Image Steganography Under Multiple Performance Metrics: A Comparative Study

Image steganography performance depends on a trade-off among embedding capacity, imperceptibility, structural fidelity, robustness, security, and computational cost. This paper presents a comparative evaluation framework for spatial-domain and transform-domain image steganography using four standard benchmark images: Lena, Baboon, Barbara, and Cameraman. Representative spatial methods, Least Significant Bit (LSB) substitution and Pixel Value Differencing (PVD), are compared with transform-domain methods based on the Discrete Cosine Transform (DCT), Discrete Wavelet Transform (DWT), and a hybrid DWT–DCT approach. The evaluation parameters include embedding capacity, bits per pixel, PSNR, SSIM, extraction accuracy, robustness, and execution complexity. Equations, image-wise tables, average-performance tables, and bar charts are provided. The numerical comparison tables are explicitly presented as an illustrative experimental template because measured values require implementation under identical payload, image size, and attack conditions. Keywords: Steganography; LSB; PVD; DCT; DWT; Embedding Capacity; PSNR; SSIM; Performance Evaluation