Financial Surveillance Image Management Methods and Systems

Abstract

A financial surveillance video management system includes an image database storing multiple sets of image data generated by multiple monitors, a core financial database storing multiple sets of account data corresponding to multiple accounts, and a monitoring and management server connected to the image database and the core financial database via a communication network. For each set of account data, the monitoring and management server uses an abnormal account prediction model to determine whether the account data is abnormal. For each set of account data determined to be abnormal, the monitoring and management server obtains multiple sets of target image data from the image data based on the account data. For each set of target image data, the monitoring and management server extends the retention period of the target image data.

Claims

1. A financial surveillance image management method, implemented by a management system, the management system including an image database, a financial core database, and a monitoring and management server connected to the image database and the financial core database via a communication network, the image database storing multiple sets of image data related to multiple monitors, each set of image data including a shooting location and a retention period, the financial core database storing multiple sets of account data corresponding to multiple accounts, each set of account data including multiple sets of transaction data, each set of transaction data having a transaction location, the method comprising the following steps: (A) For each set of account data, the monitoring and management server, based on the transaction data of the account data, uses an abnormal account prediction model for determining whether an account data is abnormal, to determine whether the account data is abnormal; (B) For each account data identified as abnormal, the monitoring and management server obtains multiple target image data from the image data based on the transaction location of all transaction data of the account data, wherein the distance between the shooting location of each target image data and the transaction location of one of the transaction data of the account data is less than a distance threshold; and (C) For each target image data, the monitoring and management server extends the retention period of the target image data.

2. The financial surveillance image management method as described in claim 1, wherein, Step (A) includes the following sub-steps: (A-1) For each account data, the monitoring and management server generates a prediction result including an anomaly probability based on the transaction data of the account data using the abnormal account prediction model; (A-2) The monitoring and management server determines whether the anomaly probability of the prediction result is greater than a first anomaly threshold; (A-3) When it is determined that the anomaly probability of the prediction result is not greater than the first anomaly threshold, the monitoring and management server determines that the account data is not abnormal; and (A-4) When it is determined that the anomaly probability of the prediction result is greater than the first anomaly threshold, the monitoring and management server determines that the account data is abnormal.

3. The financial surveillance image management method as described in claim 2, wherein, Following sub-step (A-2), the following sub-steps are also included: (A-5) When it is determined that the probability of an anomaly in the prediction result is not greater than the first anomaly threshold, the monitoring and management server will also classify the account information as low-risk account information; (A-6) When it is determined that the probability of an anomaly in the prediction result is greater than the first anomaly threshold, the monitoring and management server will also determine whether the probability of an anomaly in the prediction result is greater than a second anomaly threshold greater than the first anomaly threshold; (A-7) When it is determined that the probability of an anomaly in the prediction result is not greater than the second anomaly threshold, the monitoring and management server will classify the account information as medium-risk account information; and (A-8) When it is determined that the probability of an anomaly in the prediction result is greater than the first anomaly threshold, the monitoring and management server will classify the account information as high-risk account information; In step (B), for each high-risk account data, the monitoring and management server obtains multiple high-risk target image data from the image data; for each medium-risk account data, the monitoring and management server obtains multiple medium-risk target image data from the image data; in step (C), the extended retention period corresponding to each high-risk target image data is longer than the extended retention period corresponding to each medium-risk target image data.

4. The financial surveillance image management method as described in claim 3, wherein, In step (C), for each medium-risk target image data, the monitoring and management server multiplies the retention period of the medium-risk target image data by a medium-risk multiple greater than 1; for each high-risk target image data, the monitoring and management server multiplies the retention period of the high-risk target image data by a high-risk multiple greater than the medium-risk multiple.

5. The financial surveillance image management method as described in claim 1, the management system further includes a geographic information database connected to the surveillance management server via the communication network, the geographic information database storing multiple geographic location information data corresponding to the monitors respectively, and after step (A) further includes the following step: (D) the surveillance management server generates a financial fraud map based on the transaction location of all account data that are identified as abnormal, and the geographic location information data.

6. A financial surveillance video management system, comprising: an image database storing multiple sets of image data generated by multiple monitors, each set of image data including a shooting location and a retention period; a financial core database storing multiple sets of account data corresponding to multiple accounts, each set of account data including multiple transaction data, each transaction data having a transaction location; and a monitoring management server connected to the image database and the financial core database via a communication network; wherein... For each account data, the monitoring and management server uses an abnormal account prediction model to determine whether the account data is abnormal based on the transaction data of the account data. For each account data that is determined to be abnormal, the monitoring and management server obtains multiple target image data from the image data based on the transaction locations of all transaction data of the account data. The distance between the shooting location of each target image data and the transaction location of one of the transaction data of the account data is less than a distance threshold. For each target image data, the monitoring and management server extends the retention period of the target image data.

7. The financial surveillance video management system as described in claim 6, wherein, For each account, the monitoring and management server uses the abnormal account prediction model to generate a prediction result including an abnormal probability based on the transaction data of the account. The monitoring and management server determines whether the abnormal probability of the prediction result is greater than a first abnormal threshold. When the abnormal probability of the prediction result is determined to be greater than the first abnormal threshold, the monitoring and management server determines that the account is abnormal. When the abnormal probability of the prediction result is determined not to be greater than the first abnormal threshold, the monitoring and management server determines that the account is not abnormal.

8. The financial surveillance video management system as described in claim 7, wherein, When the probability of an anomaly in the predicted result is determined to be no greater than the first anomaly threshold, the monitoring and management server will still classify the account information as low-risk account information. When the probability of anomaly in the predicted result is determined to be greater than the first anomaly threshold, the monitoring and management server will further determine whether the probability of anomaly in the predicted result is greater than a second anomaly threshold greater than the first anomaly threshold. When the probability of anomaly in the predicted result is determined to be greater than the second anomaly threshold, the monitoring and management server will classify the account information as high-risk account information. When the probability of anomaly in the predicted result is determined to be no greater than the first anomaly threshold, the monitoring and management server will classify the account information as medium-risk account information. For each high-risk account information, the monitoring and management server will obtain multiple high-risk target image data from the image data. For each medium-risk account information, the monitoring and management server will obtain multiple medium-risk target image data from the image data. The extended retention period for each high-risk target image data is longer than the extended retention period for each medium-risk target image data.

9. The financial surveillance video management system as described in claim 8, wherein, For each medium-risk target image, the monitoring and management server multiplies the retention period of the medium-risk target image by a medium-risk multiple greater than 1. For each high-risk target image, the monitoring and management server multiplies the retention period of the high-risk target image by a high-risk multiple greater than the medium-risk multiple.

10. The financial surveillance video management system as described in claim 6 further comprises: a geographic information database connected to the surveillance management server via the communication network, storing multiple sets of geographic location information data corresponding to the respective monitors; wherein, The monitoring and management server generates a financial fraud map based on the transaction locations of all accounts identified as having anomalies, as well as such geographical location information.

Images