Detecting digital audio drugs using deep learning

Al Husaini, Mohammed Abdulla Salim and Al Husaini, Yousuf Nasser and Habaebi, Mohamed Hadi (2025) Detecting digital audio drugs using deep learning. In: 2024 IEEE International Conference on Computing (ICOCO), 12-14 December 2024, KL Malaysia.

PDF - Published Version Restricted to Repository staff only Download (375kB) | Request a copy

Abstract

Digital drugs, auditory illusions created by playing slightly different frequencies in each ear, can influence mental states. Experiments were conducted using MATLAB 2023b on hardware with specifications of processor intel core i7 and graphic card NVIDIA GeForce GTX 4070. A dataset has a total of 7,000 audio files, divided into 5,000 audio drug files embedded with binaural beats and 1,000 original audio files from various categories. This database used to train and evaluate deep learning model to detect and classify audio drugs. Inception MV4 model was trained using SGDM optimizer over 3 epochs with different values of learning rates, achieving high performance metrics and demonstrating its efficacy in classification tasks. Inception MV4 model achieved average accuracies 99.9733% with learning rates 1e-3 and 1e-4, and an average accuracy 98.8833% with a learning rate 1e-5.

Item Type:	Proceeding Paper (Slide Presentation)
Uncontrolled Keywords:	Digital Drugs, Inception MV4, Deep Convolutional Neural Network, Learning Rate
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7800 Electronics. Computer engineering. Computer hardware. Photoelectronic devices > TK7885 Computer engineering
Kulliyyahs/Centres/Divisions/Institutes (Can select more than one option. Press CONTROL button):	Kulliyyah of Engineering > Department of Electrical and Computer Engineering Kulliyyah of Engineering
Depositing User:	Dr. Mohamed Hadi Habaebi
Date Deposited:	10 Apr 2025 16:39
Last Modified:	10 Apr 2025 16:39
URI:	http://irep.iium.edu.my/id/eprint/120522

Actions (login required)

View Item