An Internet of Things-based recruitment system

Abstract

The present invention discloses a recruitment system based on the Internet of Things, which relates to the technical field of Internet of Things recruitment. The present invention includes: an online recruitment monitoring module, an anti-cheating analysis module, a cheating warning processing module, and a local database. By judging whether the applicant's line of sight frequently shifts in a certain direction, and considering the situation where the enterprise recruiter's line of sight is not on the applicant, the present invention analyzes the comprehensive cheating warning coefficient of the applicant to improve the accuracy of cheating behavior judgment. Considering that peeking usually takes a certain amount of time, sometimes the applicant just shifts the line of sight to other directions due to nervousness, but the duration is usually not too long. Based on this, a more detailed analysis of the applicant's cheating behavior is carried out, and the video images that may have cheating behavior are screened to facilitate the enterprise recruiter to query, thereby improving the recruitment efficiency and reducing the workload of the enterprise recruiter.

Description

Technical Field

[0001] The present invention relates to the technical field of Internet of Things recruitment, and in particular to a recruitment system based on the Internet of Things. Background Art

[0002] Online video interview is an important interview method in current corporate recruitment. Corporate recruiters often ask candidates online questions and take exams. However, due to the lack of offline supervision, corporate interviewers often need to review a large number of video images during the interview to prevent cheating, and review whether the candidates have cheated, which leads to reduced recruitment efficiency. Therefore, it is necessary to reduce the workload of corporate interviewers by studying an anti-cheating analysis system during recruitment.

[0003] For example, the prior art, such as the invention patent application with the publication number: CN114943517A, discloses an electronic recruitment matching method, system and storage medium based on the Internet of Things, and the method includes: using a preset model of pre-set valid resume information to match and screen the preliminary resume information; parsing the valid resume information, determining the sorting of the valid resume information according to the number of characters in each field; and pushing all valid resumes in order according to the sorting results. Another example of the prior art, such as the invention patent application with the publication number: CN115983812A, discloses a recruitment system based on the Internet of Things, and the method includes: improving the applicant's attention to flyers, effectively reducing the random discarding of flyers, facilitating enterprises to recruit employees suitable for positions, and effectively improving the push effect of resume flyers. After the air bag strip is deflated, the net bag is restricted to separate the sticky rubber pad to prevent the sticky rubber pad from affecting the storage of the resume flyer, and the subsequent lifting cover presses the resume flyer that has not been taken away back to the equipment box, which is convenient for the recycling and reuse of the resume flyer.

[0004] It can be seen from the above scheme that the current recruitment system based on the Internet of Things lacks certain attention to judging whether the applicant has cheated by analyzing the direction of the eyes and the distance of pupil movement. When cheating, the applicant mostly uses communication devices such as mobile phones and cheats by sneaking glances. In the entire interview process, it is easy to sneak glances at the communication device many times, so that the applicant often sneaks glances in a specific direction and the pupil moves a large distance. If this is not analyzed, the accuracy of the applicant's cheating judgment is reduced. At the same time, there is a lack of attention to the comprehensive analysis of the eyes of the company's recruitment officer and the eyes of the applicant. During the interview process, the company's recruitment officer will inevitably not look at the applicant. The applicant often chooses to cheat by sneaking glances at this time and gets good results, which makes the company unable to accurately recruit high-quality employees, reduces the quality of corporate recruitment, and increases the workload of the company's recruitment officer's subsequent review. Summary of the invention

[0005] The object of the present invention is to provide a recruitment system based on the Internet of Things, which solves the problems existing in the background technology.

[0006] To solve the above technical problems, the present invention adopts the following technical solutions: The present invention provides a recruitment system based on the Internet of Things, including: an online recruitment monitoring module, which is used to obtain the video images of the enterprise recruiter at each examination time point and the video images of the applicant at each examination time point during the interview examination period when the enterprise conducts online video recruitment.

[0007] An anti-cheating analysis module, which is used to screen each target time point of the applicant, capture the image coordinates of each eye contour feature point and the image coordinates of each pupil feature point of the applicant at each target time point, analyze the pupil offset direction and pupil offset distance of the applicant at each target time point, and evaluate the eye anomaly coefficient of the applicant at each target time point.

[0008] A cheating warning processing module, which is used to judge whether the applicant has a suspicion of eye cheating. If so, analyze each cheating prediction time point of the applicant, extract the video images of the applicant at each cheating prediction time point, and send them to the enterprise recruiter.

[0009] The beneficial effects of the present invention are as follows: (1) The online recruitment monitoring module of the present invention facilitates subsequent analysis by obtaining the video images of the enterprise recruiter and the applicant.

[0010] (2) The anti-cheating analysis module of the present invention analyzes the pupil offset direction and pupil offset distance of the applicant at each target time point to judge whether the applicant has a situation where the line of sight frequently shifts in a certain direction. Considering the situation where the enterprise recruiter's line of sight is not on the applicant, the above situations are comprehensively analyzed, thereby improving the accuracy of the judgment on whether the applicant has a cheating peek behavior.

[0011] (3) The cheating warning processing module of the present invention improves the accuracy of the cheating behavior judgment by analyzing the comprehensive cheating warning coefficient of the applicant. Considering that peeking often takes a certain amount of time, sometimes the applicant just shifts the line of sight to other directions due to nervousness, but the duration is often not too long, and the eye warning coefficient is analyzed based on this, and a more detailed analysis of the applicant's cheating behavior is carried out, and the video images that may have cheating behavior are screened, which is convenient for the enterprise recruiter to query, thereby improving the recruitment efficiency and reducing the workload of the enterprise recruiter. Description of the Drawings

[0012] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. 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.

[0013] Figure 1 It is a schematic diagram of the system modules of the present invention. Specific embodiments

[0014] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying 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 belong to the scope of protection of the present invention.

[0015] Refer to Figure 1 As shown, the present invention provides a recruitment system based on the Internet of Things, including: an online recruitment monitoring module, an anti-cheating analysis module, a cheating warning processing module, and a local database.

[0016] It should be noted that the online recruitment monitoring module is connected to the anti-cheating analysis module, the anti-cheating analysis module is connected to the cheating warning processing module, and the local database is connected to the anti-cheating analysis module and the cheating warning processing module.

[0017] It should also be noted that the local database is used to store the distance values between each pupil feature point and each eye contour feature point of the applicant in the normal state, the parameter adjustment values at the normal time point and the vigilant time point, the range of each offset direction, the threshold of the comprehensive cheating warning coefficient, the preset analysis duration, and the threshold of the eye warning coefficient corresponding to the comprehensive cheating warning coefficient interval.

[0018] The online recruitment monitoring module is used to obtain the video images of the enterprise recruiter at each examination time point and the video images of the applicant at each examination time point during the online video recruitment by the enterprise.

[0019] In a specific embodiment, the method for obtaining the video images of the enterprise recruiter at each examination time point and the video images of the applicant at each examination time point during the interview examination period is as follows: obtaining the video images of the corresponding enterprise recruiter at each examination time point and the video images of the applicant at each examination time point through the video recording devices of the enterprise recruiter and the applicant.

[0020] It should be noted that before obtaining the video images, authorization reminders are given to the enterprise recruiters and candidates. After obtaining the authorization of the enterprise recruiters and candidates, the video images are obtained.

[0021] The online recruitment monitoring module of the present invention facilitates subsequent analysis by obtaining the video images of enterprise recruiters and candidates.

[0022] The anti-cheating analysis module is used to screen the target time points of the candidates, capture the image coordinates of each eye contour feature point and each pupil feature point of the candidates at each target time point, analyze the pupil offset direction and pupil offset distance of the candidates at each target time point, and evaluate the eye anomaly coefficient of the candidates at each target time point.

[0023] In a specific embodiment, the method for screening the target time points of the candidates is as follows: By using the existing image open / close eye recognition technology and the video images of the candidates at each examination time point, the target time points of the candidates can be screened.

[0024] In a specific embodiment of the present invention, the method for capturing the image coordinates of each eye contour feature point and each pupil feature point of the candidates at each target time point is as follows: Based on the video images of the candidates at each examination time point, the video images of the candidates at each target time point are extracted.

[0025] Taking the lower left corner of the video image of the candidate at each target time point as the coordinate origin, the direction from the lower left corner to the lower right corner as the positive x-axis direction, and the direction from the lower left corner to the upper left corner as the positive y-axis direction, an x-y coordinate system is established.

[0026] Through the eye contour image capture technology, the image coordinates of each eye contour feature point and its position of the candidate in the video image at each target time point are obtained, and based on the positions of each eye contour feature point of the candidate in the video image at each target time point, they are mapped one by one to the x-y coordinate system, so as to obtain the image coordinates of each eye contour feature point of the candidate at each target time point.

[0027] Through the pupil contour image capture technology, the image coordinates of each pupil feature point of the candidate at each target time point are obtained in the same way.

[0028] In a specific embodiment of the present invention, the method for analyzing the pupil offset direction and pupil offset distance of the candidate at each target time point is as follows: Based on the image coordinates of each pupil feature point and each eye contour feature point of the candidate at each target time point, the distance values between each pupil feature point and each eye contour feature point of the candidate at each target time point are calculated , where n represents the number of each target time point. , where m is a positive integer greater than 2, i represents the number of each pupil feature point, , where j is a positive integer greater than 2, p represents the number of each eye contour feature point, , where q is a positive integer greater than 2.

[0029] Calculate the reduction change ratio of the distance between each pupil feature point and each eye contour feature point of the applicant at each target time point, extract the pupil feature point and the eye contour feature point with the largest reduction change ratio of the distance at each target time point of the applicant, and mark them as the target pupil feature point and the target eye contour feature point, so as to obtain the target pupil feature point and the target eye contour feature point of the applicant at each target time point. Use the direction from the target pupil feature point to the target eye contour feature point of the applicant at each target time point as the pupil offset direction, so as to obtain the pupil offset direction of the applicant at each target time point.

[0030] Extract the distance values between the target pupil feature point and the target eye contour feature point of the applicant at each target time point, and calculate the pupil offset distance of the applicant at each target time point based on this.

[0031] In a specific embodiment, the method for calculating the reduction change ratio of the distance between each pupil feature point and each eye contour feature point of the applicant at each target time point is as follows: Obtain the distance values between each pupil feature point and each eye contour feature point of the applicant in the normal state from the local database , and calculate the reduction change ratio of the distance between each pupil feature point and each eye contour feature point of the applicant at each target time point .

[0032] It should be noted that the distance values between each pupil feature point and each eye contour feature point of the applicant in the normal state are obtained by acquiring the video image of the applicant in the normal state before the interview test and calculating the distance values between each pupil feature point and each eye contour feature point of the applicant at each target time point.

[0033] In a specific embodiment, the method for calculating the pupil offset distance of the applicant at each target time point is as follows: Based on the distance values between each pupil feature point and each eye contour feature point of the applicant in the normal state, extract the distance value between the target pupil feature point and the target eye contour feature point of the applicant in the normal state , and based on the distance values between the target pupil feature point and the target eye contour feature point of the applicant at each target time point , calculate the pupil offset distance of the applicant at each target time point .

[0034] In a specific embodiment of the present invention, calculating the distance values between each pupil feature point and each eye contour feature point of the applicant at each target time point, the specific calculation method is as follows: based on the image coordinates of each pupil feature point of the applicant at each target time point and the image coordinates of each eye contour feature point , calculate the distance values between each pupil feature point and each eye contour feature point of the applicant at each target time point .

[0035] In a specific embodiment of the present invention, evaluating the eye anomaly coefficient of the applicant at each target time point, the specific evaluation method is as follows: according to the video images of the enterprise recruiter at each examination time point, screen the normal time points and vigilant time points of the applicant

[0036] Based on the pupil offset direction of the applicant at each target time point, analyze the offset distance hazard coefficient of the applicant at each target time point, and extract the offset distance hazard coefficient of the applicant at each normal time point and the offset distance hazard coefficient of each vigilant time point , where t represents the number of each normal time point , w is a positive integer greater than 2, u represents the number of each vigilant time point , v is a positive integer greater than 2

[0037] Obtain the tuning parameter value A for normal time points and the tuning parameter value B for vigilant time points from the local database, and calculate the eye anomaly coefficient of the applicant at each normal time point , and calculate the eye anomaly coefficient of the applicant at each vigilant time point .

[0038] Summarize the eye anomaly coefficients of the applicant at each target time point

[0039] In a specific embodiment of the present invention, screening the normal time points and vigilant time points of the applicant, the specific screening method is as follows: according to the video images of the enterprise recruiter at each examination time point, screen the effective pupil capture time points and non-captured pupil time points of the enterprise recruiter

[0040] In a specific embodiment, screening the effective pupil capture time points and non-captured pupil time points of the enterprise recruiter, the specific screening method is as follows: based on the existing pupil contour image capture technology, the effective pupil capture time points and non-captured pupil time points of the enterprise recruiter can be screened

[0041] If a certain target time point of the applicant is the same as a certain effective pupil capture time point of the enterprise recruiter, then mark this target time point as a normal time point, so as to screen the normal time points of the applicant

[0042] If a certain target time point of an applicant coincides with a certain pupil non - capture time point of an enterprise recruiter, then mark this target time point as a vigilant time point, and thus screen the vigilant time points of the applicant.

[0043] In a specific embodiment of the present invention, when analyzing the offset distance hazard coefficient of an applicant at each target time point, the specific analysis method is as follows: Obtain each offset direction range from the local database. If the pupil offset direction of the applicant at a certain target time point is included in a certain offset direction range, then take this offset direction range as the analysis offset direction range of the applicant at this target time point, and thus screen the analysis offset direction ranges of the applicant at each target time point, map to obtain each target time point of the applicant in each analysis offset direction range, and count the number of target time points of the applicant in each analysis offset direction range , where r represents the number of each analysis offset direction range, , s is a positive integer greater than 2, and calculate the offset frequency hazard coefficient of the applicant in each analysis offset direction range , where e represents the natural constant and s represents the number of analysis offset direction ranges.

[0044] It should be noted that each of the offset direction ranges is divided with the pupil center as the origin, starting from the due north direction, with an angle range of 22.5° for each division, so as to obtain each offset direction range.

[0045] If the pupil offset direction of the applicant at a certain target time point is included in a certain analysis offset direction range, then take the offset frequency hazard coefficient of this analysis offset direction range as the offset direction hazard coefficient of this target time point, and thus obtain the offset direction hazard coefficients of the applicant at each target time point .

[0046] Based on the pupil offset distance of the applicant at each target time point , calculate the offset distance hazard coefficient of the applicant at each target time point , where m is the number of target time points.

[0047] The anti - cheating analysis module of the present invention analyzes the pupil offset direction and pupil offset distance of the applicant at each target time point, determines whether the applicant has a situation where the line of sight frequently shifts in a certain direction, and considering the situation where the enterprise recruiter's line of sight is not on the applicant, comprehensively analyzes the above - mentioned situations, so as to improve the accuracy of judging whether the applicant has the behavior of cheating by stealing glances.

[0048] The cheating warning processing module is used to determine whether a candidate has a suspicion of eye cheating. If so, it analyzes each estimated cheating time point of the candidate, extracts the video images of the candidate at each estimated cheating time point, and sends them to the enterprise recruiter.

[0049] In a specific embodiment of the present invention, to determine whether a candidate has a suspicion of eye cheating, the specific determination method is as follows: based on the eye abnormality coefficients of the candidate at each target time point, the comprehensive cheating warning coefficient of the candidate is analyzed.

[0050] The threshold of the comprehensive cheating warning coefficient is obtained from the local database. If the comprehensive cheating warning coefficient of the candidate is greater than the threshold of the comprehensive cheating warning coefficient, it is determined that the candidate has a suspicion of eye cheating.

[0051] In a specific embodiment of the present invention, to analyze the comprehensive cheating warning coefficient of the candidate, the specific analysis method is as follows: the preset analysis duration is obtained from the local database, a certain target time point of the candidate is randomly selected as the time point to be analyzed, and each target time point within the preset analysis duration before the time point to be analyzed of the candidate is used as each comparison time point, and at the same time, each target time point within the preset analysis duration after the time point to be analyzed of the candidate is used as each comparison time point, so as to obtain the time point to be analyzed of the candidate and its each comparison time point.

[0052] It should be noted that the preset analysis duration is set by enterprise personnel and can be set to, for example, 1s, 2s, 3s, etc. If the enterprise personnel are more strict about the examination, the preset analysis duration can be set smaller, such as 1s.

[0053] Based on the eye abnormality coefficients of the candidate at each target time point , the eye abnormality coefficient D of the candidate at the time point to be analyzed and the eye abnormality coefficients at each comparison time point are extracted , where N represents the number of each comparison time point, , M is a positive integer greater than 2, and the eye warning coefficient of the candidate at the time point to be analyzed is calculated , where M represents the number of comparison time points, and so on, to obtain the eye warning coefficients of the candidate at each target time point , and the comprehensive cheating warning coefficient of the candidate is calculated .

[0054] In a specific embodiment of the present invention, to analyze each estimated cheating time point of the candidate, the specific analysis method is as follows: the eye warning coefficient thresholds corresponding to each comprehensive cheating warning coefficient interval are obtained from the local database, and based on the calculated comprehensive cheating warning coefficient of the candidate, the eye warning coefficient threshold of the candidate is mapped.

[0055] If the eye warning coefficient of an applicant at a certain target time point is greater than the eye warning coefficient threshold, then mark the target time point as a cheating prediction time point, so as to screen out each cheating prediction time point of the applicant.

[0056] It should be noted that the larger the comprehensive cheating warning coefficient, the smaller the corresponding eye warning coefficient threshold. This data is set by the staff. When the comprehensive cheating warning coefficient is larger, it means that the applicant is more likely to cheat, and more video images need to be handed over to the enterprise recruiter for review. At this time, the eye warning coefficient threshold for interception needs to be set smaller to obtain more video images.

[0057] The cheating warning processing module of the present invention improves the accuracy of cheating behavior judgment by analyzing the comprehensive cheating warning coefficient of the applicant. Considering that peeking often has a certain duration, sometimes the applicant just makes a line-of-sight deviation to other directions due to nervousness, but the duration is usually not too long. Based on this, the eye warning coefficient is analyzed to conduct a more detailed analysis of the applicant's cheating behavior, and screen out the video images that may have cheating behavior, which is convenient for the enterprise recruiter to query, thus improving the recruitment efficiency and reducing the workload of the enterprise recruiter.

[0058] The above content is only an example and explanation of the concept of the present invention. Those skilled in the art of this technology can make various modifications or supplements to the described specific embodiments or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, they should all belong to the protection scope of the present invention.

Claims

1. An Internet of Things-based recruitment system, characterized in that, Including: An online recruitment monitoring module, which is used to obtain the video images of the enterprise recruiter at each exam time point and the video images of the applicant at each exam time point during the interview exam period when the enterprise conducts online video recruitment; An anti-cheating analysis module, which is used to screen each target time point of the applicant, capture the image coordinates of each eye contour feature point and the image coordinates of each pupil feature point of the applicant at each target time point, analyze the pupil deviation direction and pupil deviation distance of the applicant at each target time point, and evaluate the eye anomaly coefficient of the applicant at each target time point; The method for evaluating the eye anomaly coefficient of the applicant at each target time point is specifically as follows: According to the video images of the enterprise recruiter at each exam time point, screen each normal time point and each vigilant time point of the applicant; Analyze the hazard coefficient of the deviation distance of the applicant at each target time point according to the pupil deviation direction of the applicant at each target time point, and extract the hazard coefficient g of the deviation distance of the applicant at each normal time point t and the hazard coefficient h of the deviation distance at each warning time point u , where t represents the number of each normal time point, t = 1, 2,..., w, w is a positive integer greater than 2, and u represents the number of each warning time point, u = 1, 2,..., v, v is a positive integer greater than 2; Obtain the normal time point parameter value A and the vigilant time point parameter value B from the local database, and calculate the eye anomaly coefficient λ of the applicant at each normal time point t = g t * A, and calculate the eye anomaly coefficient of the applicant at each vigilant time point Summarize the eye anomaly coefficients of the applicant at each target time point; The method for specifically analyzing the deviation distance hazard coefficient of the applicant at each target time point is as follows: Obtain the range of each offset direction from the local database. If the pupil offset direction of an applicant at a certain target time point is included in a certain offset direction range, then use this offset direction range as the analysis offset direction range of the applicant at this target time point, so as to screen the analysis offset direction ranges of the applicant at each target time point, map to obtain each target time point of the applicant in each analysis offset direction range, and count the number d of target time points of the applicant in each analysis offset direction range r , where r represents the number of each analysis offset direction range, r = 1, 2,..., s, s is a positive integer greater than 2, and calculate the offset frequency hazard coefficient of the applicant in each analysis offset direction range where e represents the natural constant and s represents the number of analysis offset direction ranges; If the pupil deviation direction of an applicant at a certain target time point is included in a certain analysis deviation direction range, the deviation frequency hazard coefficient of this analysis deviation direction range is taken as the deviation direction hazard coefficient of this target time point, so as to obtain the deviation direction hazard coefficient f of the applicant at each target time point n , where n represents the number of each target time point, n = 1, 2,..., m, and m is a positive integer greater than 2; According to the pupil offset distance ε of the candidate at each target time point n , calculate the offset distance hazard coefficient of the candidate at each target time point where m is the number of target time points; A cheating warning processing module, which is used to determine whether the applicant has a suspicion of eye cheating. If so, analyze each cheating prediction time point of the applicant, extract the video images of the applicant at each cheating prediction time point, and send them to the enterprise recruiter.

2. The recruitment system based on the Internet of Things according to claim 1, characterized in that, The method for specifically capturing the image coordinates of each eye contour feature point and each pupil feature point of the applicant at each target time point is as follows: Based on the video images of the applicant at each exam time point, extract the video images of the applicant at each target time point; Taking the lower left corner of the video image of the applicant at each target time point as the coordinate origin, the direction from the lower left corner to the lower right corner as the positive x-axis direction, and the direction from the lower left corner to the upper left corner as the positive y-axis direction, establish an x-y coordinate system; Through the eye contour image capture technology, obtain each eye contour feature point and its position of the applicant in the video image at each target time point, and map them one by one to the x-y coordinate system according to the positions of each eye contour feature point of the applicant in the video image at each target time point, so as to obtain the image coordinates of each eye contour feature point of the applicant at each target time point; Through the pupil contour image capture technology, similarly obtain the image coordinates of each pupil feature point of the applicant at each target time point.

3. The recruitment system based on the Internet of Things according to claim 1, characterized in that, The method for specifically analyzing the pupil deviation direction and pupil deviation distance of the applicant at each target time point is as follows: Based on the image coordinates of each pupil feature point and each eye contour feature point of the applicant at each target time point, calculate the distance value β between each pupil feature point and each eye contour feature point of the applicant at each target time point nip , where i represents the number of each pupil feature point, i = 1, 2,..., j, j is a positive integer greater than 2, p represents the number of each eye contour feature point, p = 1, 2,..., q, q is a positive integer greater than 2; Calculate the distance reduction change ratio between each pupil feature point and each eye contour feature point of the applicant at each target time point, extract the pupil feature point and eye contour feature point with the largest distance reduction change ratio at each target time point of the applicant, and mark them as the target pupil feature point and the target eye contour feature point, so as to obtain the target pupil feature point and the target eye contour feature point of the applicant at each target time point. Take the direction from the target pupil feature point to the target eye contour feature point of the applicant at each target time point as the pupil deviation direction, so as to obtain the pupil deviation direction of the applicant at each target time point; Extract the distance value between the target pupil feature point and the target eye contour feature point of the applicant at each target time point, and calculate the pupil deviation distance of the applicant at each target time point accordingly.

4. The recruitment system based on the Internet of Things according to claim 3, characterized in that, Calculating the distance values between each pupil feature point and each eye contour feature point of the applicant at each target time point, and the specific calculation method is as follows: According to the image coordinates (x ni , y ni ) of each pupil feature point and the image coordinates (x np , y np ) of each eye contour feature point of the applicant at each target time point, calculate the distance values between each pupil feature point and each eye contour feature point of the applicant at each target time point 5. The recruitment system based on the Internet of Things according to claim 1, characterized in that, Screening each normal time point and each vigilant time point of the applicant, and the specific screening method is as follows: Based on the video images of the enterprise recruiter at each examination time point, screening each pupil effective capture time point and each pupil non-capture time point of the enterprise recruiter; If a certain target time point of the applicant is the same as a certain pupil effective capture time point of the enterprise recruiter, then mark this target time point as a normal time point, so as to screen each normal time point of the applicant; If a certain target time point of the applicant is the same as a certain pupil non-capture time point of the enterprise recruiter, then mark this target time point as a vigilant time point, so as to screen each vigilant time point of the applicant.

6. The recruitment system based on the Internet of Things according to claim 1, characterized in that Judging whether the applicant has a suspicion of eye contact cheating, and the specific judgment method is as follows: Analyzing the comprehensive cheating warning coefficient of the applicant based on the abnormal eye contact coefficient of the applicant at each target time point; Obtaining the comprehensive cheating warning coefficient threshold from the local database. If the comprehensive cheating warning coefficient of the applicant is greater than the comprehensive cheating warning coefficient threshold, it is judged that the applicant has a suspicion of eye contact cheating.

7. An Internet of Things-based recruitment system according to claim 6, characterized in that, Analyzing the comprehensive cheating warning coefficient of the applicant, and the specific analysis method is as follows: Obtaining the preset analysis duration from the local database, randomly selecting a certain target time point of the applicant as the time point to be analyzed, taking each target time point within the preset analysis duration before the time point to be analyzed of the applicant as each comparison time point, and at the same time taking each target time point within the preset analysis duration after the time point to be analyzed of the applicant as each comparison time point, so as to obtain the time point to be analyzed of the applicant and each of its comparison time points; According to the eye anomaly coefficient E of the applicant at each target time point n , extract the eye anomaly coefficient D of the applicant at the time point to be analyzed and the eye anomaly coefficients C at each comparison time point N , where N represents the number of each comparison time point, N = 1, 2,..., M, and M is a positive integer greater than 2. Calculate the eye warning coefficient of the applicant at the time point to be analyzed , where M represents the number of comparison time points, and so on, to obtain the eye warning coefficient F of the applicant at each target time point n , calculate the comprehensive cheating warning coefficient of the applicant 8. An Internet of Things-based recruitment system according to claim 7, characterized in that Analyzing each cheating prediction time point of the applicant, and the specific analysis method is as follows: Obtaining the eye contact warning coefficient thresholds corresponding to each comprehensive cheating warning coefficient interval from the local database, and mapping to obtain the eye contact warning coefficient threshold of the applicant according to the comprehensive cheating warning coefficient of the applicant; If the eye contact warning coefficient of the applicant at a certain target time point is greater than the eye contact warning coefficient threshold, then mark this target time point as a cheating prediction time point, so as to screen each cheating prediction time point of the applicant.

Images