A Novel efficient Block Based Segmentation Algorithm for Compound Image Compression
Compound image comprises of the combination of text, graphics and natural images. Compression and transmission of these compound images are the essential processes in the real time applications where segmentation plays a crucial role. Compound image transmission for real-time applications require that the compression technique should not only attains high compression ratio, moreover has low complexity, high PSNR value and a required level of security. Several approaches are being proposed in the past for the process of segmentation and compression of compound images. This paper proposes a novel method for block-based compound image compression. The pictorial blocks and text/graphics blocks are separated from the compound image. The pictorial blocks are compressed using discrete haar wavelet transformation and text/graphics colors are mapped to primary colors using color quantization algorithm and the resulting index values are then compressed using lossless Huffman coding technique. The compressed image is then encrypted using Advanced Encryption Standard algorithm. The proposed model provides several benefits like low complexity, high compression ratio, and a high PSNR value .Moreover it ensures that the transmission is fast and highly secured. Experimental results conclude that the proposed block based segmentation algorithm provides better results, than most of the other image compression techniques.
Keywords: Compound image compression, Haar wavelet transformation, Huffman coding, Advanced Encryption Standard algorithm 1. INTRODUCTION
Internet, fax, mobiles phones etc are the most common communication medium in use nowadays. The information used in these media is in digital format. Transmission of these digital forms of documents is done in a fraction of seconds. High compression and bit rates are required during transmission, to avoid expenses and delay. Thus the data compression becomes unavoidable. Electronic document images are often a mixture of data types, such as text, background, and foreground. These types of images are named as Compound images . Compound images must be compressed and transmitted in a secure way . Especially, with the popularity of cloud computing, compound images must be displayed on remote clients, wireless projectors, and thin clients. Some clients are not able to provide them from files directly. Compression and transmission of compound images renders a common solution to such clients. Though, in this solution, effective compression of compound images has become a widespread and vital problem . In the absence of data compression algorithms, it is surely impossible to transmit a large volume of data over the up to date bandwidth-constrained networks in real time . In this paper a hybrid compound image compression technique using block based segmentation is proposed. Initially the compound image is classified into 16x16 non-overlapping blocks and then segmented into pictorial and text/graphics blocks using color number thresholding technique. The pictorial blocks are then compressed using discrete haar wavelet transformation. The text/graphics blocks are first quantized based on their text/graphics pixel’s color probability and then compressed using Huffman coding. Both the compressed blocks are then merged into a single image which is then provided as input to the Advanced Encryption Standard algorithm that transforms the compressed image into an encrypted meaningless image. Now the compressed and encrypted image is transmitted to the receiver. The process is reversed for decryption and decompression at the receiver side. Thus the transmitted compound image is retrieved at the receiver side. Various compound image samples are used as input for the experimental analysis of the proposed methodology. Experimental results reveal the robustness of the proposed method. 2. RELATED...
 Wenpeng Ding, Yan Lu and Feng Wu, "Enable Efficient Compound Image Compression in H.264/AVC Intra Coding", IEEE International Conference on Image Processing, Vol. 2, pp: 337 - 340, 2007.
 Joan Daemen, Vincent Rijmen, "The Block Cipher Rijndael", Proceeding CARDIS '98 Proceedings of the The International Conference on Smart Card Research and Applications, pp: 277 - 284, 2000.
 Jagadhish H.and Lohit M. Kadlaskar, "A NEW LOSSLESS METHOD OF IMAGE COMPRESSION AND DECOMPRESSION USING HUFFMAN CODING TECHNIQUES", Journal of Theoretical and Applied Information Technology, pp: 18-23, 2010.
Tony Lin, Pengwei Hao, Sang Uk Lee, "Efficient Coding of Computer Generated Compound Images", IEEE International Conference onImage Processing, Vol.1, pp: 561-564, 2005.
 Xin Li, Shawmin Lei, "Block-based Segmentation and Adaptive Coding for Visually Lossless Compression of Scanned Documents", ICIP, Vol.3, pp: 450-453, 2001.
Please join StudyMode to read the full document