İkili imgeler için blok tabanlı yeni bir kimlik doğrulama yöntemi

Digital watermarking techniques have been commonly used for binary image authentication. The digital watermarking techniques for binary images are different from other watermarking techniques because the pixel values of the binary images are 0 and 1. Digital watermarking techniques for binary images is more difficult than watermarking in graylevel or colored images, because the human vision system can perceive small changes in binary images. In this article, a new fragile watermarking method is proposed for authentication for binary images. In this paper, a new fragile block based watermarking method for binary images authentication is proposed. The proposed method consists of dividing into blocks, determining embeddable blocks, feature extraction, data hiding, data extraction, data verification and tamper detection steps. Firstly, the binary image is divided into blocks. According to rules, embeddable blocks are determined. Embeddable blocks are divided into 4 sub-blocks which are 3 feature extraction blocks and a data hiding block. The feature value is obtained by using the feature extraction blocks. Watermark embedding location (index) and authentication bit is calculated by using feature value. The pixel values of the data hiding blocks are determined by using feature and these values modified with authentication bits. In watermark extraction phase, feature values are calculated by using feature extraction. If the feature value of the block is the same as the value of the pixel, the authentication process is completed. Otherwise, the attack is detected. Blocks in variable sizes are used to measure the performance of the proposed method. Experimental results have shown that the proposed method has high visual quality, capacity and this method can be used in practice. Authentication bit generation and watermark embedding steps are given below. Step 1: Obtain image. Step 2: Determine block size Step 3: Divide image to sub-blocks by using nonoverlapping blocks. Step 4: Determine embeddable blocks. Pixels of embeddable blocks which are not all 0 and 1. In the proposed method, feature extractor blocks and data hiding block are used. Step 4: I used Eq. 1-7 to generate embeddable bit and data hiding indices. To extract feature, Eq. 1-7 are used. Step 5: The obtained authentication bit and index are used for data hiding. Data hiding process is described in Eq. 8. Data extraction and tamper detection steps are given below. Step 1: Use Eq. 1-7 for feature extraction. Step 2: If determined pixel value is equal to authentication bit, image authentication is successful. Step 3: If determined pixel value is not equal to authentication bit, go to tamper detection. In this method, 2M x 2N sized blocks are used for image authentication. 4 M x N sized blocks are obtained from 2M x 2N blocks. The first 3 blocks are used for feature extraction. The last block is used for data hiding. In the proposed method, 4 x 4, 6 x 6, 8 x 8 and 10 x 10 sized macro blocks are used for performance testing. 2 x 2, 3 x 3, 4 x 4 and 5 x 5 size blocks are also used for feature extraction and data hiding in these macro blocks. The characteristics of the proposed method are given below. The proposed method can extract authentication bits without any map or data. This method is blind watermarking and image authentication method. In a macro block, blocks are divided into feature extraction blocks and data hiding block. Authentication bit and its index are determined by using feature extraction blocks. Execution cost of the proposed method is low. The proposed method is an extensible method. (The users can define different features.) The proposed method can use sub-blocks in various sizes. The key is obtained by using embeddable pixels. The proposed method has tamper detection ability by using the key. The proposed method has high visual quality and high payload capacity. Visual quality and data hiding capacity are used for performance test. PSNR (Peak Signal Noise-to Ratio) and payload capacity are used in this test. Experimental results showed that the proposed method resulted successfully and the proposed method obtained higher capacity and visual quality than previously proposed methods in the literature.