Wyniki wyszukiwania - BazTech

1

An Inverse Halftoning Technique Using Modified Look-Up Tables

Chang C.-C., Chan C.-S.

Fundamenta Informaticae

|

2005

|

Vol. 68, nr 3

207-220

EN

In this paper, we shall propose a method to invert halftone images. We shall use a modified look-up table to reconstruct gray level images from halftone images. In the modified look-up table training phase, we shall use a sliding window to get halftone patterns and their corresponding gray-level lists. The modified look-up table can be obtained by calculating the centroid of the gray-level lists. In the image reconstruction phase, we shall use the sliding window to slide around the pixel to reconstruct so that we can get candidate gray values. The reconstructed gray value can be obtained by calculating the centroid of the candidate gray values. Experimental results have shown that our method indeed can get satisfied results.

2

A New Density-Based Scheme for Clustering Based on Genetic Algorithm

Lin C.-Y., Chang C.-C., Lin C.-C.

Fundamenta Informaticae

|

2005

|

Vol. 68, nr 4

315--331

EN

Density-based clustering can identify arbitrary data shapes and noises. Achieving good clustering performance necessitates regulating the appropriate parameters in the density-based clustering. To select suitable parameters successfully, this study proposes an interactive idea called GADAC to choose suitable parameters and accept the diverse radii for clustering. Adopting the diverse radii is the original idea employed to the density-based clustering, where the radii can be adjusted by the genetic algorithm to cover the clusters more accurately. Experimental results demonstrate that the noise and all clusters in any data shapes can be identified precisely in the proposed scheme. Additionally, the shape covering in the proposed scheme is more accurate than that in DBSCAN.

3

Embedding Invisible Watermarks into Digital Images Based on Side-Match Vector Quantization

Wu H.-C., Chang C.-C.

Fundamenta Informaticae

|

2004

|

Vol. 63, nr 1

89--106

EN

This paper proposes an effective digital watermarking scheme to protect the intellectual property rights of digital images. This scheme embeds representative watermarks into images during the process of side-match vector quantization (SMVQ) compresses and extracts the digital watermark effectively in order to determine the ownership of the image when a dispute over intellectual property rights occurs. The method proposed in this paper embeds a watermark based on the grouping of the codewords in a state codebook, and it stresses the fact that the SMVQ compressed image and the decompressed image both contain watermarks. In addition, the decompressed image has strong resistance against being damaged. Even under JPEG lossy compression, the watermark can still be restored effectively. The proposed method enables the watermark to be embedded directly into a compressed image that is much smaller than the original one. Once applied to the network, the method can save time in image transmission as well as storage space.

4

An Adaptive Steganographic Scheme for Color Images

Chang C.-C., Tsai P., Lin M.-H.

Fundamenta Informaticae

|

2004

|

Vol. 62, nr 3,4

275--289

EN

In this paper, a new palette-based image steganography is proposed. Palette-based images such as GIF files have been widely used on the Internet and web pages such that all browsers can recognize them. Palette-based images can be used as good carriers for transmitting secret messages. In this scheme, the Least Significant Bits (LSBs) of the palette-based image are utilized in order to hide the secret message by using the computational information hiding technique. The adaptive hiding capacity can be computed according to different sizes of the embedding block. Experimental results also show that existing artifacts in the traditional LSB-based method are unnoticeable. The number of modification pixels in the embedding process is decreased, while good stego-image quality is maintained, and the cover image is left unblemished by camouflaged information. Furthermore, the feature of adaptive hiding capacity provides the possibility of protecting different sizes of embedded secret messages.