<p>This paper presents an inversion-transposition-based transform-domain scrambling technique that achieves high imperceptibility of the stego image while ensuring robust security of the secret image. This scrambling technique leverages a hybrid discrete cosine transform (DCT) and discrete wavelet transform (DWT) model with a unique scrambling mechanism based on matrix inversion and transposition. Combined with adaptive thresholding, this approach simultaneously achieves high imperceptibility of the stego image and reliable protection of the secret image against unauthorized interpretation. Experimental analysis on the basis of various parameters shows the effectiveness of the proposed technique. The proposed method decomposed the cover image using DWT, followed by DCT in order to stabilize the energies, after which the devised scrambling inversion-transposition technique is applied on the secret image to enhance structural randomness and improve resistance against unauthorized interpretation, while maintaining high imperceptibility of the stego image. The two processed images are embedded into one another using adaptive masking and thresholding. This approach adjusts the embedding strength according to local image features, resulting in a stego image that shows no visible difference from the original cover image when tested on standard images.</p>