REAL-TIME EXAM SURVEILLANCE SYSTEM

Abstract

The present invention discloses a real-time exam surveillance system (100), including at least one high-resolution camera (110) positioned to capture live video feeds from an examination hall. Embodiments may also include a YOLO-CNN module (120) configured to process the video feeds in real time. In some embodiments, the YOLO (You Only Look Once)model detects objects, and the Convolutional Neural Network (CNN)identifies unauthorized devices and suspicious behaviours indicative of cheating. Embodiments may also include a feedback loop (130) within the CNN module to dynamically adjust detection thresholds based on abnormalities to enhance detection accuracy and minimize false positives. Embodiments may also include a university portal integration module (140) for real-time updates. In some embodiments, details of detected incidents, including timestamp (142) and evidence (144), may be recorded and displayed on student profiles.

Specification

Description:The following specification particularly describes the invention and the manner in which it is to be performed.
CROSS-REFERENCE TO RELATED APPLICATIONS AND PRIORITY
[001] The present application does not claim priority from any patent application.
PREAMBLE
[002] The following specification particularly describes the invention and the manner in which it is to be performed.
TECHNICAL FIELD
[003] The present subject matter described herein, in general, relates to real-time surveillance systems for examination environments, and more particularly, to an automated system utilizing YOLO-CNN technology for detecting unauthorized objects and suspicious behaviours indicative of cheating during exams.
BACKGROUND
[004] In recent years, academic institutions have increasingly faced challenges in maintaining the integrity of examinations due to the growing sophistication of cheating methods. Traditional exam monitoring systems primarily rely on human invigilators to oversee students, but these methods are often inefficient, especially in large examination halls or online assessments. Human proctors are prone to fatigue, and their ability to monitor every student in real time is limited. Consequently, unauthorized use of electronic devices, such as mobile phones, smartwatches, and concealed earphones, has become prevalent, allowing students to discreetly access information during exams.
[005] Existing automated systems for exam monitoring attempt to address this issue by employing basic object detection technologies or by monitoring students through webcams during online exams. However, these systems often suffer from high rates of false positives or delayed detection, which significantly limits their effectiveness. Additionally, many prior systems are not adaptable to in-person examinations and require extensive manual review of recorded footage to confirm incidents, which is time-consuming and impractical for large-scale exam environments.
[006] Moreover, existing solutions tend to lack integration with academic portals or incident reporting mechanisms, which complicates the process of logging, tracking, and responding to detected cheating attempts. Current approaches are often unable to provide real-time notifications to invigilators, which delays intervention and allows cheating incidents to go unnoticed until after the exam has concluded.
[007] The prior art IN202421040714A (hereafter referred to as patent '714) discloses an IoT-based Noise Detection & Alert System designed to monitor noise levels and enhance security across diverse environments. Using noise sensors integrated with a Raspberry Pi, the system continuously tracks ambient noise in real-time. It compares noise levels against set thresholds to detect disturbances and send automated alerts. The system also incorporates image capture to provide visual context for incidents, aiding targeted interventions and documentation. Its versatility makes it ideal for educational, commercial, and residential settings, offering proactive noise pollution control, intruder detection, and real-time security responses for safer environments. However, the patent '714 fails to mention the use of real-time video processing through a YOLO-CNN model for detecting unauthorized devices and suspicious behaviours in exams. Patent '714 further lacks integration with university portals for updating student profiles and sending immediate alerts to invigilators. Furthermore, the patent '714 does not address dynamic threshold adjustment for enhanced detection accuracy or scalability to monitor multiple exam halls simultaneously, limiting its application compared to the present invention.
[008] The prior art US11232686B2 (hereafter referred to as patent '686) discloses a method may include transmitting a video stream of a live scene over a network at a real-time transmission speed and detecting an event associated with the video stream being transmitted. The method may include transmitting the video stream over the network at a speed lower than the real-time transmission speed during the event. Transmitting the video stream at the speed lower than the real-time transmission speed may introduce a time stretch for the video stream to be played in slow motion. The method may include reducing a bit rate of the video stream after the event and transmitting the video stream with the reduced bitrate over the network after the event to compensate for the time stretch. However, the patent '685 fails to mention a real-time exam surveillance system (100) that dynamically detects unauthorized devices and suspicious behaviours using a module, while integrating with university portals for instant updates and alerts to invigilators.
[009] The prior art US20210366072A1 (hereafter referred to as patent '072) discloses a threat detection system that shows a user an incident as it develops in real time by leveraging artificial intelligence (AI) to more accurately focus cameras and highlight the areas of concern within those feeds, providing a much more efficient user interface to the operator. These annotated feeds and feed focused triggering events can also be connected to third party systems. This timeline of events and evidence (small annotated clip of detection when available) is archived and can be reviewed at a later date, containing an accurate timeline of the incident as it progressed. However, the patent '072 fails to mention a targeted approach for real-time exam surveillance, focusing instead on general threat detection. The present invention utilizes a module for accurate detection of unauthorized devices and suspicious behaviours, with dynamic feedback and scalable monitoring capabilities.
[0010] The present invention addresses the above shortcomings of the prior art. However, the invention is entirely different from the prior art in terms of novelty and technological advancements.
OBJECT
[0011] The object of the present invention is to provide a real-time exam surveillance system that leverages YOLO-CNN technology to detect unauthorized devices and suspicious behaviours during exams, ensuring accurate, automated monitoring and maintaining exam integrity through instant alerts and detailed incident reporting.
SUMMARY
[0012] The disclosure of the present invention discloses a real-time exam surveillance system (100), including at least one high-resolution camera (110) positioned to capture live video feeds from an examination hall. Embodiments may also include a YOLO-CNN module (120) configured to process the video feeds in real time. In some embodiments, the YOLO (You Only Look Once) model detects objects, and the Convolutional Neural Network (CNN)identifies unauthorized devices and suspicious behaviour indicative of cheating.
[0013] Embodiments may also include a feedback loop (130) within the CNN module to dynamically adjust detection thresholds based on abnormalities to enhance detection accuracy and minimize false positives. Embodiments may also include a university portal integration module (140) for real-time updates. In some embodiments, details of detected incidents, including timestamp (142) and evidence (144), may be recorded and displayed on student profiles. Embodiments may also include a notification system (150) that sends real-time alerts to invigilators through a user interface (UI), providing incident details including time (152), location (154), and suspected cheating activity (156).
[0014] In some embodiments, the high-resolution camera (110) may be configured to capture panoramic views of the entire examination hall, enabling comprehensive monitoring of multiple desks simultaneously. In some embodiments, the system as claimed may include a processing unit, selected from a GPU or dedicated server, configured to handle the computational requirements for real-time video analysis and object detection by the YOLO-CNN module (120).
[0015] In some embodiments, the YOLO model detects objects such as mobile phones or electronic devices placed under desks, and the CNN algorithm may be trained to recognize subtle behaviour, such as body posture changes associated with cheating attempts. In some embodiments, the system as claimed may include a reporting module configured to generate comprehensive logs of detected cheating incidents, which may be accessible through the university portal for post-examination review by authorities.
[0016] In some embodiments, the system as claimed may include a non-transitory computer-readable medium containing instructions which, when executed by a processor, cause the processor to capture live video from examination environments. Embodiments may also include process the video feeds in real time using a YOLO-CNN model to detect unauthorized objects and suspicious behaviour. Embodiments may also include automatically update student profiles with cheating incident details and send real-time alerts to invigilators with relevant evidence (144) for immediate action.
[0017] Embodiments of the present disclosure may also include a method for real-time detection and reporting of cheating in examination environments, including capturing live video feeds from examination rooms using high-resolution cameras (110). Embodiments may also include processing the video feeds using a YOLO-CNN algorithm to detect unauthorized objects, such as electronic devices, and identify suspicious behaviour indicative of cheating.
[0018] Embodiments may also include updating student profiles on a university portal with the detected incident details, including timestamp (142), object identification, and video evidence (144). Embodiments may also include generating and sending real-time alerts to invigilators. In some embodiments, the alert includes event details and evidence (144) for immediate review and action. In some embodiments, the method as claimed may include scalability and adaptability features. In some embodiments, the system may be capable of monitoring multiple examination halls simultaneously without requiring significant reconfiguration, and can adapt to varying seating arrangements and exam hall layouts.
BRIEF DESCRIPTION
[0019] The foregoing detailed description of embodiments is better understood when read in conjunction with the appended drawings. For the purpose of illustrating of the present subject matter, an example of construction of the present subject matter is provided as figures; however, the invention is not limited to the specific method disclosed in the document and the figures.
[0020] Figure. 1 is a block diagram illustrating a real-time exam surveillance system, according to some embodiments of the present disclosure.
[0021] Figure. 2 is a flowchart illustrating a method, according to some embodiments of the present disclosure.
DETAILED DESCRIPTION
[0022] Some of the embodiments of this disclosure, illustrating all its features will now be discussed in detail. The words "comprising," "having," "containing," and "including," and other forms thereof, are intended to be equivalent in meaning and be open-ended in that an item or items following any one of these words is not meant to be an exhaustive listing of such item or items or meant to be limited to only the listed item or items. It must also be noted that as used herein and in the appended claims, the singular forms "a," "an," and "the" include plural references unless the context clearly dictates otherwise.
[0023] Before the present a real-time exam surveillance system designed to enhance the integrity of examinations by utilizing YOLO-CNN technology for automated detection of cheating activities in academic environments is described, it is to be understood that this method is not limited to the particular system(s), and methodologies described, as there can be multiple possible embodiments that are not expressly illustrated in the present disclosure. It is also to be understood that the terminology used in the description is to describe the particular implementations or versions or embodiments only, and is not intended to limit the scope of the present application. This summary is provided to introduce aspects involved in a real-time exam surveillance system designed to enhance the integrity of examinations by utilizing YOLO-CNN technology for automated detection of cheating activities in academic environments. This summary is not intended to identify essential features of the claimed subject matter nor is it intended for use in determining or limiting the scope of the claimed subject matter.
[0024] The present invention discloses a real-time exam surveillance system (100) designed to enhance the integrity of examinations by utilizing YOLO-CNN technology for automated detection of cheating activities in academic environments. The system integrates high-resolution cameras (110) strategically placed in examination halls to capture live video feeds of students and their surroundings. These video feeds are processed in real time using a YOLO (You Only Look Once) Convolutional Neural Network (CNN) model, which is specifically programmed to detect unauthorized objects, such as mobile phones or earphones, and recognize suspicious behaviours, such as body posture changes indicative of cheating. The YOLO model excels in identifying objects, while the CNN is fine-tuned to detect subtle behavioural anomalies. A feedback loop (130) mechanism, which allows the system to dynamically adjust detection thresholds based on abnormalities, thereby reducing false positives and enhancing the system's precision. When unauthorized behaviour or objects are detected, the system immediately updates the student's profile on the university portal, recording incident details such as timestamps (142), object identification, and evidence (144) like video footage. The system also includes a notification module that sends real-time alerts to invigilators through a user interface, allowing them to monitor or intervene in suspected cheating incidents promptly. The invention is scalable, adaptable to multiple examination halls, and does not require extensive reconfiguration if the layout changes. This automated surveillance approach provides a highly reliable, efficient solution to ensure the integrity of examinations, reducing reliance on human proctors and minimizing the risk of cheating.
[0025] In one embodiment, the system comprises strategically placed high-resolution cameras (110), capable of capturing wide-angle or panoramic views of the examination hall. These cameras are installed in such a way that they cover all areas where students are seated, ensuring clear, unobstructed visibility of each student's desk. The cameras continuously capture real-time video streams with high clarity, enabling the system to monitor for unauthorized objects or suspicious activities during the exam. The captured footage provides a basis for accurate object and behaviour detection, which is processed in real time by the YOLO-CNN module (120).
[0026] In one embodiment, the YOLO-CNN module (120) is configured to analyze the live video feeds received from the high-resolution cameras (110). The YOLO (You Only Look Once) model detects objects such as mobile phones, earphones, or other unauthorized devices that may be concealed on or near a student's desk. Concurrently, the CNN is trained to recognize suspicious behaviours, such as a student's body posture or hand movements indicative of cheating, including attempts to use hidden electronic devices. The combination of object detection and behavioural analysis allows the system to quickly and accurately identify potential cheating incidents.
[0027] In one embodiment, the system incorporates a feedback loop (130) within the CNN module that allows for real-time adjustments to the detection algorithm. When the system identifies an abnormality, such as a potential false positive or an undetected object, the feedback loop (130) refines the detection parameters, dynamically modifying the threshold values. This ensures that the CNN can better differentiate between legitimate actions, such as a student shifting in their seat, and suspicious behaviour, such as unauthorized device use. As a result, the system continuously improves its detection accuracy while reducing the occurrence of false positives.
[0028] In one embodiment, the system is integrated with a university portal that allows for the real-time reporting of detected cheating incidents. When the system identifies suspicious activity, it records key details such as the time (152) of the incident, the type of unauthorized device or behaviour detected, and any associated evidence (144), such as screenshots or video clips. These details are automatically linked to the student's profile on the university portal, enabling exam administrators and authorities to review the evidence (144). The integration ensures that all detected incidents are logged and accessible for post-exam evaluation, maintaining transparency and integrity in the examination process.
[0029] In one embodiment, the system includes a notification system (150) that is configured to send real-time alerts to invigilators whenever a suspicious activity or unauthorized device is detected. The notification system (150) is connected to a user interface (UI) that displays critical information about the incident, including the time (152) of detection, the exact location (154) within the examination hall, and a description of the suspected cheating activity (156). These real-time alerts enable invigilators to promptly respond to the situation, either by physically intervening or by monitoring the student more closely through the surveillance system, thereby preventing cheating in real time.
[0030] In one embodiment, referring to Figure. 1 is a block diagram that describes a real-time exam surveillance system (100), according to some embodiments of the present disclosure. In some embodiments, the real-time exam surveillance system (100) may include at least one high-resolution camera (110) positioned to capture live video feeds from an examination hall, a YOLO-CNN module (120) configured to process the video feeds in real time, a feedback loop (130) within the CNN module to dynamically adjust detection thresholds based on abnormalities to enhance detection accuracy and minimize false positives, and a university portal integration module (140) for real-time updates. The real-time exam surveillance system (100) may also include a notification system (150) that sends real-time alerts to invigilators through a user interface (UI), providing incident details.
[0031] In some embodiments, the YOLO (You Only Look Once)model detects objects, and the Convolutional Neural Network (CNN) identifies unauthorized devices and suspicious behaviours indicative of cheating. The university portal integration module (140) may include timestamp (142). The university portal integration module (140) may also include evidence (144), may be recorded and displayed on student profiles. Details of detected incidents. The notification system (150) may include time (152), location (154), and suspected cheating activity (156).
[0032] In some embodiments, the high-resolution camera (110) may be configured to capture panoramic views of the entire examination hall, enabling comprehensive monitoring of multiple desks simultaneously. In some embodiments, the YOLO model detects objects such as mobile phones or electronic devices placed under desks, and the CNN algorithm may be trained to recognize subtle behaviours, such as body posture changes associated with attempts.
[0033] In some embodiments, the system as claimed. In some embodiments, the system as claimed. A non-transitory computer-readable medium. Capture live video from examination environments. Process the video may feed in real time using a YOLO-CNN model to detect unauthorized objects and suspicious behaviours. Automatically update student profiles with cheating incident details and send real-time alerts to invigilators with relevant evidence (144) for immediate action.
[0034] In one embodiment, referring to Figure. 2 is a flowchart that describes a method, according to some embodiments of the present disclosure. In some embodiments, at (210), the method may include capturing live video feeds from examination rooms using high-resolution cameras (110). At (220), the method may include processing the video feeds using a YOLO-CNN algorithm to detect unauthorized objects, such as electronic devices, and identify suspicious behaviour indicative of cheating. At (230), the method may include updating student profiles on a university portal with the detected incident details, including timestamp (142), object identification, and video evidence (144). At (240), the method may include generating and sending real-time alerts to invigilators. The alert may include event details and evidence (144) for immediate review and action. In some embodiments, the method as claimed. Scalability and adaptability may feature. The system may be capable of monitoring multiple examination halls simultaneously without requiring significant reconfiguration, and can adapt to varying seating arrangements and exam hall layouts.
, Claims:1. A real-time exam surveillance system (100), comprising:
a. at least one high-resolution camera (110) positioned to capture live video feeds from an examination hall;
b. a YOLO-CNN module (120) configured to process the video feeds in real time, wherein the YOLO (You Only Look Once) model detects objects, and the Convolutional Neural Network (CNN) identifies unauthorized devices and suspicious behaviours indicative of cheating;
c. a feedback loop (130) within the CNN module to dynamically adjust detection thresholds based on abnormalities to enhance detection accuracy and minimize false positives;
d. a university portal integration module (140) for real-time updates, wherein details of detected incidents, including timestamp (142) and evidence (144), are recorded and displayed on student profiles; and
e. a notification system (150) that sends real-time alerts to invigilators through a user interface (UI), providing incident details including time (152), location (154), and suspected cheating activity (156).
2. The system as claimed in claim 1, wherein the high-resolution camera (110) is configured to capture panoramic views of the entire examination hall, enabling comprehensive monitoring of multiple desks simultaneously.
3. The system as claimed in claim 1, further comprising a processing unit, selected from a GPU or dedicated server, configured to handle the computational requirements for real-time video analysis and object detection by the YOLO-CNN module (120).
4. A method for real-time detection and reporting of cheating in examination environments, comprising:
a. capturing live video feeds from examination rooms using high-resolution cameras (110) (210);
b. processing the video feeds using a YOLO-CNN algorithm to detect unauthorized objects, such as electronic devices, and identify suspicious behaviour indicative of cheating (220);
c. updating student profiles on a university portal with the detected incident details, including timestamp (142), object identification, and video evidence (144) (230); and
d. generating and sending real-time alerts to invigilators, wherein the alert includes event details and evidence (144) for immediate review and action (240).
5. The method as claimed in claim 4, further comprising scalability and adaptability features, wherein the system is capable of monitoring multiple examination halls simultaneously without requiring significant reconfiguration, and can adapt to varying seating arrangements and exam hall layouts.
6. The system as claimed in claim 1, wherein the YOLO model detects objects such as mobile phones or electronic devices placed under desks, and the CNN algorithm is trained to recognize subtle behaviours, such as body posture changes associated with cheating attempts.
7. The system as claimed in claim 1, further comprising a reporting module configured to generate comprehensive logs of detected cheating incidents, which are accessible through the university portal for post-examination review by authorities.
8. The system as claimed in claim 1, further comprising a non-transitory computer-readable medium containing instructions which, when executed by a processor, cause the processor to:
a. capture live video from examination environments;
b. process the video feeds in real time using a YOLO-CNN model to detect unauthorized objects and suspicious behaviours; and
c. automatically update student profiles with cheating incident details and send real-time alerts to invigilators with relevant evidence (144) for immediate action.

Documents