Please use this identifier to cite or link to this item:
|Title:||Abnormal behaviours detection||Authors:||Lyu, Qing Yang||Keywords:||Engineering::Electrical and electronic engineering||Issue Date:||2022||Publisher:||Nanyang Technological University||Source:||Lyu, Q. Y. (2022). Abnormal behaviours detection. Final Year Project (FYP), Nanyang Technological University, Singapore. https://hdl.handle.net/10356/157370||Abstract:||With the continuous development of the artificial intelligence, the computer vison now has become more and more popular. There is a trend that many fields besides industrial sectors now start to adopt the computer vision. It helps to boost the efficiency in our life This project aims to use machine learning to design and develop a computer vision software to help invigilator to detect and identify abnormal behaviours during the examinations and tests. Once the software recognized suspicious behaviours, it will alter invigilator to check and review. It will aid invigilator to prevent any missing candidates’ suspicious behaviours. This software contains 3 core modules which are integrated by me to support it able to detect abnormal behaviours in the exam scenario. These core modules are YOLOv5 which is object detection algorithm that can detect any custom objects that trained by user, DeepSort algorithm which is able to track any recognized objects by YOLO algorithm. The algorithm will assign IDs to the objects for user to track. Lastly, in order to prevent candidate cheating by voice, the voice recognition function is powered by Azure Speech, which is able to detect any speech at any time in the background. In result, this software is able to identify abnormal behaviours such as people missing, multiple persons inside the screen, looking around and looking up. It is able to recognise the prohibited items in the exam such as cell phones, notebooks, textbooks and laptops.||URI:||https://hdl.handle.net/10356/157370||Fulltext Permission:||restricted||Fulltext Availability:||With Fulltext|
|Appears in Collections:||EEE Student Reports (FYP/IA/PA/PI)|
Updated on May 23, 2022
Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.