In this paper, a simple scheme for histogram
modification-based lossless information embedding
(HM-LIE) is proposed. The proposed scheme is free
from memorizing side information, i.e., blind. A HMLIE
scheme modifies particular pixel values in an image
in order to embed information in it on the basis
of its histogram, i.e., tonal distribution. The scheme
recovers the original image as well as extracts embedded
information from a distorted image conveying
embedded information. Most HM-LIE schemes
should memorize a set of image-dependent side information
per image. The proposed scheme does not
have to memorize such information to avoid costly
identification of the distorted image carrying embedded
information because of the introduction of two
mechanisms. One is estimating side information on
the basis of a simple statistic, and the other is concealing
not only main information but also a part of
the side information in the image. These approaches
make the proposed scheme superior to the conventional
blind schemes in terms of the quality of images
conveying embedded information.