Page 74 - 2024F

UEC Int’l Mini-Conference No.53 67

Evolutionary Multi Objective Transfer Optimization.

Fermin DEL VALLE, Hiroyuki SATO
Department of Informatics
Hiroyuki Sato Laboratory, The University of Electro-Communications
Tokyo, Japan
delvallevegaferminalberto@gmail.com

Keywords: NSGA-II, Compressive Sensing, Chaotic Mixing, Secure Audio Processing.

Abstract
The increasing demand for secure and efficient digital audio storage and transmission requires
advanced methods that integrate encryption and compression. This study implements the Non-
Dominated Sorting Genetic Algorithm II (NSGA-II) to optimize encryption parameters for audio
signals using Compressive Sensing (CS) and Chaotic Mixing.
The proposed system segments audio into frames and applies the Discrete Cosine Transform
(DCT) for sparse representation. CS enables simultaneous compression and encryption through a
key-dependent sensing matrix, while Chaotic Mixing enhances security by scrambling the matrix
elements using additional keys. NSGA-II optimizes encryption settings by balancing two primary
objectives: maximizing randomness and minimizing correlation with the original signal.
Experimental results demonstrate that the system effectively secures audio, achieving a Pearson
Correlation Coefficient (PCC) of 0.97 with correct keys and 0.01 with incorrect keys. Spectrogram
and waveform analyses confirm the integrity of the reconstructed signal when decrypted correctly.
These findings suggest that NSGA-II-driven optimization enhances audio security while preserving
storage efficiency. Future research will explore advanced chaotic systems and alternative optimization
techniques.

∗
The author is supported by (JUSST) JASSO Scholar-
ship.