input the transaction features into a pre-constructed machine learning model and obtain an output risk value; determine that the transaction situation is an abnormal transaction when the risk value is greater than or equal to a preset risk threshold.

[0081] Optionally, the device further includes a retrieval information point generation module, configured to: extract a transaction code from transaction-related information, generate chapter breakpoints based on the transaction code; add the chapter breakpoints to the captured video corresponding to the current transaction; after the transaction is completed, obtain the key transaction information of the current transaction, and insert the key transaction information into the chapter breakpoints to generate retrieval information points, where the key transaction information includes a transaction serial number, a customer number, a teller number, a transaction code, a start time of handling, and an end time of handling.

[0082] The technical solution of the embodiment of the present invention can, by dynamically adjusting video saving parameters, achieve on-demand allocation of video storage resources, reduce storage waste during unattended periods, and extend the effective video storage duration. By combining multi-dimensional information with model determination of transaction situations, the comprehensiveness and accuracy of abnormal transaction identification can be improved, covering risk scenarios that are difficult to capture by manual rules. By locking and saving abnormal videos, it can be ensured that key risk videos are not overwritten, improving the reliability and efficiency of risk event traceability.

[0083] A video saving strategy determination device provided by an embodiment of the present invention can execute a video saving strategy determination method provided by any embodiment of the present invention, and has corresponding functional modules and beneficial effects for executing the method.

[0084] Embodiment 4

[0085] Figure 4 FIG. shows a schematic structural diagram of an electronic device 10 that can be used to implement the embodiments of the present invention. The electronic device is intended to represent various forms of digital computers, such as, for example, a laptop computer, a desktop computer, a workbench, a personal digital assistant, a server, a blade server, a mainframe computer, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as, for example, a personal digital assistant, a cellular phone, a smart phone, a wearable device (such as a helmet, glasses, a watch, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present invention described and / or claimed herein.

[0086] As Figure 4As shown, the electronic device 10 includes at least one processor 11 and a memory communicatively connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc. The memory stores a computer program executable by the at least one processor. The processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 into the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0087] Multiple components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a magnetic disk, an optical disc, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0088] The processor 11 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the processor 11 include but are not limited to a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as a video saving strategy determination method.

[0089] In some embodiments, a video saving strategy determination method can be implemented as a computer program tangibly embodied in a computer-readable storage medium, such as the storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed onto the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the video saving strategy determination method described above can be executed. Alternatively, in other embodiments, the processor 11 can be configured to execute a video saving strategy determination method by any other appropriate means (e.g., by means of firmware).

[0090] The various embodiments of the systems and techniques described above in this specification can be implemented in digital electronic circuitry, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on a chip (SOCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include: being implemented in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which may be a special-purpose or general-purpose programmable processor that receives data and instructions from, and transmits data and instructions to, a storage system, at least one input device, and at least one output device.

[0091] The computer programs for implementing the methods of the present invention can be written in any combination of one or more programming languages. These computer programs can be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus, such that the computer programs, when executed by the processor, cause the functions / operations specified in the flowchart and / or block diagram to be implemented. The computer programs can be executed entirely on the machine, partly on the machine, as a stand-alone software package partly on the machine and partly on a remote machine or entirely on the remote machine or server.

[0092] In the context of the present invention, a computer-readable storage medium can be a tangible medium that can contain or store a computer program for use by or in connection with an instruction execution system, apparatus, or device. The computer-readable storage medium can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. Alternatively, the computer-readable storage medium can be a machine-readable signal medium. More specific examples of the machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0093] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or a trackball) through which the user can provide input to the electronic device. Other kinds of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0094] The systems and techniques described herein can be implemented in a computing system including backend components (e.g., as a data server), or a computing system including middleware components (e.g., an application server), or a computing system including frontend components (e.g., a user computer having a graphical user interface or a web browser through which the user can interact with an implementation of the systems and techniques described herein), or a computing system including any combination of such backend components, middleware components, or frontend components. The components of the system can be interconnected to each other by digital data communication in any form or medium (e.g., a communication network). Examples of communication networks include: local area network (LAN), wide area network (WAN), blockchain network, and the Internet.

[0095] The computing system can include a client and a server. The client and the server are generally far from each other and usually interact through a communication network. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or a cloud host, which is a host product in the cloud computing service system and solves the defects of difficult management and weak business scalability existing in traditional physical hosts and VPS services.

[0096] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps recited in the present invention can be executed in parallel, sequentially, or in a different order, as long as the desired results of the technical solution of the present invention can be achieved, and this is not limited herein.

[0097] The above specific embodiments do not constitute a limitation on the protection scope of the present invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for determining a video saving strategy, characterized in that, Including: Obtain scene-related parameters, and dynamically adjust video saving parameters according to the scene-related parameters, where the scene-related parameters include system usage and personnel detection situation; When the system detects a transaction, obtain transaction-related information of the current transaction, and determine the transaction situation according to the scene-related parameters, the transaction-related information, and a pre-constructed machine learning model; When the transaction situation is an abnormal transaction, mark the captured video corresponding to the current transaction as an abnormal video, and lock and save the abnormal video.

2. The method according to claim 1, wherein The dynamically adjusting the video saving parameters according to the scene-related parameters includes: Judge whether the system usage is using the system. If so, determine that the video saving parameters are a preset first bit rate and first frame rate; Otherwise, when no person is detected in the personnel detection situation, determine that the video saving parameters are a preset second bit rate and second frame rate; when a person is detected in the personnel detection situation, determine that the video saving parameters are a preset third bit rate and first frame rate, where the first bit rate is less than the third bit rate and greater than the second bit rate, and the second frame rate is less than the first frame rate.

3. The method according to claim 1, wherein The determining the transaction situation according to the scene-related parameters, the transaction-related information, and a pre-constructed machine learning model includes: When the system usage is not using the system and a person is detected in the personnel detection situation, determine that the transaction situation is an abnormal transaction; Obtain a preset important code list, and determine the transaction situation according to the preset important code list and the transaction-related information; Predict the transaction-related information through a pre-constructed machine learning model to determine the transaction situation.

4. The method according to claim 3, wherein The determining the transaction situation according to the preset important code list and the transaction-related information includes: Extract a transaction code and a transaction amount from the transaction-related information; When the transaction code is in the important code list, determine that the transaction situation is an abnormal transaction; When the transaction amount is greater than a preset amount threshold, determine that the transaction situation is an abnormal transaction.

5. The method according to claim 3, characterized in that, The predicting the transaction-related information through a pre-constructed machine learning model to determine the transaction situation includes: Extract transaction features from the transaction-related information; Input the transaction features into a pre-constructed machine learning model, and obtain the output risk value; When the risk value is greater than or equal to a preset risk threshold, determine that the transaction situation is an abnormal transaction.

6. The method according to claim 1, characterized in that, The method further includes: Extract a transaction code from the transaction-related information, and generate a chapter break point according to the transaction code; Add the chapter break point to the captured video corresponding to the current transaction; After the transaction is completed, obtain the transaction key information of the current transaction, and insert the transaction key information into the chapter break point to generate a retrieval information point, where the transaction key information includes a transaction serial number, a customer number, a teller number, a transaction code, a handling start time, and a handling end time.

7. A video saving strategy determination device, characterized in that Including: A video saving parameter determination module, configured to obtain scene-related parameters, and dynamically adjust video saving parameters according to the scene-related parameters, where the scene-related parameters include system usage and personnel detection situation; A transaction situation determination module, configured to obtain transaction-related information of the current transaction when the system detects the occurrence of a transaction, and determine the transaction situation according to the scenario-related parameters, the transaction-related information, and a pre-constructed machine learning model; An abnormal video locking module, configured to mark the captured video corresponding to the current transaction as an abnormal video and lock and save the abnormal video when the transaction situation is an abnormal transaction.

8. An electronic device, characterized in that, The electronic device includes: At least one processor; And a memory communicatively connected to the at least one processor; Wherein, the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the method according to any one of claims 1-6.

9. A computer storage medium, characterized in that, The computer storage medium stores computer instructions, and the computer instructions are used to implement the method according to any one of claims 1-6 when executed by a processor.

10. A computer program product, characterized in that, The computer program product includes a computer program, and the computer program implements the method according to any one of claims 1-6 when executed by a processor.

Images