Secret Image Communication Scheme Based on Visual Cryptography and Tetrolet Tiling Patterns

Visual cryptographic scheme is specially designed for secret image sharing in the form of shadow images. The basic idea of visual cryptography is to construct two or more secret shares from the original image in the form of chaotic image. In this paper, a novel secret image communication scheme based on visual cryptography and Tetrolet tiling patterns is proposed. The proposed image communication scheme will break the secret image into more shadow images based on the Tetrolet tiling patterns. The secret image is divided into 4×4 blocks of tetrominoes and employs the concept of visual cryptography to hide the secret image. The main feature of the proposed scheme is the selection of random blocks to apply the tetrolet tilling patterns from the fundamental tetrolet pattern board. Single procedure is used to perform both tetrolet transform and the scheme of visual cryptography. Finally, the experimental results showcase the proposed scheme is an extraordinary approach to transfer the secret image and reconstruct the secret image with high visual quality in the receiver end.

residual reconstruction method for compressed sensing videos. They analyzed sparse representation using Karhunen-Loeve transform and multi frame reference. A modified threshold selection scheme based on Particle Swarm Optimization (PSO) and support vector machine (SVM) offered an accuracy of 90. 93% [He, Yu, Hong et al. (2018)]. Zhang et al. [Zhang, Yang, Li et al. (2018)] proposed a detection mechanism for seam carved images using uniform local binary patterns. These methods are used to embed watermark image into two cover images. Gurunathan et al. [Gurunathan and Rajagopalan (2020)] used the cuckoo search algorithm to find an optimal substitution matrix for transforming the message to be hidden. Xiao et al. [Xiao, Yang, Li et al. (2019)] proposed a lesion extraction method with reversible data hiding scheme. Patient details will be embedded into the lesion part of the image which achieves privacy preserving. Yan et al. [Yan, Xiang and Hua (2019)] proposed analysis-by-synthesis (AbS) framework to improve the quality of reconstructed images. The rest of the paper is organized as follows: In Section 2 brief description on the tetrominoes, construction of 4×4 random tetrolet patterns for the visual secret sharing and reconstruction procedures are presented. The experimental results and various analysis of image quality are presented in Section 3. Finally conclusions are presented in Section 4.

Methodologies used for image sharing and reconstruction 2.1 The main concepts
Many secret sharing algorithms were proposed for image protection and secure transmission. Traditional visual secret sharing schemes, pixel expansion, non-pixel expansion techniques, extended visual cryptography, dynamic visual cryptography and progressive visual secret sharing are the major techniques used to construct the shadow images. However, investigators faced many problems while attempting to implement these crypto schemes such as very hard to configure the matrix elements and quality loss in the reconstructed image. To deal with the aforesaid problems, it is most pressing requirement to find the nonexpandable block based technique for protecting secret images. In this paper, we propose a novel secret image communication scheme based on visual cryptography and tetrolet tiling patterns. The base idea of the proposed scheme is to identify the 4×4 block of tetrolet and non tetrolet patterns to construct the shadow images from the original image. The secret image is divided into equal size of odd and even blocks to allocate tetrominoes in the appropriate blocks. In each sector every alternate pattern is sealed by tetrolet tiles and rest of the patterns are marked as zero value matrices. After compilation of the both tetrolet and non tetrolet tiles, the patterns are chained together to construct the shadow images. Finally, by using the overlapping operation the original image has reconstructed without quality loss.

Symbols
Symbols in the proposed schemes are defined as follows: • I: The secret image with pixel intensity , • OddB: odd block • S: Shadow image with pixel intensity .

Framework for tetrolet patterns
The base concept of tetrolet is composing of different types of tetrominoes with proper assembling sequence. Tetrominoes are Haar based wavelet model by connecting equal sized squares. Four equal sizes of tetrominoes are used to make the shape of tetrolet pattern. Consider a digital image I with size of M×N. Let I={( , ): , = 0, … , − 1} ⊂ 2 be the index set of a digital image and the four connected neighborhood pixels and denoted as: (1) Sample connectivity is illustrated in the following Fig. 1. Considering the one dimensional tetromino, the indexing of four elements is described as: 2) The rule of tetromino for the index position is defined as, L:�⎯⎯⎯� {0, 1, 2, 3} (3)

Tetrolet in visual cryptography
A Digital image is composed of 'n' number of finite elements with fixed size. Each element is perfectly arranged by its pixel index position to make a two dimensional digital image. Consider a digital image I with size of 16×16 contains 256 pixels and position range from I1,1 to I16,16. Generally the tetrolet tiles are 4×4 based square matrices so that the secret image is divided into fixed number of 4×4 non overlapped blocks for tetrolet tiling process. Furthermore, the chaotic sequence is a most important factor of visual cryptography for constructing secret shares. Therefore, a sub block of original image is classified into two types of clusters as Tetrolet pattern and Non Tetrolet pattern. Visual representation of cluster combination area is presented in the Fig. 6.

Preliminaries for tetrolet tiles
The security of secret image is achieved by the determination of tetrolet patterns, non tetrolet patterns, number of clusters and number of blocks. Arithmetic operations are needed to calculate the perquisites for the proposed communication scheme. Tab. 4 shows that calculation of cluster and block size for the secret image.

Construction of visual cryptography secret shares
After the determination of cluster size, the number of tetrolet tiles is identified for preprocessing. The proposed construction of visual secret sharing scheme fixed tetrominoes in the respective blocks and rest of the non tetrolet blocks are fixed by zero matrices. The steps involved in generating the shadow images are as follows: Input: M×N size of secret image I Output: M×N size of secret share image S, Framed by both tetro and non-tetro patterns Step 1: Read the Original Image.
Step 2: Calculate the block and cluster size by the algorithm shown in the Tab. 4.
Step 3: Initialize the variables for block size, sequential processing for row and columns.
Step 6: Calculate the reminder value R for row process & reminder value C for column process.
Case 1: Apply tetrolet tiles for odd series of row blocks. Step 7: For i=1: SECT.
Secret Share1=combine (sector1: sector N) End Step 9: Repeat the process to generate more shares with different tetrolet tiles.
Step 10: Display the generates secret shares.

Revealing process
After the generation of multiple secret shares, the secure transmission is ensured at the sender area. In the revealing phase, the receiver must collect all the secret shares from various tetrolet participants and stack the secret shares one by one to reveal the original image. Without huge computational complexities; our proposed scheme performs the overlapping operation to reconstruct the secret image with high visual quality. The steps involved in revealing process are given in Tab. 5.

Experimental results
In this section, we have conducted experimental test for several domains. Tested images, comparisons, histogram analysis and image quality metrics are presented here. All the experiments are run on an Intel core i3-3110M CPU @ 2.40 GHZ with 2 GB of internal memory space. A gallery of gray scale images is show in the Fig. 7.

Experiments with 8 secret shares
A radiographic image is successfully communicated to another participant by the proposed secret sharing scheme. Eight secret shares are generated by using tetrolet tiling patterns as shown in Fig. 8. For security reason the secret image has divide into 4×4 block and further it is categorized by two factors like, tetro tile and non tetro tiles. Every secret share contains 256 tetro tiles and 16 tetro tiles in each row. All the information pixels are stored into tetro tile patterns and non tetro tiles contain no information. By stacking all the shadow images one by one, the secret image has revealed with high visual quality. All the share images retain the same size as the original image for tetrolet processing.

Experiments with 12 secret shares
In addition, the number of shadow images has increased; therefore twelve secret shares are generated by applying non tetro and tetrolet tiling patterns shown in Fig. 9. Every secret share contains 256 tetro tiles, 768 non tetro tiles. Each cell in this pattern has separated by odd and even sectors. An odd row sector contains 16 tetro and 16 non tetro tiles and even row sector contains 32 non tetro tiles to construct the secret shares.   We analyzed various image quality factors to determine the quality of reconstructed image and the observations is shown in Tab. 6. Considering the mean squared and peak signal noise ratio, the proposed scheme gives the minimum squared error and better peak signal noise ratio. The pixel intensity variations of the secret shares are illustrated in Fig. 10 and Fig. 11. The similarity measurement is calculated and shown in Tab

Conclusion
Secret sharing scheme is an astonishing technique to protect the confidential information for any type of applications. Many visual secret sharing schemes have been developed with high level mathematical computations that cause overlapping issues. In this paper, a novel secret communication scheme based on visual cryptography and tetrolet tiling patterns is proposed. At first, the original image is divided into 4×4 sub blocks. The alternative rows and columns of sub blocks are grouped into tetrolet and non-tetrolet blocks. Using the tetrominoes patterns, the proposed scheme encodes the tetrolet blocks of original image into meaningful secret shares. The non tetrolet blocks are completely replaced by zero. Finally, a new shadow image is obtained in regular intervals with different tetrolet patterns. The experimental results show that the secret images were reconstructed successfully with high visual quality. The visual representation of histogram generation is also presented in experimental section. The histogram shows the different level of pixel intensity for different tested images and its corresponding secret share images. The similarity between the original image and share images are also computed by correlation coefficients. The similarity checking has been done in all the orientations like vertical, horizontal and diagonal. Comparisons done between the proposed scheme and other approaches are also listed in the experimental section. Thus, the proposed secret communication scheme provides better security for any type of application with more number of participants involved in secret sharing domain. In future, proposed scheme can be extended for color images with different tetrolet patterns on the same share images.