Jia Uddin

Work place: Artificial Intelligence and Big Data Department, Woosong University, Daejeon, South Korea

E-mail: jia.uddin@wsu.ac.kr

Website:

Research Interests: Deep Learning, Machine Learning

Biography

Jia Uddin is as an Assistant Professor, Department of AI and Big Data, Endicott College, Woosong University, Daejeon, South Korea. He received Ph.D. in Computer Engineering from University of Ulsan, South Korea, M.Sc. in Electrical Engineering (Specialization: Telecommunications), Blekinge Institute of Technology, Sweden, and B.Sc. in Computer and Communication Engineering, International Islamic University Chittagong, Bangladesh. He was a visiting faculty at School of computing, Staffordshire University, United Kingdom, Telkom University, Indonesia, and University of Foggia, Italy. He is an Associate Professor (now on leave), Department of Computer Science and Engineering, Brac University, Dhaka, Bangladesh. His research interests are Industrial Fault Diagnosis, Machine Learning/Deep Learning based prediction and detection using multimedia signals.

Author Articles
Drone Detection from Video Streams Using Image Processing Techniques and YOLOv7

By Muhammad K. Kabir Anika N. Binte Kabir Jahid H. Rony Jia Uddin

DOI: https://doi.org/10.5815/ijigsp.2024.02.07, Pub. Date: 8 Apr. 2024

For ensuring the safety issues, a country should establish a secure monitoring system around the most important places. Due to the huge development in unmanned aerial vehicles (UAV), drone detection is a vital part of the safety monitoring system for reducing threats from neighboring countries or terrorist groups. This paper presents a deep learning-based drone detection method. A You Only Look Once (YOLO) v7 architecture is used to train on the dataset. The training dataset consists of drone images in various environments. The trained model was tested on multiple videos of drones from YouTube. Experimental results demonstrate that the model exhibited a recall of 0.9656 and a precision of 0.9509. In addition, the performance of the model compares with the state-of-art models with YOLOv8, YOLO-NAS, Faster-RCNN architectures and it outperforms the other models by maintaining a more stable precision and recall curve.

[...] Read more.
Other Articles