KR20100119914A - The method and system to achieve the reality of examinee in on-line test using a webcam - Google Patents

Abstract

PURPOSE: An authenticity holding method of an online evaluation examinee and a system thereof using a webcam by comparing pictures are provided to implement a real-time test supervising effect by comparing the pictures. CONSTITUTION: A reference image storage unit stores a digital image of an examinee. A test preparation photo storage unit stores the photograph of the examinee which is before test. A compared photo storage unit stores the process of the captured photograph. A recognition module program storage unit stores the program evaluated the fraudulent action.

Description

The method and system to achieve the reality of examinee in on-line test using a webcam}

      The present invention relates to the authenticity of the online test taker, and more specifically, the authenticity of the online test taker who has a real-time test supervision effect by comparing photographs obtained before and during the test and evaluating whether or not cheating is judged by the sameness of face and environment. It relates to a method and a system to secure.

 Advanced evaluation fairness is required according to the increasing demand for online evaluation for convenience and long-distance reasons, and cheating that may occur in one way will undermine the test fairness and unfairness of the test taker and test taker, thereby ensuring the authenticity of oneself. Only the importance of online evaluation is growing.

 Current participants (especially cyber universities) may consider disabling or enabling the use of function keys in the online evaluation, recording the use of the function keys, and using the function key detection function as warning or cheating. We are passively responding to anticipated misconduct as a means of securing data for the public, or by constantly changing test systems and evaluation methods. However, as the number of test users increases, some participants are concerned about fairness, and prospective consumers who are interested in the online evaluation are hesitant about the evaluation.

 Fingerprint recognition and iris recognition technology is already advanced, but it requires a separately manufactured sensor and requires a certain operation to give the evaluator time to respond to the authentication. This technique is insufficient for identity verification to be applied in an online test with the advantage of remoteness.

 Republic of Korea Patent Publication No. 2007-0050029 (Application No. 10-2007-0039984) illustrates the application to attendance certification and examination of online training using a steel cut using a camera connected to the terminal. It is a method of requesting a specific action to obtain a frontal picture, and when applied to the test, attention is distributed and it is inconvenient to apply to the test.

 In addition, Troy University, Alabama, USA (Patent No. 10-1998-0709046) uses an expensive device that combines a microphone, a camera, and fingerprint recognition to test subjects' video and sound data in real time. It is also controlled by the evaluator's server. In case of ineligible reasons during the test, the test will be stopped immediately and there is much controversy.In the case of real-time detection of ineligible reasons, the administrator should observe the test management status on the main server all the time. The server burden will be high enough.

 SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problems. In order to evaluate whether a subject of a particular subject actually took a test in an online test, the time interval and front of the situation under test are preferably minimized. The goal is to ensure the authenticity of the online test taker, which makes the situation that the candidates take the exams naturally, and the test takers do not have to be psychologically burdened.

 Candidates must consider both the temporal and spatial position of the candidate to gain real-time oversight of the online exam, and the candidate should value the ease and fairness of the assessment. Examine the candidate and the surrounding environment at appropriate time intervals throughout the test time, and use the face recognition algorithm to obtain an accurate picture.In the case of a large number of candidates, consider the burden on the main server of the transmission picture, which is the main data of the test supervision of the present invention. do.

To this end, a picture of the candidate is captured and transmitted to the server according to a predetermined time or event using a webcam having a resolution capable of capturing an area including the table where the monitor is placed and the upper body of the candidate. In order to apply it to face recognition, it is necessary to obtain an accurate picture. Adjust the number of test questions that appear on your monitor to get a face-to-face picture, and create the next question icon on each page to encourage clicks or to capture a picture when you click on an answer so that you can stare at the monitor and capture an accurate picture. use. In addition, since there is an individual difference in the ability to solve the test questions, it also serves to distribute the server burden caused by the transmission of the captured picture. As the number of test takers increases, the resulting pictures are compressed and uploaded to the server to minimize server burden. In the preliminary preparation stage, the photographs captured by the candidate, the comparison pictures captured according to the time and event during the test, and the registration images of the self image submitted at the time of enrollment are quickly compared with the face recognition algorithm, so that the environment of the test and the face of the candidate can be compared. By judging, it is possible to judge cheating by evaluating the absence, pluralization, and testimonial of the candidate, and finally by using the reference material where the monitor is placed by the manager.

As a suggested method of ensuring authenticity of the candidate, using a webcam that captures a large area, including the table on which the monitor is placed and the upper body of the candidate, may result in the use of a person other than the candidate. It is possible to observe the use of a book or any reference material, so that the actual classroom is very similar to the test supervision situation (the shorter the capturing time, the more so it can be adjusted according to the computer's computing power and data processing speed). Can be reproduced. In addition, due to the characteristics of the face recognition algorithm, the recognition rate improves as the picture that is taken in front of the user improves the recognition rate, so that the candidate can adjust the number of test questions so that the candidate can watch the monitor to create the next item icon to click or to capture the picture when clicking the answer. The advantage is that you can get a picture. Based on the environment of the single photograph obtained in the test preparation confirmation stage, the recognition module is used to identify pluralization and absence and judge the identity of the candidate by face recognition to prevent cheating by preventing the wrongdoing of the candidate and the candidate. Will be removed to ensure your authenticity. In addition, face verification (1: 1, for example, access control or payment verification) or face identification (1: N, criminal search) is not complicated because it is a comparison between a specific person and a large number, but the present invention is only a self-photograph of a specific person. It is advantageous in terms of time required and recognition performance.

In the present invention, the test taker is tested without any psychological burden of being supervised without any special procedure, and the photo data stored in the main server is controlled only by the administrator, thereby eliminating the problem of personal privacy infringement.

This is an authentication step that requires the user to go through the test using the webcam activated by the execution of the program downloaded from the main server before the test. Determining the environmental evaluation and the sameness of pluralization, nonexistence, etc. by comparing the image stored in the main server with the unique number stored in the main server according to the unique number. Evaluating cheating, and storing the result of the administrator reviewing the suspected cheating picture determined by the computer once again.

Hereinafter, with reference to the accompanying drawings, preferred embodiments of the present invention will be described in detail. However, the spirit of the present invention is not limited to the embodiments presented, and those skilled in the art who understand the spirit of the present invention may easily implement other embodiments by adding, changing, deleting, and adding components within the scope of the same idea. It will be able to reproduce it, but it will also be said that it falls within the scope of the idea of the present invention.

1) Reference image registration step

At the beginning of the course registration or at the beginning of the semester, the candidate's digital image (picture, etc.) must be registered in the unique number and used as a reference image for future judgment.

2) Preliminary Steps

   The webcam should be placed in a location where the candidate's upper body and table area can be photographed and send and receive data to and from the computer. Candidates access the main server, enter their unique information (name, student number, etc.), and download a program that activates the webcam to their computer to capture, compress, and transmit photos. Alternatively, the program can be embedded into a webcam that will be connected to the candidate's computer. If the number of testers is small, the transmission file compression function and the decompression function for evaluating fraud may be disabled.

3) Test preparation confirmation step

   If you install the webcam activation program before the exam, a video of the webcam is displayed on the candidate's computer screen. When the candidate's torso and table area appear and the candidate is positioned in the center as much as possible, click the Send icon to adjust the webcam so that the picture is captured and sent to the main server before the test can be taken. When you transfer, the movie screen automatically disappears.

4) Save the comparison picture

   The moment when the picture is captured is determined by combining the time and event situation. Adjusting the number of questions will allow you to solve problems per screen and click on a photo capture event occurrence icon such as the "next question" icon. If you intentionally lengthen the time interval (by leaving a seat or not proceeding with the reference material) because the candidate captures the picture with the click of a candidate, the blind spot of the test supervision situation occurs and you cannot take a picture for a long time. This can be caused. To prevent this, the candidate will automatically capture the photo by firing an event at regular intervals based on a timer from the time the test is taken.

5) Selection of suspicious suspects by face recognition algorithm

   The test preparation step pictures stored according to the unique numbers and the comparison pictures and reference images captured during the test are extracted and simplified by the image and signal processing algorithms, and compared with the face recognition algorithm. Examination preparation stage Photo and comparison pictures to evaluate the pluralization of two or more people present in the picture (consulting the test, A in Fig. 2) or the absence of absence (no seat, B in Fig. 2) in the picture An environmental assessment is made, and the identity of the reference image and the comparative pictures (C, D in Figure 2) are evaluated to determine whether or not the registered person actually takes the test (substitute test) or cheating at the time of enrollment. When the feature data based on the basis vector is extracted and compared, the facial data is fast but the false rejection rate is high. Photo acquisition steps 3 and 4 are so important.

6) Judgment and Confirmation by Manager

   Since face recognition is more sensitive to the environment such as illumination and photographing direction than other biotechnology, the subject may be massed in order to determine the misconduct in real time due to its inaccurate accuracy. The list of suspected cheating judged by the face recognition algorithm can be more easily determined by the administrator comparing the actual pictures taken. In addition, it is determined at this stage whether or not to use the reference material on the table (on the B desk of Figure 2).

7) Save Results Step

   The result of step 6 is stored in the unique number and used by the questionnaire for future processing.

1 is an overall schematic diagram of a method for securing authenticity of an online test taker according to an exemplary embodiment of the present invention.

2 is an example related to a comparison result of an environment and a face obtained by a face recognition algorithm and the presence or absence of a reference book reviewed by a manager.

Claims (4)

A preliminary preparation step for the candidate to connect to the main server and download a program to the target computer to enable the webcam to capture, compress, and transmit the picture to the main server; A test preparation step of transmitting the photographs captured by the candidates to the main server to view the test; Comparing picture storage step of transmitting the captured picture in accordance with a certain interval and event conditions from the time after the candidate starts the test; A step of selecting a suspected misconduct to extract, simplify, and compare feature data by a face recognition algorithm; A determination and confirmation step of viewing the photographs directly by the manager of the photographs having a high probability of misconduct determined in the step; And a result storage step of storing the directly retrieved result of the step in a unique number on a use key for future test taker's processing. The method of claim 1, A method of authenticity that is simplified by embedding a program in the webcam that can capture, compress, and transfer pictures in the preliminary preparation step. The method of claim 1, Adjusting the number of test questions according to the event situation in the step of storing the comparison picture by combining the next question icon and the answer submission icon to induce clicks to capture a picture that accurately monitors the monitor and has the effect of real-time supervision. A webcam activation program storage unit for the candidate to access the server and to download and execute the program on the target computer; A reference image storage unit for storing a photograph (digital image) of a candidate submitted at the time of enrollment; A test preparation confirmation picture storage unit for storing a photograph captured and transmitted by the candidate immediately before the test after the webcam is activated; A comparison picture storage unit for storing a captured picture compressed and transmitted by combining a predetermined time and an event situation during a test; Compressed test prep photo, program to decompress comparison photos, A recognition module program storage unit for storing the program for extracting and simplifying feature data by decompressing photographs and reference images by image and signal processing algorithms and evaluating fraud by face recognition algorithms; And Through the above recognition module program, the authenticity securing system including a result storage unit for storing the result of reading the photo directly by the administrator to the possibility of cheating in a unique number in order to use in the future process.

Images