An online management system for pet insurance based on image recognition

Abstract

The present invention discloses an online management system for pet insurance based on image recognition, belonging to the technical field, which includes an electronic insurance policy preprocessing module, a policy area recognition module, a policy area information storage module, and a cloud storage module. The present invention can conveniently perform active partitioning and information recognition on electronic insurance policy pictures, improving the efficiency of picture information acquisition and the speed of subsequent classification storage, making the entire management system more efficient; it can associate four types of information in the electronic insurance policy, making it more convenient for staff to manage and view, effectively improving the management efficiency; it can conveniently identify the pet type result corresponding to the current electronic insurance policy based on the preset graphic identifier in the pet type information area picture, without manual participation during the recognition process, and completely solves the problems such as low recognition accuracy existing in traditional pet type recognition through nose prints, etc.

Description

Technical Field

[0001] The present invention relates to the technical field of image recognition processing, and particularly relates to an online management system for pet insurance based on image recognition. Background Art

[0002] Pets are very common in modern life. A pet refers to a creature that people raise for spiritual purposes rather than economic purposes. Traditional pets refer to animals of the class Mammalia or Aves, which are raised for play and companionship. In real life, pets include fish, reptiles, amphibians, insects, and even plants, which are used for viewing, companionship, and relieving people's mental stress.

[0003] With the development of the times, the scope of pets is very wide, including animals, plants, virtual pets, electronic pets, etc. However, the laws of most countries still limit pets to the category of animals. Raising pets as an opportunity for humans to get close to nature can meet the psychological needs of humans and is a very healthy and normal hobby.

[0004] Pet insurance refers to a commercial insurance behavior in which the applicant pays insurance premiums for the insured pet according to the contract agreement, and the insurer undertakes the liability of paying insurance benefits for the injury or loss of the pet caused by the occurrence of an accident that may occur as agreed in the contract, or when the insured pet dies, is disabled, is ill, or reaches the age or term agreed in the contract.

[0005] With the increase in the number of pets and the growth of expenses for pet medical treatment, surgery, etc., the expenditure pressure of pet owners in the medical aspect is increasing continuously, which objectively stimulates the insurance demand of pet owners. In addition, the reduction or abolition of registration fees and annual inspection fees has also further expanded the potential scope of pet insurance, and the improvement of the status of pets has also increased the subjective insurance demand of the owners.

[0006] However, in the existing data management technology of pet insurance, pet information, applicant information, beneficiary information, etc. need to be manually entered, which is very troublesome to operate, and the pet insurance-related information after entry is also too redundant and inconvenient to manage. For this reason, an online management system for pet insurance based on image recognition is proposed. Summary of the Invention

[0007] The technical problem to be solved by the present invention is: how to solve the problem that in the existing data management technology of pet insurance, pet information, applicant information, beneficiary information, etc. need to be manually entered, which is very troublesome to operate, and the pet insurance-related information after entry is also too redundant and inconvenient to manage, and provides an online management system for pet insurance based on image recognition.

[0008] The present invention solves the above technical problem through the following technical solutions. The present invention includes:

[0009] An electronic insurance policy preprocessing module, which is used to preprocess the electronic insurance policy pictures of each pet insurance and send the preprocessed electronic insurance policy pictures to the policy area recognition module;

[0010] The policy area recognition module is used to recognize four types of information areas in the electronic insurance policy picture. After obtaining the area pictures of the four types of information, it sends the area pictures of the four types of information to the policy area information storage module. The four types of information include pet type information, pet picture information, applicant information, and beneficiary information;

[0011] The policy area information storage module is used to recognize the pet type information in the electronic insurance policy picture and recognize the other three types of information, and associate and store the pet picture information with the applicant information, beneficiary information, and pet type information;

[0012] The cloud storage module is used to upload the data content contained in the first-level database and the second-level database of the policy area information storage module to the cloud in real time.

[0013] Furthermore, in the electronic insurance policy preprocessing module, the preprocessing includes image denoising and image enhancement processing.

[0014] Furthermore, the policy area recognition module includes a pet type information area recognition unit, a pet picture information area recognition unit, an applicant information area recognition unit, and a beneficiary information area recognition unit; the pet type information area recognition unit is used to recognize the pet type information area in the electronic insurance policy picture through a trained first target recognition network and obtain the pet type information area picture; the pet picture information area recognition unit is used to recognize the pet picture information area in the electronic insurance policy picture through a trained second target recognition network and obtain the pet picture information area picture; the applicant information area recognition unit is used to recognize the applicant information area in the electronic insurance policy picture through a trained third target recognition network and obtain the applicant information area picture; the beneficiary information area recognition unit is used to recognize the beneficiary information area in the electronic insurance policy picture through a trained fourth target recognition network and obtain the beneficiary information area picture.

[0015] Furthermore, the bottom sheet area information storage module includes a pet type information recognition unit, an area information processing unit, and an information association unit; the pet type information recognition unit is used to recognize the preset graphic identifiers in the pet type information area picture, obtain the pet type result corresponding to the current electronic insurance bottom sheet, and establish a first-level database with the pet type result; the area information processing unit is used to segment the pet head picture in the pet picture information area picture along its contour, obtain the segmented pet head picture, that is, the pet head picture corresponding to the current electronic insurance bottom sheet, store it in the established second-level database, and is used to recognize the text information in the applicant information area picture and the beneficiary information area picture, obtain the applicant information and beneficiary information corresponding to the current electronic insurance bottom sheet, and generate applicant information entries and beneficiary information entries respectively; the information association unit is used to associate the applicant information entries and beneficiary information entries corresponding to the current electronic insurance bottom sheet with the pet head picture and store them in the second-level database.

[0016] Furthermore, there are two ways to upload the electronic insurance bottom sheet pictures of each pet insurance into the system. The first is to upload by taking pictures of the electronic insurance bottom sheet pictures; the second is to directly upload the original pictures of the electronic insurance bottom sheet.

[0017] Furthermore, any one of the RCNN, SSD, and Yolo networks is selected as the original target recognition network for the first target recognition network, the second target recognition network, the third target recognition network, and the fourth target recognition network.

[0018] Furthermore, the first target recognition network, the second target recognition network, the third target recognition network, and the fourth target recognition network identify each information area according to the size characteristics of the rectangular frames at the edges of each information area in the electronic insurance bottom sheet picture.

[0019] Furthermore, the specific process of the pet type information recognition unit recognizing the preset graphic identifier includes the following steps:

[0020] S11: First, obtain the pet type information area picture, use the trained fifth target recognition network to recognize the type of the preset graphic identifier in the picture, obtain the pet type information corresponding to the current electronic insurance bottom sheet picture, establish a first-level database with the pet type result, and generate an index about the pet type information.

[0021] S12: Use the contour detection function in OpenCV to detect the outer contour of the recognized preset graphic identifier, calculate the number of line segments outside the preset graphic identifier according to the contour recognition result, and different numbers of line segments represent the sub-classification information under different pet types, and generate a sub-classification index under the same pet type in the first-level database.

[0022] The present invention has the following advantages compared with the prior art: For the online management system of pet insurance based on image recognition, through the set bottom form area recognition module, it can conveniently perform active partitioning and information recognition on the electronic insurance bottom form pictures, greatly improving the efficiency of obtaining picture information and the speed of subsequent classification and storage, making the entire management system more efficient; through the set information association unit, it can associate the four types of information in the electronic insurance bottom form, making it more convenient for staff to manage and view, and effectively improving the management efficiency; through the set pet type information recognition unit, it can conveniently identify the pet type result corresponding to the current electronic insurance bottom form according to the preset graphic identifier in the pet type information area picture. The recognition process does not require manual participation and completely solves the problems such as low recognition accuracy in traditional pet type recognition through nose prints, etc. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 is a block diagram of the online management system of pet insurance based on image recognition in an embodiment of the present invention;

[0024] Figure 2 is a schematic structural diagram of the electronic insurance bottom form picture in an embodiment of the present invention;

[0025] Figure 3 a in

[0026] Figure 3 is a schematic diagram of a circular preset graphic identifier in an embodiment of the present invention;

[0027] Figure 3 b in is a schematic diagram of a square preset graphic identifier in an embodiment of the present invention;

[0028] The following is a detailed description of the embodiments of the present invention. This embodiment is implemented on the premise of the technical solution of the present invention, and detailed implementation manners and specific operation processes are given. However, the protection scope of the present invention is not limited to the following embodiments.

[0029] As Figure 1 shown, this embodiment provides a technical solution: An online management system of pet insurance based on image recognition, including an electronic insurance bottom form preprocessing module, a bottom form area recognition module, a bottom form area information storage module, and a cloud storage module;

[0030] In this embodiment, the electronic insurance policy preprocessing module is used to preprocess the electronic insurance policy pictures of each pet insurance. The preprocessing includes image denoising and image enhancement processing, so as to obtain electronic insurance policy pictures with clearer features, which is more convenient for subsequent recognition processing. The preprocessed electronic insurance policy pictures are sent to the policy area recognition module for recognition processing.

[0031] In this embodiment, the policy area recognition module includes a pet type information area recognition unit, a pet picture information area recognition unit, an applicant information area recognition unit, and a beneficiary information area recognition unit. The pet type information area recognition unit is used to recognize the pet type information area in the electronic insurance policy picture through the trained first target recognition network and obtain the pet type information area picture. The pet picture information area recognition unit is used to recognize the pet picture information area in the electronic insurance policy picture through the trained second target recognition network and obtain the pet picture information area picture. The applicant information area recognition unit is used to recognize the applicant information area in the electronic insurance policy picture through the trained third target recognition network and obtain the applicant information area picture. The beneficiary information area recognition unit is used to recognize the beneficiary information area in the electronic insurance policy picture through the trained fourth target recognition network and obtain the beneficiary information area picture. After obtaining the area pictures of the four types of information, the area pictures of the above four types of information are sent to the policy area information storage module for classification, recognition, and storage. By setting the policy area recognition module, the electronic insurance policy pictures can be conveniently partitioned and information recognized actively, greatly improving the efficiency of picture information acquisition and the speed of subsequent classification and storage, making the entire management system more efficient.

[0032] In this embodiment, the bottom sheet area information storage module includes a pet type information recognition unit, an area information processing unit, and an information association unit; the pet type information recognition unit is used to recognize the preset graphic identifier in the pet type information area picture, obtain the pet type result corresponding to the current electronic insurance bottom sheet, and establish a first-level database with the pet type result; the area information processing unit is used to segment the pet head picture in the pet picture information area picture along its contour, obtain the segmented pet head picture, that is, the pet head picture corresponding to the current electronic insurance bottom sheet, store it in the established second-level database, and is used to recognize the text information in the applicant information area picture and the beneficiary information area picture, obtain the applicant information and beneficiary information corresponding to the current electronic insurance bottom sheet, and generate applicant information entries (txt format) and beneficiary information entries (txt format) respectively; the information association unit is used to associate the applicant information entries and beneficiary information entries corresponding to the current electronic insurance bottom sheet with the pet head picture and store them in the second-level database; through the set information association unit, the four types of information in the electronic insurance bottom sheet can be associated, which is more convenient for the staff to manage and view, and effectively improves the management efficiency; through the set pet type information recognition unit, it is possible to conveniently recognize the pet type result corresponding to the current electronic insurance bottom sheet according to the preset graphic identifier in the pet type information area picture. The recognition process does not require manual participation and completely solves the problems such as low recognition accuracy in the traditional method of recognizing pet types through nose prints, etc.

[0033] In this embodiment, the cloud storage module is used to upload the data content contained in the first-level database and the second-level database of the bottom sheet area information storage module to the cloud in real time for subsequent traceability or investigation.

[0034] In this embodiment, there are two ways to upload the electronic insurance bottom sheet pictures of each pet insurance into the system. The first is to upload by taking pictures of the electronic insurance bottom sheet pictures. In this upload method, due to the influence of the shooting environment, etc., it is necessary to perform image denoising and image enhancement processing on the obtained electronic insurance bottom sheet pictures. The second is to directly upload the electronic insurance bottom sheet pictures (i.e., the original pictures). In this upload method, generally, it is not necessary to perform image denoising and image enhancement processing on the electronic insurance bottom sheet pictures.

[0035] In this embodiment, the image denoising method can be either arithmetic mean filtering or Gaussian filtering, but is not limited to the above methods.

[0036] In this embodiment, when the first target recognition network, the second target recognition network, the third target recognition network, and the fourth target recognition network are trained, first, a large number of pictures are manually labeled and used as a training set to be input into the original target recognition network for training. After optimizing the network model, the network model parameters are determined, that is, the network model structure is determined, and then the structures of the first target recognition network, the second target recognition network, the third target recognition network, and the fourth target recognition network are determined.

[0037] In this embodiment, the original target recognition network of the first target recognition network, the second target recognition network, the third target recognition network, and the fourth target recognition network selects any one of the RCNN, SSD, and Yolo networks, but is not limited to the above types.

[0038] In this embodiment, the first target recognition network, the second target recognition network, the third target recognition network, and the fourth target recognition network identify each information area according to the size characteristics of the rectangular frames at the edges of each information area in the electronic insurance policy image.

[0039] Specifically, as Figure 2 shown, it is a schematic structural diagram of an electronic insurance policy image. In the figure, the pet picture information area 1 contains a pet head picture, the pet type information area 2 contains a recognizable preset graphic logo, the policyholder information area 3 contains the printed policyholder information, and the beneficiary information area 4 contains the beneficiary information. The lengths and widths of the rectangular frames at the edges of the pet picture information area 1, the pet type information area 2, the policyholder information area 3, and the beneficiary information area 4 gradually increase and are all fixed values, which is convenient for the first target recognition network, the second target recognition network, the third target recognition network, and the fourth target recognition network to identify.

[0040] As Figure 3 shown in a of, it is a schematic diagram of a circular preset graphic logo, corresponding to indicating that the pet type is "dog". As Figure 3 shown in b of, it is a schematic diagram of a square preset graphic logo, corresponding to indicating that the pet type is "cat". As Figure 3 shown in c of, it is a schematic diagram of a triangular preset graphic logo, corresponding to indicating that the pet type is "duck", including but not limited to the above types of preset graphic logos. Since there are too many existing pet types, it is temporarily impossible to list them all.

[0041] In this embodiment, the specific process of the pet type information recognition unit for recognizing the preset graphic logo includes the following steps:

[0042] S11: First, obtain the picture of the pet type information area, use the trained fifth target recognition network to recognize the types of preset graphic identifiers in the picture, and obtain the pet type information corresponding to the current electronic insurance policy form picture. Establish a first-level database based on the pet type results and generate an index for the pet type information.

[0043] S12: Use the contour detection function in OpenCV to detect the external contour of the recognized preset graphic identifier, and calculate the number of line segments outside the preset graphic identifier according to the contour recognition result ( Figure 3 The type when only one line segment is shown in the figure). Different numbers of line segments represent the sub-category information under different pet types, and generate a sub-category index under the same pet type in the first-level database. Generating two-level indexes makes it more convenient to view and manage.

[0044] For example Figure 3 a in it indicates that the pet type corresponding to the current electronic insurance policy form picture is "dog", and the breed of the dog is "Chinese rural dog"; Figure 3 b in it indicates that the pet type corresponding to the current electronic insurance policy form picture is "cat", and the breed of the cat is "orange cat", Figure 3 c in it indicates that the pet type corresponding to the current electronic insurance policy form picture is "duck", and the breed of the duck is "Call duck".

[0045] It should be noted that the fifth target recognition network can recognize three types of targets (the shapes of three preset graphic identifiers), and is also trained based on any one of the RCNN, SSD, and Yolo networks.

[0046] In summary, the above-mentioned pet insurance online management system based on image recognition can conveniently perform active partitioning and information recognition on the electronic insurance policy form picture through the set bottom form area recognition module, greatly improving the efficiency of obtaining picture information and the speed of subsequent classification storage, making the entire management system more efficient; through the set information association unit, the four types of information in the electronic insurance policy form can be associated, making it more convenient for staff to manage and view, and effectively improving the management efficiency; through the set pet type information recognition unit, it can conveniently identify the pet type result corresponding to the current electronic insurance policy form according to the preset graphic identifier in the pet type information area picture. The recognition process does not require manual participation and completely solves the problems such as low recognition accuracy in traditional pet type recognition through nasal patterns.

[0047] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include at least one such feature. In the description of the present invention, "a plurality" means at least two, such as two, three, etc., unless otherwise specifically defined.

[0048] In the description of this specification, the description with reference to terms such as "one embodiment", "some embodiments", "examples", "specific examples", or "some examples", etc. means that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0049] Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention.

Claims

1. An online management system for pet insurance based on image recognition, characterized in that, Including: An electronic insurance policy preprocessing module, which is used to preprocess the electronic insurance policy pictures of each pet insurance and send the preprocessed electronic insurance policy pictures to the policy area recognition module; The policy area recognition module is used to recognize four types of information areas in the electronic insurance policy picture. After obtaining the area pictures of the four types of information, it sends the area pictures of the four types of information to the policy area information storage module. The four types of information include pet type information, pet picture information, applicant information, and beneficiary information; The policy area information storage module is used to recognize the pet type information in the electronic insurance policy picture, and recognize the other three types of information, and associate and store the pet picture information with the applicant information, beneficiary information, and pet type information; The cloud storage module is used to upload the data content contained in the first-level database and the second-level database of the policy area information storage module to the cloud in real time; The specific process of the policy area recognition module for recognizing the preset graphic identifier includes the following steps: S11: First, obtain the area picture of the pet type information, use the trained fifth target recognition network to recognize the type of the preset graphic identifier in the picture, and obtain the pet type information corresponding to the current electronic insurance policy picture. Establish a first-level database with the pet type result and generate an index about the pet type information; S12: Use the contour detection function in OpenCV to detect the external contour of the recognized preset graphic identifier, calculate the number of line segments outside the preset graphic identifier according to the contour recognition result. Different numbers of line segments represent the sub-classification information under different pet types, and generate a sub-classification index under the same pet type in the first-level database.

2. The online management system for pet insurance based on image recognition according to claim 1, characterized in that: In the electronic insurance policy preprocessing module, the preprocessing includes image denoising and image enhancement processing.

3. The online management system for pet insurance based on image recognition according to claim 2, characterized in that: The policy area recognition module includes a pet type information area recognition unit, a pet picture information area recognition unit, an applicant information area recognition unit, and a beneficiary information area recognition unit; the pet type information area recognition unit is used to recognize the pet type information area in the electronic insurance policy picture through the trained first target recognition network and obtain the area picture of the pet type information; the pet picture information area recognition unit is used to recognize the pet picture information area in the electronic insurance policy picture through the trained second target recognition network and obtain the area picture of the pet picture information; the applicant information area recognition unit is used to recognize the applicant information area in the electronic insurance policy picture through the trained third target recognition network and obtain the area picture of the applicant information; the beneficiary information area recognition unit is used to recognize the beneficiary information area in the electronic insurance policy picture through the trained fourth target recognition network and obtain the area picture of the beneficiary information.

4. The online management system for pet insurance based on image recognition according to claim 3, characterized in that: The bottom single area information storage module includes a pet type information recognition unit, an area information processing unit, and an information association unit; the pet type information recognition unit is used to recognize the preset graphic identifier in the pet type information area picture, obtain the pet type result corresponding to the current electronic insurance bottom single, and establish a first-level database with the pet type result; the area information processing unit is used to segment the pet head picture in the pet picture information area picture along its contour, obtain the segmented pet head picture, that is, the pet head picture corresponding to the current electronic insurance bottom single, store it in the established second-level database, and is used to recognize the text information in the applicant information area picture and the beneficiary information area picture, obtain the applicant information and beneficiary information corresponding to the current electronic insurance bottom single, and generate applicant information entries and beneficiary information entries respectively; the information association unit is used to associate the applicant information entries and beneficiary information entries corresponding to the current electronic insurance bottom single with the pet head picture and store them in the second-level database.

5. The online management system for pet insurance based on image recognition according to claim 4, characterized in that: There are two ways to upload the electronic insurance bottom single pictures of each pet insurance into the system. The first is to upload by taking pictures of the electronic insurance bottom single pictures; the second is to directly upload the original pictures of the electronic insurance bottom single pictures.

6. The online management system for pet insurance based on image recognition according to claim 5, characterized in that: The original target recognition networks of the first target recognition network, the second target recognition network, the third target recognition network, and the fourth target recognition network are selected from any one of the RCNN, SSD, and Yolo networks.

7. The online management system for pet insurance based on image recognition according to claim 6, characterized in that: The first target recognition network, the second target recognition network, the third target recognition network, and the fourth target recognition network recognize each information area according to the size characteristics of the rectangular frames at the edges of each information area in the electronic insurance bottom single picture during recognition.

Images