Image steganography camouflages secret messages in images by tampering image contents. There is a natural desire for hiding maximum secret information with the least possible distortions in the host image. This requires an algorithm that intelligently optimizes the capacity keeping the required imperceptibility of the image. This paper presents an image steganography scheme that preserves an adaptively chosen block of dominant coefficients from each Discrete Cosine Transform coefficients, whereas the rest of the coefficients are replaced with normalized secret image pixel values. Secret image pixel value are normalized in an adaptively chosen range. Embedding such kind of normalized data in adaptively chosen non-square L- shaped blocks utilize maximum embedding space available in each block that consequently results in maximizing payload capacity, while maintaining the image quality. This scheme achieved payload capacity up to 21.5 bit per pixel (bpp), while maintaining image quality of 38.24 dB peak signal to noise ratio.