OBJECT IDENTIFICATION IN DEVELOPMENT OF CHEATING IN EXAM DETECTION USING GESTURE RECOGNITION
Cheating during examinations has become a growing concern in every educational institution due to its impact on the quality of education and eventually producing more human resources that are dishonest and without competitive spirit. Despite already assigning supervisors during the examinations, the...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Project |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/36337 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | Cheating during examinations has become a growing concern in every educational institution due to its impact on the quality of education and eventually producing more human resources that are dishonest and without competitive spirit. Despite already assigning supervisors during the examinations, there are still students who are able to cheat without getting caught. That is because human errors are very likely to occur during supervising, such as losing focus, being unaware of when a student is cheating, etc. Therefore, this final project presents a system for supervising exams continuously while also detecting any indications of cheating performed by students.
The project idea is based on a surveillance camera system, but added with cheating indication detection feature and the ability to send notification to supervisor. This system uses YOLO, an object detection algorithm that utilizes CNN to classify objects that a student uses for cheating and finger sign gestures. The testing result shows that detection using YOLO has some successfully detectable objects (70-80% accuracy) while some others could hardly be detected (17-29% accuracy), and in some cases could be mistaken for other objects. |
|---|