Ensuring data security and privacy is essential when transmitting sensitive visual information over insecure networks. Reversible Data Hiding (RDH) enables confidential data embedding in digital images while allowing complete recovery of the original image after extraction. However, many existing RDH techniques suffer from visible distortion and vulnerability to steganalysis. This work proposes a security-enhanced RDH framework integrating median prediction and grayscale invariance for imperceptible and fully reversible embedding. Median prediction exploits local pixel correlations, while grayscale invariance preserves intensity consistency, reducing perceptual distortion and improving security. Experiments on benchmark grayscale images (Lena, Barbara, Cameraman, Peppers) achieve PSNR > 40 dB. Statistical metrics including entropy, SNR, and Chi-square analysis confirm strong resistance to detection and data tampering, supporting scalable and secure visual communication implemented in Python.