Peningkatan Kinerja Steganografi Multi-Stego pada Citra Digital melalui Segmentasi Entropi dan Secret Sharing

Elqolby, Nazhifah (2025) Peningkatan Kinerja Steganografi Multi-Stego pada Citra Digital melalui Segmentasi Entropi dan Secret Sharing. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6025241059-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (5MB) | Request a copy

Abstract

Keamanan dalam transmisi informasi merupakan aspek krusial dalam menjaga privasi data, sehingga mendorong berkembangnya steganografi sebagai teknik perlindungan data rahasia. Namun, metode steganografi konvensional masih menghadapi tantangan dalam menyeimbangkan kapasitas penyematan dan kualitas visual citra stego, yang dapat menimbulkan distorsi mencolok dan meningkatkan risiko terdeteksi. Penelitian ini mengusulkan suatu model steganografi baru bernama ELStego, yang memanfaatkan pendekatan multi-stego dengan skema post-embedding secret sharing untuk meningkatkan efisiensi penyematan sekaligus imperceptibility citra stego. Model ini menerapkan segmentasi citra berbasis entropi untuk membagi citra sampul ke dalam blok-blok secara dinamis. Blok-blok tersebut kemudian dipilih secara optimal dengan algoritma hybrid, yaitu gabungan dari Fruit Fly Optimization Algorithm (FOA) dan Improved Seeker Optimization Algorithm (ISOA), yang dirumuskan sebagai algoritma Fruit Fly Optimization hybridized with Improved Seeker Optimization (FOIS). Proses penyematan dilakukan menggunakan teknik histogram shifting yang diarahkan oleh seleksi piksel berbasis FOIS, memungkinkan penyisipan data yang adaptif dan efisien. Setelah data rahasia disisipkan ke dalam citra sampul dan menghasilkan satu citra stego, proses secret sharing berbasis polinomial diterapkan pada citra tersebut untuk membagi isi tersembunyinya menjadi beberapa share. Hasilnya berupa lima citra stego (multi-stego) yang identik secara visual tetapi menyimpan bagian informasi berbeda, dan hanya dapat direkonstruksi apabila jumlah minimum k citra tersedia. Evaluasi dilakukan menggunakan dataset citra dari Signal and Image Processing Institute (SIPI) serta bit rahasia acak. Hasil eksperimen menunjukkan bahwa ELStego mampu menghasilkan kualitas citra stego yang tinggi, dengan nilai Peak Signal-to-Noise Ratio (PSNR) rata-rata berkisar antara 50,69 dB (Aerial dan Tank, k=2) hingga 56,43 dB (Airplane, k=3). Tidak seperti pendekatan steganografi konvensional yang hanya menggunakan satu citra stego, ELStego mendistribusikan data rahasia ke dalam beberapa citra stego secara terdistribusi dan aman, sehingga meningkatkan kapasitas, ketahanan, dan kualitas visual.

======================================================================================================================================================

Security in information transmission is a crucial aspect of preserving data privacy, which has driven the development of steganography as a technique for concealing secret data. However, conventional steganographic methods still face challenges in balancing embedding capacity and the visual quality of stego images, often resulting in noticeable distortion and increasing the risk of detection. This study proposes a novel steganographic model called ELStego, which adopts a multi-stego approach with post-embedding secret sharing to enhance both embedding efficiency and the imperceptibility of stego images. The model employs entropy-based image segmentation to dynamically divide the cover image into variable-sized blocks. These blocks are then optimally selected using a hybrid algorithm that combines the Fruit Fly Optimization Algorithm (FOA) and the Improved Seeker Optimization Algorithm (ISOA), formulated as Fruit Fly Optimization hybridized with Improved Seeker Optimization (FOIS). Data embedding is carried out using the histogram shifting technique, guided by FOIS-based pixel selection, enabling adaptive and efficient data hiding. After the secret data is embedded into the cover image, producing a single stego image, a polynomial-based secret sharing scheme is applied to divide the hidden content into multiple shares. This results in five visually identical stego images (multi-stego) that each store a distinct portion of the concealed data. The original secret can only be reconstructed if at least k stego images are available. Experiments were conducted using image datasets from the Signal and Image Processing Institute (SIPI) along with random secret bits. The results demonstrate that ELStego achieves high stego image quality, with average Peak Signal-to-Noise Ratio (PSNR) values ranging from 50.69 dB (Aerial and Tank, k=2) to 56.43 dB (Airplane, k=3). Unlike conventional single-stego approaches, ELStego distributes the secret data across multiple stego images in a secure and redundant manner, significantly improving payload capacity, robustness, and visual quality.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	algoritma fois, keamanan informasi, multi-stego, penyembunyian data, pergeseran histogram, steganografi, data hiding, fois algorithm, histogram shifting, information security, multi-stego, steganography.
Subjects:	T Technology > T Technology (General) T Technology > T Technology (General) > T57.5 Data Processing T Technology > T Technology (General) > T57.83 Dynamic programming T Technology > T Technology (General) > T58.5 Information technology. IT--Auditing
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Informatics Engineering > 55101-(S2) Master Thesis
Depositing User:	Nazhifah Elqolby
Date Deposited:	04 Aug 2025 10:01
Last Modified:	04 Aug 2025 10:01
URI:	http://repository.its.ac.id/id/eprint/126164

Actions (login required)

View Item