In this paper robust image authentication integrated with semi-fragile pixel-wise tamper localization is analyzed. A new pixel-wise transformation robust to blurring/sharpening while fragile to all other image processing operations is proposed. A new method featuring binary and percentage measures with novel ability to integrate human opinion for image authenticity evaluation is presented. Protection for all bits in the pixel is advantage as well as small size of the signature – less than 10% of initial image.