Skintone detection based Steganography using Wavelet Transform
Manisha Kude, Manjusha Borse · 2016
This paper present a unique method in which secret data can be embedded within skin, as it is not that much critical to HVS (Human Visual System). Biometric features like skin-ton is beneficial for this method. In this process, the secret data is embedded in specified regions, in place of embedding data anywhere in image. At first take input image and then perform skin-tone detection on Hue, saturation, value (HSV) color model. Secondly, mask image is converted in frequency domain. This is achieved by applying Haar-Discrete Wavelet Transform (DWT), leading to four sub-bands. Sub-band LL (Low-Low band) is used for processing the data. Different levels of DWT are used with Haar transforms. Then the payload (number of bits that can be embedded) is calculated. By tracing skin pixel in one of the given high frequency sub-band, secret data embedding is Performed. First cropped the image then performed all steps for embedding process. Cropping results provides more security than without cropping, because at the decoding side this cropped region works as a key.