CN112613436B - Examination cheating detection method and device - Google Patents
Examination cheating detection method and device Download PDFInfo
- Publication number
- CN112613436B CN112613436B CN202011584152.2A CN202011584152A CN112613436B CN 112613436 B CN112613436 B CN 112613436B CN 202011584152 A CN202011584152 A CN 202011584152A CN 112613436 B CN112613436 B CN 112613436B
- Authority
- CN
- China
- Prior art keywords
- frame image
- examinee
- cheating
- human body
- video
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/20—Movements or behaviour, e.g. gesture recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- General Health & Medical Sciences (AREA)
- Human Computer Interaction (AREA)
- Computational Linguistics (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Multimedia (AREA)
- Social Psychology (AREA)
- Psychiatry (AREA)
- Apparatus For Radiation Diagnosis (AREA)
Abstract
The application discloses an examination cheating detection method and device, wherein the examination cheating detection method comprises the following steps: acquiring a video to be detected; judging whether a candidate with limb movement change exists between an ith frame image and an (i-1) th frame image of the video to be detected; under the condition that it is determined that a candidate with limb motion change exists between an ith frame image and an (i-1) th frame image, aiming at each candidate in part or all of the ith frame image, acquiring a two-dimensional test vector corresponding to the candidate, and determining whether the candidate is cheating according to the two-dimensional test vector corresponding to the candidate; wherein, two-dimensional examination taking vector includes: the one-dimensional limb vectors of the examinee in the ith frame image, the (i+m) th frame image, the (i+2m) th frame image, … … and the (i+mn-m) th frame image comprise: coordinates of key points of the upper body of the examinee, which are exposed above the tabletop; and adding (mn-m) to i, and continuing to execute the judging step. According to the embodiment of the application, the calculation efficiency is improved.
Description
Technical Field
The application relates to the field of image processing, in particular to an examination cheating detection method and device.
Background
In order to analyze the wind examination problem in the examination, such as cheating behaviors of examinees, invigorator and the like, after various examinations such as college exams, adult self-tests, academic level tests and the like are studied, examination video playback is required. The examination video playback task requires a significant amount of time and monetary costs, especially for large-scale examination. Therefore, a large examination video data analysis system is needed, so that behaviors of examinees can be automatically analyzed, and problems existing in the examination can be analyzed, so that the system plays a vital role in fairness and fairness of the examination on one hand, plays a deterrent role in cheating behaviors on the other hand, and the occurrence of cheating of the examination is reduced from the source.
The related examination cheating detection method has the problems of low calculation efficiency and low recognition accuracy.
Disclosure of Invention
The application provides an examination cheating detection method and device, which can improve the calculation efficiency.
The first aspect of the present application provides an examination cheating detection method, including:
acquiring a video to be detected;
judging whether a test taker with limb movement change exists between an ith frame image and an (i-1) th frame image of the video to be detected;
under the condition that it is determined that a candidate with limb motion change exists between an ith frame image and an (i-1) th frame image of the video to be detected, for each candidate in part or all of the candidate in the ith frame image, acquiring a two-dimensional test vector corresponding to the candidate, and determining whether the candidate is cheating according to the two-dimensional test vector corresponding to the candidate; wherein, the two-dimensional examination-taking vector comprises: the one-dimensional limb vector of the examinee in the ith frame image, the (i+m) th frame image, the (i+2m) th frame image, the … … and the (i+mn-m) th frame image comprises: coordinates of key points of the upper body of the human body above the table top of the examinee; i is an integer greater than or equal to 1, m is an integer greater than or equal to 1, and n is an integer greater than or equal to 2;
And adding (mn-m) to i, and continuously executing the step of judging whether a candidate with limb motion change exists between the ith frame image and the (i-1) th frame image of the video to be detected.
In some exemplary embodiments, in a case where it is determined that there is no examinee having a limb motion change between the i-th frame image and the (i-1) -th frame image of the video to be detected, the method further includes:
and adding 1 to i, and continuing to execute the step of judging whether a candidate with limb motion change exists between the ith frame image and the (i-1) th frame image of the video to be detected.
In some exemplary embodiments, before the determining whether there is a candidate with a change in limb motion between the i-th frame image and the (i-1) -th frame image of the video to be detected, the method further includes:
performing human body detection on an ith frame image in the video to be detected to obtain a human body detection result image of each examinee in the ith frame image, and performing human body detection on an (i-1) th frame image in the video to be detected to obtain the human body detection result image of each examinee in the (i-1) th frame image; wherein, the human body detection result graph comprises: the device comprises a minimum circumscribed rectangular frame for exposing the upper half of the human body above the desktop of the examinee;
The judging whether the examinee with limb movement change exists between the ith frame image and the (i-1) th frame image of the video to be detected comprises the following steps:
determining whether a subject with limb movement change exists between the ith frame image and the (i-1) th frame image according to the human body detection result images of all the subjects in the ith frame image and the human body detection result images of all the subjects in the (i-1) th frame image.
In some exemplary embodiments, the determining whether there is a limb movement change between the i-th frame image and the (i-1) -th frame image according to the human body detection result graphs of all the examinees in the i-th frame image and the human body detection result graphs of all the examinees in the (i-1) -th frame image includes:
calculating, for each of the examinees in the i-th frame image, a similarity between the human body detection result images of the same examinee in the i-th frame image and the (i-1) -th frame image;
under the condition that the similarity is larger than a first preset threshold value, confirming that the examinee has limb movement change;
and under the condition that the similarity is smaller than or equal to the first preset threshold value, confirming that the examinee has no limb movement change.
In some exemplary embodiments, the key points of the upper body of the human body include: the key points of the nose, the key points of the left eye, the key points of the right eye, the key points of the left ear, the key points of the right ear, the key points of the neck, the key points of the left shoulder, the key points of the right shoulder, the key points of the left elbow, the key points of the right elbow, the key points of the left hand and the key points of the right hand.
In some exemplary embodiments, the determining whether the examinee is cheating according to the two-dimensional examination vector corresponding to the examinee includes:
and determining whether the characterization of the examinee is cheating or not according to the two-dimensional examination vector corresponding to the examinee and the cheating confidence of the examinee.
In some exemplary embodiments, the determining whether the test taker is cheating based on the test taker's confidence in the cheating includes:
under the condition that the cheating confidence of the examinee is larger than a second preset threshold value, determining that the examinee is cheating;
and under the condition that the cheating confidence of the examinee is smaller than or equal to the second preset threshold value, determining that the examinee is not cheating.
A second aspect of the present application provides an examination cheating detection apparatus, including:
The video acquisition module is used for acquiring a video to be detected;
the limb motion change judging module is used for judging whether a candidate with limb motion change exists between an ith frame image and an (i-1) th frame image of the video to be detected;
the cheating determining module is used for acquiring a two-dimensional test taking vector corresponding to each of the examinees in part or all of the ith frame image under the condition that the examinees with limb action changes exist between the ith frame image and the (i-1) th frame image of the video to be detected, and determining whether the examinees are cheating according to the two-dimensional test taking vector corresponding to the examinees; wherein, the two-dimensional examination-taking vector comprises: the one-dimensional limb vector of the examinee in the ith frame image, the (i+m) th frame image, the (i+2m) th frame image, the … … and the (i+mn-m) th frame image comprises: coordinates of key points of the upper body of the human body above the table top of the examinee; i is an integer greater than or equal to 1, m is an integer greater than or equal to 1, and n is an integer greater than or equal to 2;
the limb movement change judging module is also used for: and adding (mn-m) to i, and continuously executing the step of judging whether a candidate with limb motion change exists between the ith frame image and the (i-1) th frame image of the video to be detected.
In some exemplary embodiments, the limb movement change determination module is further configured to:
and under the condition that it is determined that no examinee with limb movement change exists between the ith frame image and the (i-1) th frame image of the video to be detected, adding 1 to i, and continuously executing the step of judging whether the examinee with limb movement change exists between the ith frame image and the (i-1) th frame image of the video to be detected.
In some exemplary embodiments, the limb movement change determination module is further configured to:
performing human body detection on an ith frame image in the video to be detected to obtain a human body detection result image of each examinee in the ith frame image, and performing human body detection on an (i-1) th frame image in the video to be detected to obtain the human body detection result image of each examinee in the (i-1) th frame image; wherein, the human body detection result graph comprises: the device comprises a minimum circumscribed rectangular frame for exposing the upper half of the human body above the desktop of the examinee;
determining whether a subject with limb movement change exists between the ith frame image and the (i-1) th frame image according to the human body detection result images of all the subjects in the ith frame image and the human body detection result images of all the subjects in the (i-1) th frame image.
The application has the following advantages:
according to the embodiment of the application, a simple and rapid method is adopted to judge whether a candidate with limb motion change exists between the ith frame image and the (i-1) th frame image, and under the condition that the candidate with limb motion change exists between the ith frame image and the (i-1) th frame image, whether the candidate is cheating or not is further determined based on the two-dimensional response test vector corresponding to the candidate, namely, whether the candidate is cheating or not is detected by adopting a more accurate and slightly complex method, so that the calculation efficiency is improved.
In some exemplary embodiments, under the condition that it is determined that no candidate with variation of limb actions exists between the ith frame image and the (i-1) th frame image of the video to be detected, it is considered that all candidates in the ith frame image are not cheating, that is, whether the candidate is cheating is further determined based on the two-dimensional test vectors corresponding to the candidate, so that most of pictures without cheating are filtered, and calculation efficiency is improved.
In some exemplary embodiments, a human body detection model is trained based on a first video of a real examination scene, human body detection is performed on an ith frame image based on the trained human body detection model to obtain a human body detection result diagram of each examinee in the ith frame image, human body detection is performed on an (i-1) th frame image in the video to be detected to obtain the human body detection result diagram of each examinee in the (i-1) th frame image, and whether an examinee with limb movement change exists is further judged.
In some exemplary embodiments, the cheating classifier is trained based on the second video of the real examination scene, and the cheating confidence of the examinee is obtained based on the trained cheating classifier, so that whether the examinee is cheating or not is judged, and the recognition rate of various environments and complex scenes can be ensured because the cheating classifier is obtained based on the second video training of the real examination scene.
Drawings
The accompanying drawings are included to provide a further understanding of the application and are incorporated in and constitute a part of this specification, illustrate the application and, together with the description, do not limit the application.
FIG. 1 is a flowchart of an examination cheating detection method according to an embodiment of the present application;
FIG. 2 is a schematic diagram of key points of the upper body of the human body according to an embodiment of the present application;
FIG. 3 is a flowchart of an examination cheating detection method provided by an example of an embodiment of the present application;
fig. 4 is a block diagram of an examination cheating detection device according to another embodiment of the present application.
In the drawings:
1: key point 2 of right hand: key point of left hand
3: key point 4 of right elbow: key point of left elbow
5: key point 6 of right shoulder: key point of left shoulder
7: key point 8 of neck: key point of right ear
9: key point 10 of left ear: key point of nose
11: key point 12 for right eye: key point of left eye
Detailed Description
The following detailed description of specific embodiments of the present application refers to the accompanying drawings. It should be understood that the detailed description is presented herein for purposes of illustration and explanation only and is not intended to limit the present application.
As used in this disclosure, the term "and/or" includes any and all combinations of one or more of the associated listed items.
The terminology used in the present disclosure is for the purpose of describing particular embodiments only and is not intended to be limiting of the disclosure. As used in this disclosure, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.
When the terms "comprises," "comprising," and/or "including" are used in this disclosure, they specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.
Embodiments of the present disclosure may be described with reference to plan and/or cross-sectional views with the aid of idealized schematic diagrams of the present disclosure. Accordingly, the example illustrations may be modified in accordance with manufacturing techniques and/or tolerances.
Unless otherwise defined, all terms (including technical and scientific terms) used in this disclosure have the same meaning as commonly understood by one of ordinary skill in the art. It will be further understood that terms, such as those defined in commonly used dictionaries, should be interpreted as having a meaning that is consistent with their meaning in the context of the relevant art and the present disclosure, and will not be interpreted in an idealized or overly formal sense unless expressly so defined herein.
Fig. 1 is a flowchart of an examination cheating detection method according to an embodiment of the present application.
As shown in fig. 1, one embodiment of the present application proposes an examination cheating detection method, including:
step 100, obtaining a video to be detected.
And step 101, judging whether a candidate with limb movement change exists between the ith frame image and the (i-1) th frame image of the video to be detected.
In some exemplary embodiments, before the determining whether there is a candidate with a change in limb motion between the i-th frame image and the (i-1) -th frame image of the video to be detected, the method further includes:
performing human body detection on an ith frame image in the video to be detected to obtain a human body detection result image of each examinee in the ith frame image, and performing human body detection on an (i-1) th frame image in the video to be detected to obtain the human body detection result image of each examinee in the (i-1) th frame image; wherein, the human body detection result graph comprises: the device comprises a minimum circumscribed rectangular frame for exposing the upper half of the human body above the desktop of the examinee;
The judging whether the examinee with limb movement change exists between the ith frame image and the (i-1) th frame image of the video to be detected comprises the following steps:
determining whether a subject with limb movement change exists between the ith frame image and the (i-1) th frame image according to the human body detection result images of all the subjects in the ith frame image and the human body detection result images of all the subjects in the (i-1) th frame image.
In some exemplary embodiments, a pre-trained human body detection model may be used to perform human body detection on an i-th frame image to obtain a human body detection result image of each examinee in the i-th frame image, and perform human body detection on an (i-1) -th frame image in the video to be detected to obtain the human body detection result image of each examinee in the (i-1) -th frame image.
In some exemplary embodiments, the human detection model may be a human detection network or a human detector.
In some exemplary embodiments, human detection networks include, but are not limited to, YOLOv3, YOLOv4, single-step multi-frame object detection (SSD, single Short MultiBox Detector), fast-return convolutional neural network (fast-RCNN, faster Recurrent Convolutional Neural Network), support vector machine (SVM, support Vector Machine), and the like detection networks.
In some exemplary embodiments, the human body detector includes, but is not limited to, a detector of Yolov3, yolov4, SSD, faster-RCNN, SVM, and the like.
In some exemplary embodiments, the human detection model may be trained using the following methods:
acquiring a large number of first videos of real examination scenes, and manually marking the first videos, wherein a marking target is a minimum external rectangular frame comprising the upper body of each examinee exposed above a desktop; and training a human body detection model according to the marked first video.
In some exemplary embodiments, exposing the upper body of the subject above the table top comprises: human head, upper limbs, hands.
In some exemplary embodiments, in the process of training the human body detection model, a frame of image of the first video of the real examination scene is used as input of the human body detection model, and a minimum circumscribed rectangular frame corresponding to each examinee in the frame of image is used as output of the human body detection model, and the minimum circumscribed rectangular frame can also be called a human body detection result diagram.
In some exemplary embodiments, the human detection model is optimized and accelerated by adopting a TensorRT, NCNN method and the like in the process of training the human detection model.
In some exemplary embodiments, the determining whether there is a limb movement change between the i-th frame image and the (i-1) -th frame image according to the human body detection result graphs of all the examinees in the i-th frame image and the human body detection result graphs of all the examinees in the (i-1) -th frame image includes:
calculating, for each of the examinees in the i-th frame image, a similarity between the human body detection result images of the same examinee in the i-th frame image and the (i-1) -th frame image;
under the condition that the similarity is larger than a first preset threshold value, confirming that the examinee has limb movement change;
and under the condition that the similarity is smaller than or equal to the first preset threshold value, confirming that the examinee has no limb movement change.
In some exemplary embodiments, the similarity may be calculated in a variety of ways, and the specific calculation method is not used to limit the protection scope of the embodiments of the present application. For example, the similarity is calculated by calculating the average value of the sums of pixel differences corresponding to all pixels of the two human body detection result maps.
In some exemplary embodiments, the first preset threshold T1 may be empirically set, with larger T1 being more video pictures filtered out and vice versa.
In some exemplary embodiments, since the success rate of the cheating identification is very difficult to reach 100%, in practical use, a part of non-cheating identification is allowed to be mistakenly identified as cheating, and then the first preset threshold T1 is slightly smaller than the first preset threshold T1 through manual confirmation.
102, under the condition that it is determined that a candidate with limb motion change exists between an ith frame image and an (i-1) th frame image of the video to be detected, for each of a part or all of the candidates in the ith frame image, acquiring a two-dimensional test vector corresponding to the candidate, and determining whether the candidate is cheating according to the two-dimensional test vector corresponding to the candidate; wherein, the two-dimensional examination-taking vector comprises: the one-dimensional limb vector of the examinee in the ith frame image, the (i+m) th frame image, the (i+2m) th frame image, the … … and the (i+mn-m) th frame image comprises: coordinates of key points of the upper body of the human body above the table top of the examinee; i is an integer greater than or equal to 1, m is an integer greater than or equal to 1, and n is an integer greater than or equal to 2.
In this embodiment of the present application, for each of the examinees having a limb motion change in the ith frame image, a two-dimensional test taking vector corresponding to the examinee may be obtained, and whether the examinee is cheating or not may be determined according to the two-dimensional test taking vector corresponding to the examinee;
And the two-dimensional examination vectors corresponding to the examinees can be acquired for each examinee in the ith frame of image, and whether the examinees cheat or not is determined according to the two-dimensional examination vectors corresponding to the examinees.
Fig. 2 is a schematic diagram of key points of the upper body of the human body in the embodiment of the present application. As shown in fig. 2, in some exemplary embodiments, the key points of the upper body of the human body include: the key points of the nose, the key points of the left eye, the key points of the right eye, the key points of the left ear, the key points of the right ear, the key points of the neck, the key points of the left shoulder, the key points of the right shoulder, the key points of the left elbow, the key points of the right elbow, the key points of the left hand and the key points of the right hand.
Then the firstThe examinee's one-dimensional limb vector in the i-frame image can be expressed as ((x) 0,i ,y 0,i ) (x 1,i ,y 1,i )…(x 11,i ,y 11,i ) A) is provided; wherein, (x) j，i ，y j，i ) The coordinates of the j-th key point of the examinee in the i-th frame image are represented, j=0, 1,2, …,11.
The two-dimensional examination vector corresponding to the examinee can be expressed as:
in some exemplary embodiments, the determining whether the examinee is cheating according to the two-dimensional examination vector corresponding to the examinee includes:
and determining whether the characterization of the examinee is cheating or not according to the two-dimensional examination vector corresponding to the examinee and the cheating confidence of the examinee.
In some exemplary embodiments, the value of the cheating confidence is between 0 and 1, and the cheating confidence represents the probability of whether the examinee is cheating.
In some exemplary embodiments, the determining whether the test taker is cheating based on the test taker's confidence in the cheating includes:
under the condition that the cheating confidence of the examinee is larger than a second preset threshold value, determining that the examinee is cheating;
and under the condition that the cheating confidence of the examinee is smaller than or equal to the second preset threshold value, determining that the examinee is not cheating.
In some exemplary embodiments, a pre-trained cheating classifier may be employed to determine the cheating confidence of the test taker. Namely, the two-dimensional test vectors corresponding to the test taker are used as the input of the cheating classifier to obtain the cheating confidence of the test taker.
In some exemplary embodiments, the cheating classifier may be trained in the following manner:
acquiring a large number of second videos of real examination scenes, and manually marking the second videos, wherein a marking target is information of whether each examinee is cheating or not in continuous mn frame images;
acquiring coordinates of key points of the upper body of the human body above the desktop of each examinee in an ith frame image in a second video, and obtaining one-dimensional limb vectors of each examinee in the ith frame image; the same method is adopted to obtain one-dimensional limb vectors of each examinee in the (i+m) th frame image, one-dimensional limb vectors of each examinee in the (i+2m) th frame image, … … and one-dimensional limb vectors of each examinee in the (i+mn-m) th frame image; forming a two-dimensional test taking vector corresponding to the test taker by using one-dimensional limb vector of the same test taker in the ith frame image, the (i+m) th frame image, the (i+2m) th frame image, the … … th (i+mn-m) th frame image;
Aiming at each examinee, taking the two-dimensional test vectors corresponding to the examinee as the input of the cheating classifier, and taking the information of whether the examinee is cheating in continuous mn frame images as the output of the cheating classifier; training the cheating classifier according to the two-dimensional examination vectors corresponding to all the examinees and the information representing whether the cheating is performed.
In some example embodiments, the cheating classifier may be a classifier trained based on a convolutional neural network (CNN, convolutional Neural Network), VGG (Visual Geometry Group Network), resnet, resNeXt, mobileNetV3, or the like.
In some exemplary embodiments, the second preset threshold T2 may be empirically set, with the larger T2 being more likely to miss a cheating segment, and conversely the more likely to occur to misinterpret a non-cheating segment as a cheating segment.
In some exemplary embodiments, since the success rate of the cheating identification is very difficult to reach 100%, in practical use, a part of non-cheating identification is allowed to be mistakenly identified as cheating, and then the user confirms the operation once, so that the second preset threshold T2 can be selected to be slightly smaller.
And step 103, adding (mn-m) to the i, and continuing to execute the step of judging whether a candidate with limb motion change exists between the i-th frame image and the (i-1) -th frame image of the video to be detected.
In some exemplary embodiments, in a case where it is determined that there is no examinee having a limb motion change between the i-th frame image and the (i-1) -th frame image of the video to be detected, the method further includes:
and adding 1 to i, and continuing to execute the step of judging whether a candidate with limb motion change exists between the ith frame image and the (i-1) th frame image of the video to be detected.
According to the embodiment of the application, a simple and rapid method is adopted to judge whether a candidate with limb motion change exists between the ith frame image and the (i-1) th frame image, and under the condition that the candidate with limb motion change exists between the ith frame image and the (i-1) th frame image, whether the candidate is cheating or not is further determined based on the two-dimensional response test vector corresponding to the candidate, namely, whether the candidate is cheating or not is detected by adopting a more accurate and slightly complex method, so that the calculation efficiency is improved.
In some exemplary embodiments, under the condition that it is determined that no candidate with variation of limb actions exists between the ith frame image and the (i-1) th frame image of the video to be detected, it is considered that all candidates in the ith frame image are not cheating, that is, whether the candidate is cheating is further determined based on the two-dimensional test vectors corresponding to the candidate, so that most of pictures without cheating are filtered, and calculation efficiency is improved.
In some exemplary embodiments, a human body detection model is trained based on a first video of a real examination scene, human body detection is performed on an ith frame image based on the trained human body detection model to obtain a human body detection result diagram of each examinee in the ith frame image, human body detection is performed on an (i-1) th frame image in the video to be detected to obtain the human body detection result diagram of each examinee in the (i-1) th frame image, and whether an examinee with limb movement change exists is further judged.
In some exemplary embodiments, the cheating classifier is trained based on the second video of the real examination scene, and the cheating confidence of the examinee is obtained based on the trained cheating classifier, so that whether the examinee is cheating or not is judged, and the recognition rate of various environments and complex scenes can be ensured because the cheating classifier is obtained based on the second video training of the real examination scene.
The following details a specific implementation of the embodiments of the present application by way of an example, which is presented for convenience of illustration only and is not intended to limit the scope of protection of the embodiments of the present application.
Example
This example describes an examination cheating detection method, and fig. 3 is a flowchart of an examination cheating detection method provided in an example of an embodiment of the present application. As shown in fig. 3, the method includes:
step 1, acquiring a large number of first videos of real examination scenes, manually marking the first videos, wherein the marking target comprises a minimum external rectangular frame of each examinee exposing the upper body of the human body above a desktop, and each examinee exposing the upper body of the human body above the desktop comprises: human head, upper limbs, hands.
In this example, when the manual annotation is performed, the manual annotation may be performed for each frame of image in the first video, or may be performed at intervals of a certain number of frames.
In this example, only the upper body of the person is considered, because there is shielding of the table.
And step 2, training a human body detection model according to the marked first video.
In this example, the human detection model may be a human detection network or a human detector.
In this example, the human detection network includes, but is not limited to, YOLOv3, YOLOv4, single-step multi-frame object detection (SSD, single Short MultiBox Detector), fast-homing convolutional neural network (fast-RCNN, faster Recurrent Convolutional Neural Network), support vector machine (SVM, support Vector Machine), and the like detection networks.
In this example, the human body detector includes, but is not limited to, a detector of Yolov3, yolov4, SSD, faster-RCNN, SVM, etc.
In this example, in the process of training the human body detection model, a frame of image of the first video of the real examination scene is used as input of the human body detection model, and a minimum circumscribed rectangular frame corresponding to each examinee in the frame of image is used as output of the human body detection model, and the minimum circumscribed rectangular frame can also be called as a human body detection result diagram.
In this example, in the process of training the human body detection model, the human body detection model is optimized and accelerated by adopting a TensorRT, NCNN method and the like.
And 3, acquiring a large number of second videos of real examination scenes, and manually marking the second videos, wherein the marking target is information of whether each examinee is cheating or not in continuous mn frame images (for example, 1 indicates that the examinee is cheating, and 0 indicates that the examinee is not cheating).
In this example, the first video and the second video may be the same video or different videos.
In this example, m is an integer greater than or equal to 1, and n is an integer greater than or equal to 2.
Step 4, acquiring coordinates of key points of each examinee in the ith frame of image, which are exposed out of the upper body of the human body above the desktop, in the second video, and obtaining one-dimensional limb vectors of each examinee in the ith frame of image; the same method is adopted to obtain one-dimensional limb vectors of each examinee in the (i+m) th frame image, one-dimensional limb vectors of each examinee in the (i+2m) th frame image, … … and one-dimensional limb vectors of each examinee in the (i+mn-m) th frame image; and forming a two-dimensional test taking vector corresponding to the test taker by using one-dimensional limb vector of the same test taker in the ith frame image, the (i+m) th frame image, the (i+2m) th frame image, the … … th (i+mn-m) th frame image.
In this example, i is an integer greater than or equal to 1.
In this example, key points of the upper body of the human body include: a key point of the nose (e.g., a point corresponding to 10 in fig. 2), a key point of the left eye (e.g., a point corresponding to 12 in fig. 2), a key point of the right eye (e.g., a point corresponding to 11 in fig. 2), a key point of the left ear (e.g., a point corresponding to 9 in fig. 2), a key point of the right ear (e.g., a point corresponding to 8 in fig. 2), a key point of the neck (e.g., a point corresponding to 7 in fig. 2), a key point of the left shoulder (e.g., a point corresponding to 6 in fig. 2), a key point of the right shoulder (e.g., a point corresponding to 5 in fig. 2), a key point of the left elbow (e.g., a point corresponding to 4 in fig. 2), a key point of the right elbow (e.g., a point corresponding to 3 in fig. 2), a key point of the left hand (e.g., a point corresponding to 2 in fig. 2), and a key point of the right hand (e.g., a point corresponding to 1 in fig. 2).
In this example, the one-dimensional limb vector of the examinee in the i-th frame image may be expressed as ((x) 0,i ,y 0,i ) (x 1,i ,y 1,i )…(x 11,i ,y 11,i ) A) is provided; wherein, (x) j，i ，y j，i ) The coordinates of the j-th key point of the examinee in the i-th frame image are represented, j=0, 1,2, …,11.
The two-dimensional examination vector corresponding to the examinee can be expressed as:
step 5, aiming at each examinee, taking the two-dimensional test vectors corresponding to the examinee as the input of the cheating classifier, and taking the information of whether the examinee is cheating in continuous mn frame images as the output of the cheating classifier; training the cheating classifier according to the two-dimensional examination vectors corresponding to all the examinees and the information representing whether the cheating is performed.
In this example, the cheating classifier may be a classifier trained on a convolutional neural network (CNN, convolutional Neural Network), VGG (Visual Geometry Group Network), resnet, resNeXt, mobileNetV3, or the like.
And 6, acquiring a video to be detected.
And 7, performing human body detection on the ith frame image in the video to be detected according to the trained human body detection model to obtain a human body detection result image (namely the minimum circumscribed rectangular frame) of each examinee in the ith frame image, and performing human body detection on the (i-1) th frame image in the video to be detected according to the trained human body detection model to obtain a human body detection result image (namely the minimum circumscribed rectangular frame) of each examinee in the (i-1) th frame image.
Step 8, judging whether a human body candidate with limb movement change exists between the ith frame image and the (i-1) th frame image according to the human body detection result image of each human body candidate in the ith frame image and the human body detection result image of each human body candidate in the (i-1) th frame image; if there is a candidate with a change in limb movement between the i-th frame image and the (i-1) -th frame image, continuing to execute the step 9; and (3) when no examinee with limb movement change exists between the ith frame image and the (i-1) th frame image, adding 1 to i, and continuing to execute the step (7).
In this example, for each examinee in the i-th frame image, the similarity between the human body detection result graphs of the same examinee in the i-th frame image and the (i-1) -th frame image is calculated; under the condition that the similarity is larger than a first preset threshold (T1), confirming that the examinee has limb movement change; and under the condition that the similarity is smaller than or equal to a first preset threshold (T1), confirming that the examinee has no limb movement change.
In this example, the similarity may be calculated in a variety of manners, and the specific calculation manner is not used to limit the protection scope of the embodiments of the present application. For example, the similarity is calculated by calculating the average value of the sums of pixel differences corresponding to all pixels of the two human body detection result maps.
In this example, the first preset threshold T1 may be empirically set, with larger T1 being more video pictures filtered out and vice versa being less.
Step 9, acquiring coordinates of key points of each examinee in the ith frame of image in the video to be detected, wherein the key points are exposed out of the upper body of the human body above the desktop, so as to obtain one-dimensional limb vectors of each examinee in the ith frame of image; the same method is adopted to obtain one-dimensional limb vectors of each examinee in the (i+m) th frame image, one-dimensional limb vectors of each examinee in the (i+2m) th frame image, … … and one-dimensional limb vectors of each examinee in the (i+mn-m) th frame image; and forming a two-dimensional test taking vector corresponding to the test taker by using one-dimensional limb vector of the same test taker in the ith frame image, the (i+m) th frame image, the (i+2m) th frame image, the … … th (i+mn-m) th frame image.
In this example, m is an integer greater than or equal to 1, and n is an integer greater than or equal to 2.
Step 10, aiming at each examinee, taking a two-dimensional test vector corresponding to the examinee as input of a cheating classifier to obtain the cheating confidence coefficient of whether the characterization of the examinee is cheating; judging whether the cheating confidence of the examinee is larger than a second preset threshold (T2), and determining that the examinee is cheating under the condition that the cheating confidence of the examinee is larger than the second preset threshold (T2); and under the condition that the cheating confidence of the examinee is smaller than or equal to a second preset threshold (T2), determining that the examinee is not cheating.
In this example, the second preset threshold T2 may be empirically set, the larger T2 the more likely that a cheating segment will be missed, and conversely the more likely that a non-cheating segment will be mistaken for a cheating segment will occur.
Step 11, adding (mn-m) to i, and continuing to execute step 7.
The above steps of the methods are divided, for clarity of description, and may be combined into one step or split into multiple steps when implemented, so long as they include the same logic relationship, and they are all within the protection scope of this patent; it is within the scope of this patent to add insignificant modifications to the algorithm or flow or introduce insignificant designs, but not to alter the core design of its algorithm and flow.
Fig. 4 is a block diagram of an examination cheating detection device according to another embodiment of the present application.
As shown in fig. 4, another embodiment of the present application proposes an examination cheating detection device, including:
a video acquisition module 401, configured to acquire a video to be detected;
a limb motion change judging module 402, configured to judge whether a candidate with a limb motion change exists between an i-th frame image and an (i-1) -th frame image of the video to be detected;
the cheating determining module 403, configured to, when determining that there is a candidate with a change in limb motion between an i-th frame image and an (i-1) -th frame image of the video to be detected, obtain, for each of some or all of the candidates in the i-th frame image, a two-dimensional test vector corresponding to the candidate, and determine whether the candidate is cheating according to the two-dimensional test vector corresponding to the candidate; wherein, the two-dimensional examination-taking vector comprises: the one-dimensional limb vector of the examinee in the ith frame image, the (i+m) th frame image, the (i+2m) th frame image, the … … and the (i+mn-m) th frame image comprises: coordinates of key points of the upper body of the human body above the table top of the examinee; i is an integer greater than or equal to 1, m is an integer greater than or equal to 1, and n is an integer greater than or equal to 2;
The limb movement change judging module 402 is further configured to: and adding (mn-m) to i, and continuously executing the step of judging whether a candidate with limb motion change exists between the ith frame image and the (i-1) th frame image of the video to be detected.
In some exemplary embodiments, the limb movement change determination module 402 is further configured to:
and under the condition that it is determined that no examinee with limb movement change exists between the ith frame image and the (i-1) th frame image of the video to be detected, adding 1 to i, and continuously executing the step of judging whether the examinee with limb movement change exists between the ith frame image and the (i-1) th frame image of the video to be detected.
In some exemplary embodiments, the limb movement change determination module 402 is further configured to:
performing human body detection on an ith frame image in the video to be detected to obtain a human body detection result image of each examinee in the ith frame image, and performing human body detection on an (i-1) th frame image in the video to be detected to obtain the human body detection result image of each examinee in the (i-1) th frame image; wherein, the human body detection result graph comprises: the device comprises a minimum circumscribed rectangular frame for exposing the upper half of the human body above the desktop of the examinee;
Determining whether a subject with limb movement change exists between the ith frame image and the (i-1) th frame image according to the human body detection result images of all the subjects in the ith frame image and the human body detection result images of all the subjects in the (i-1) th frame image.
In some exemplary embodiments, the limb movement change determining module 402 is specifically configured to implement the method according to the human body detection result graphs of all the examinees in the i-th frame image and the human body detection result graphs of all the examinees in the (i-1) -th frame image, and determine whether there is a limb movement change between the i-th frame image and the (i-1) -th frame image by:
calculating, for each of the examinees in the i-th frame image, a similarity between the human body detection result images of the same examinee in the i-th frame image and the (i-1) -th frame image;
under the condition that the similarity is larger than a first preset threshold value, confirming that the examinee has limb movement change;
and under the condition that the similarity is smaller than or equal to the first preset threshold value, confirming that the examinee has no limb movement change.
In some exemplary embodiments, the key points of the upper body of the human body include: the key points of the nose, the key points of the left eye, the key points of the right eye, the key points of the left ear, the key points of the right ear, the key points of the neck, the key points of the left shoulder, the key points of the right shoulder, the key points of the left elbow, the key points of the right elbow, the key points of the left hand and the key points of the right hand.
In some exemplary embodiments, the cheating determining module 403 is specifically configured to implement the determining whether the examinee is cheating according to the two-dimensional examination vector corresponding to the examinee in the following manner:
and determining whether the characterization of the examinee is cheating or not according to the two-dimensional examination vector corresponding to the examinee and the cheating confidence of the examinee.
In some exemplary embodiments, the cheating determining module 403 is specifically configured to implement the determining whether the candidate is cheating according to the cheating confidence of the candidate in the following manner:
under the condition that the cheating confidence of the examinee is larger than a second preset threshold value, determining that the examinee is cheating;
and under the condition that the cheating confidence of the examinee is smaller than or equal to the second preset threshold value, determining that the examinee is not cheating.
The specific implementation process of the examination cheating detection device is the same as that of the examination cheating detection method in the foregoing embodiment, and will not be repeated here.
In this embodiment, each module is a logic module, and in practical application, one logic unit may be one physical unit, or may be a part of one physical unit, or may be implemented by a combination of a plurality of physical units. In addition, in order to highlight the innovative part of the present application, elements that are not so close to solving the technical problem presented in the present application are not introduced in the present embodiment, but it does not indicate that other elements are not present in the present embodiment.
The embodiment also provides an electronic device including one or more processors; the storage device stores one or more programs, and when the one or more programs are executed by the one or more processors, the one or more processors implement the test cheating detection method provided in this embodiment, so that specific steps of the test cheating detection method are not repeated in order to avoid repeated description.
The embodiment also provides a computer readable medium, on which a computer program is stored, where the program when executed by a processor implements the test cheating detection method provided in the embodiment, and specific steps of the test cheating detection method are not repeated here to avoid repetitive description.
Those of ordinary skill in the art will appreciate that all or some of the steps, systems, functional modules/units in the apparatus, and methods disclosed above may be implemented as software, firmware, hardware, and suitable combinations thereof. In a hardware implementation, the division between the functional modules/units mentioned in the above description does not necessarily correspond to the division of physical components; for example, one physical component may have multiple functions, or one function or step may be performed cooperatively by several physical components. Some or all of the physical components may be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application specific integrated circuit. Such software may be distributed on computer readable media, which may include computer storage media (or non-transitory media) and communication media (or transitory media). The term computer storage media includes both volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data, as known to those skilled in the art. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital Versatile Disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium which can be used to store the desired information and which can be accessed by a computer. Furthermore, as is well known to those of ordinary skill in the art, communication media typically embodies computer readable instructions, data structures, program modules or other data in a modulated data signal such as a carrier wave or other transport mechanism and includes any information delivery media.
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
Those skilled in the art will appreciate that while some embodiments described herein include some features but not others included in other embodiments, combinations of features of different embodiments are meant to be within the scope of the embodiments and form different embodiments.
It is to be understood that the above embodiments are merely illustrative of the exemplary embodiments employed to illustrate the principles of the present application, however, the present application is not limited thereto. Various modifications and improvements may be made by those skilled in the art without departing from the spirit and substance of the application, and are also considered to be within the scope of the application.
Claims (7)
1. An examination cheating detection method, comprising:
acquiring a video to be detected;
judging whether a test taker with limb movement change exists between an ith frame image and an (i-1) th frame image of the video to be detected;
under the condition that it is determined that a candidate with limb motion change exists between an ith frame image and an (i-1) th frame image of the video to be detected, for each candidate in part or all of the candidate in the ith frame image, acquiring a two-dimensional test vector corresponding to the candidate, and determining whether the candidate is cheating according to the two-dimensional test vector corresponding to the candidate; wherein, the two-dimensional examination-taking vector comprises: the one-dimensional limb vector of the examinee in the ith frame image, the (i+m) th frame image, the (i+2m) th frame image, the … … and the (i+mn-m) th frame image comprises: coordinates of key points of the upper body of the human body above the table top of the examinee; i is an integer greater than or equal to 1, m is an integer greater than or equal to 1, and n is an integer greater than or equal to 2;
adding (mn-m) to i, and continuously executing the step of judging whether a candidate with limb motion change exists between the ith frame image and the (i-1) th frame image of the video to be detected;
Performing human body detection on an ith frame image in the video to be detected to obtain a human body detection result image of each examinee in the ith frame image, and performing human body detection on an (i-1) th frame image in the video to be detected to obtain the human body detection result image of each examinee in the (i-1) th frame image; wherein, the human body detection result graph comprises: the device comprises a minimum circumscribed rectangular frame for exposing the upper half of the human body above the desktop of the examinee;
the judging whether a candidate with limb movement change exists between the ith frame image and the (i-1) th frame image of the video to be detected comprises the following steps:
determining whether a subject with limb movement change exists between the ith frame image and the (i-1) th frame image according to the human body detection result images of all the subjects in the ith frame image and the human body detection result images of all the subjects in the (i-1) th frame image;
the determining whether there is a limb movement change between the i-th frame image and the (i-1) -th frame image according to the human body detection result graphs of all the examinees in the i-th frame image and the human body detection result graphs of all the examinees in the (i-1) -th frame image includes:
Calculating, for each of the examinees in the i-th frame image, a similarity between the human body detection result images of the same examinee in the i-th frame image and the (i-1) -th frame image;
under the condition that the similarity is larger than a first preset threshold value, confirming that the examinee has limb movement change;
and under the condition that the similarity is smaller than or equal to the first preset threshold value, confirming that the examinee has no limb movement change.
2. The examination cheating detection method according to claim 1, in the case where it is determined that there is no examinee whose limb movement is changed between the i-th frame image and the (i-1) -th frame image of the video to be detected, the method further comprising:
and adding 1 to i, and continuing to execute the step of judging whether a candidate with limb motion change exists between the ith frame image and the (i-1) th frame image of the video to be detected.
3. The examination cheating detection method according to any one of claims 1-2, wherein the key points of the upper body of the human body include: the key points of the nose, the key points of the left eye, the key points of the right eye, the key points of the left ear, the key points of the right ear, the key points of the neck, the key points of the left shoulder, the key points of the right shoulder, the key points of the left elbow, the key points of the right elbow, the key points of the left hand and the key points of the right hand.
4. The examination cheating detection method according to any one of claims 1-2, wherein the determining whether the examinee is cheating according to the two-dimensional examination amount corresponding to the examinee comprises:
and determining whether the characterization of the examinee is cheating or not according to the two-dimensional examination vector corresponding to the examinee and the cheating confidence of the examinee.
5. The examination cheating detection method according to claim 4, wherein the determining whether the test taker is cheating according to the test taker's confidence level of cheating comprises:
under the condition that the cheating confidence of the examinee is larger than a second preset threshold value, determining that the examinee is cheating;
and under the condition that the cheating confidence of the examinee is smaller than or equal to the second preset threshold value, determining that the examinee is not cheating.
6. An examination cheating detection device, comprising:
the video acquisition module is used for acquiring a video to be detected;
the limb motion change judging module is also used for judging whether a candidate with limb motion change exists between an ith frame image and an (i-1) th frame image of the video to be detected;
the cheating determining module is used for acquiring a two-dimensional test taking vector corresponding to each of the examinees in part or all of the ith frame image under the condition that the examinees with limb action changes exist between the ith frame image and the (i-1) th frame image of the video to be detected, and determining whether the examinees are cheating according to the two-dimensional test taking vector corresponding to the examinees; wherein, the two-dimensional examination-taking vector comprises: the one-dimensional limb vector of the examinee in the ith frame image, the (i+m) th frame image, the (i+2m) th frame image, the … … and the (i+mn-m) th frame image comprises: coordinates of key points of the upper body of the human body above the table top of the examinee; i is an integer greater than or equal to 1, m is an integer greater than or equal to 1, and n is an integer greater than or equal to 2;
The limb movement change judging module is also used for: adding (mn-m) to i, and continuously executing the step of judging whether a candidate with limb motion change exists between the ith frame image and the (i-1) th frame image of the video to be detected;
the limb motion change judging module is further configured to perform human body detection on an ith frame image in the video to be detected to obtain a human body detection result diagram of each examinee in the ith frame image, and perform human body detection on an (i-1) th frame image in the video to be detected to obtain the human body detection result diagram of each examinee in the (i-1) th frame image; wherein, the human body detection result graph comprises: the device comprises a minimum circumscribed rectangular frame for exposing the upper half of the human body above the desktop of the examinee; determining whether a subject with limb movement change exists between the ith frame image and the (i-1) th frame image according to the human body detection result images of all the subjects in the ith frame image and the human body detection result images of all the subjects in the (i-1) th frame image; the limb movement change judging module is further configured to calculate, for each of the examinees in the i-th frame image, a similarity between the human body detection result graphs of the same examinee in the i-th frame image and the (i-1) -th frame image; under the condition that the similarity is larger than a first preset threshold value, confirming that the examinee has limb movement change; and under the condition that the similarity is smaller than or equal to the first preset threshold value, confirming that the examinee has no limb movement change.
7. The examination cheating detection device according to claim 6, wherein the limb movement change judgment module is further configured to:
and under the condition that it is determined that no examinee with limb movement change exists between the ith frame image and the (i-1) th frame image of the video to be detected, adding 1 to i, and continuously executing the step of judging whether the examinee with limb movement change exists between the ith frame image and the (i-1) th frame image of the video to be detected.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011584152.2A CN112613436B (en) | 2020-12-28 | 2020-12-28 | Examination cheating detection method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011584152.2A CN112613436B (en) | 2020-12-28 | 2020-12-28 | Examination cheating detection method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112613436A CN112613436A (en) | 2021-04-06 |
| CN112613436B true CN112613436B (en) | 2023-08-04 |
Family
ID=75248412
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011584152.2A Active CN112613436B (en) | 2020-12-28 | 2020-12-28 | Examination cheating detection method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112613436B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114842271B (en) * | 2022-06-01 | 2024-07-12 | 湖南大学 | Method, device, computer equipment and storage medium for classifying invigilation video |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106203363A (en) * | 2016-07-15 | 2016-12-07 | 中国科学院自动化研究所 | Human skeleton motion sequence Activity recognition method |
| WO2017152425A1 (en) * | 2016-03-11 | 2017-09-14 | 深圳市大疆创新科技有限公司 | Method, system and device for preventing cheating in network exam, and storage medium |
| CN107491717A (en) * | 2016-06-13 | 2017-12-19 | 科大讯飞股份有限公司 | The detection method that cheats at one's exam and device |
| CN109815816A (en) * | 2018-12-24 | 2019-05-28 | 山东山大鸥玛软件股份有限公司 | A kind of examinee examination hall abnormal behaviour analysis method based on deep learning |
| CN110032992A (en) * | 2019-04-25 | 2019-07-19 | 沈阳航空航天大学 | A kind of detection method that cheats at one's exam based on posture |
| CN110781762A (en) * | 2019-09-30 | 2020-02-11 | 沈阳航空航天大学 | Examination cheating detection method based on posture |
| CN111597911A (en) * | 2020-04-22 | 2020-08-28 | 成都运达科技股份有限公司 | Method and system for rapidly extracting key frame based on image characteristics |
| CN112036299A (en) * | 2020-08-31 | 2020-12-04 | 山东科技大学 | Examination cheating behavior detection method and system under standard examination room environment |
| CN112087603A (en) * | 2020-08-31 | 2020-12-15 | 浙江蓝鸽科技有限公司 | Intelligent examination room supervision method |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11195057B2 (en) * | 2014-03-18 | 2021-12-07 | Z Advanced Computing, Inc. | System and method for extremely efficient image and pattern recognition and artificial intelligence platform |
-
2020
- 2020-12-28 CN CN202011584152.2A patent/CN112613436B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2017152425A1 (en) * | 2016-03-11 | 2017-09-14 | 深圳市大疆创新科技有限公司 | Method, system and device for preventing cheating in network exam, and storage medium |
| CN107491717A (en) * | 2016-06-13 | 2017-12-19 | 科大讯飞股份有限公司 | The detection method that cheats at one's exam and device |
| CN106203363A (en) * | 2016-07-15 | 2016-12-07 | 中国科学院自动化研究所 | Human skeleton motion sequence Activity recognition method |
| CN109815816A (en) * | 2018-12-24 | 2019-05-28 | 山东山大鸥玛软件股份有限公司 | A kind of examinee examination hall abnormal behaviour analysis method based on deep learning |
| CN110032992A (en) * | 2019-04-25 | 2019-07-19 | 沈阳航空航天大学 | A kind of detection method that cheats at one's exam based on posture |
| CN110781762A (en) * | 2019-09-30 | 2020-02-11 | 沈阳航空航天大学 | Examination cheating detection method based on posture |
| CN111597911A (en) * | 2020-04-22 | 2020-08-28 | 成都运达科技股份有限公司 | Method and system for rapidly extracting key frame based on image characteristics |
| CN112036299A (en) * | 2020-08-31 | 2020-12-04 | 山东科技大学 | Examination cheating behavior detection method and system under standard examination room environment |
| CN112087603A (en) * | 2020-08-31 | 2020-12-15 | 浙江蓝鸽科技有限公司 | Intelligent examination room supervision method |
Non-Patent Citations (3)
| Title |
|---|
| Learning skeleton representations for human action recognition;Alessia Saggese等;《Pattern Recognition Letters》;第118卷;23-31 * |
| 基于图像分析的考场视频监视系统分析与研究;刘鸾;《中国硕士学位论文全文数据库(信息科技辑)》(第2014（10）期);I136-440 * |
| 基于视频行为分析的智能监考辅助系统;李春梅等;《科技创新与应用》(第18期);8-10 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112613436A (en) | 2021-04-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111709409B (en) | Face living body detection method, device, equipment and medium | |
| CN109165552B (en) | Gesture recognition method and system based on human body key points and memory | |
| Liong et al. | Automatic micro-expression recognition from long video using a single spotted apex | |
| CN109034069B (en) | Method and apparatus for generating information | |
| US20140143183A1 (en) | Hierarchical model for human activity recognition | |
| KR102265573B1 (en) | Method and system for reconstructing mathematics learning curriculum based on artificial intelligence | |
| CN112149615B (en) | Face living body detection method, device, medium and electronic equipment | |
| CN112329663B (en) | Micro-expression time detection method and device based on face image sequence | |
| CN108229375B (en) | Method and device for detecting face image | |
| CN111310605B (en) | Image processing method and device, electronic equipment and storage medium | |
| CN112183504B (en) | Video registration method and device based on non-contact palm vein image | |
| CN113378804A (en) | Self-service sampling detection method and device, terminal equipment and storage medium | |
| CN113869229A (en) | Deep learning expression recognition method based on prior attention mechanism guidance | |
| Pang et al. | A stochastic model of selective visual attention with a dynamic Bayesian network | |
| CN114565976A (en) | Training intelligent test method and device | |
| CN109492584A (en) | A kind of recognition and tracking method and electronic equipment | |
| CN112613436B (en) | Examination cheating detection method and device | |
| CN110096989B (en) | Image processing method and device | |
| Huang et al. | Visual attention learning and antiocclusion-based correlation filter for visual object tracking | |
| CN110929583A (en) | High-detection-precision face recognition method | |
| CN106446837B (en) | A kind of detection method of waving based on motion history image | |
| CN115565097A (en) | Method and device for detecting compliance of personnel behaviors in transaction scene | |
| CN114359965A (en) | Training method and training device | |
| Jain et al. | Detecting missed and anomalous action segments using approximate string matching algorithm | |
| CN114627534A (en) | Living body discrimination method, electronic device, and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |