With thousands of images piling inside the world serves daily, authentication of any of any image at any given time around the world continues to be an undeniable issue. Mounting rise in data and information exchange, a heavy use of images has become common place. The challenges related to image authentication are increasing at an alarming rate. Digital image segmentations in watermarking has been found to be of good use in improve the performance of image authentication. Undertaken work in this paper compare time duration of segments and block based decomposition, watermark self-embedding, segment level watermark embedding and superpixel SVD based watermark embedding approach for image authentication. With this experiments and analysis, the superpixel based image segmentation with SVD has been applied to retain the quality of image of the embedment of digital watermark with acceptable PSNR value. Also time obtain for segmentation is less compare to block based method. The applied work is likely to have its prospective application in medical image segmentation, copyright protection and in digital forensic image investigation.