Evaluación del Algoritmo F5 aplicado en la Esteganografía de imágenes JPEG: Un análisis basado en métricas de calidad
Miguel Fabricio Boné-Andrade, José Javier Mendoza Loor, Luis Alfonso Núñez-Freire · Journal of Economic and Social Science Research · 2025
The study explores the application of the F5 algorithm on JPEG images to evaluate its performance in hiding information without compromising the visual quality of the images. Compression techniques based on the Discrete Cosine Transform (DCT) and matrix coding are used to minimize the perceptible alterations that may occur. The methodology includes image segmentation in 8x8 pixel blocks, quantization of coefficients and evaluation of image quality using PSNR, MSE and SSIM metrics. The experiments were performed with grayscale images, hiding messages of different lengths. The results show that the algorithm maintains high values of PSNR and SSIM, ensuring that the distortion is imperceptible. An inverse relationship between message size and visual quality is observed, although without compromising image structure. The study concludes that the F5 algorithm is effective for steganography on JPEG images, but its performance is susceptible to improvement through the application of hybrid techniques such as DCT-SVD and adaptive quantization. The need to integrate neural network approaches to strengthen resistance to detection attacks is highlighted.