Yen-Wei Chen

PCA Based Digital Watermarking

This work evaluates a novel watermarking method based on Principle Component Analysis and effectiveness of the method to some watermark attacks. A PCA is used on a block by block basis to decorrelate the image pixel, watermarks are added in the Principle Components of an image. A theoretical description of the method is included together with experimental results in order to validate the methodology presented. Simulation shows the performance of the method to be robust for image cropping and some attacks such as additive noise, low pass filtering, median filtering, and jpeg compression. This research presents a new approach to watermarking fields with good performance in image cropping, and enhancement to this system with respect to robustness against various attacks is under investigation.

ICA-based robust logo image watermarking

Digital watermarking is a technology proposed to address the issue of copyright protection for digital content. In this paper, we have developed a new robust logo watermarking technique. Watermark embedding is performed in the wavelet domain of the host image. The human visual system (HVS) is exploited by building a spatial mask based on stochastic model for content adaptive digital watermarking. Independent component analysis (ICA) is introduced to extract the logo watermark. Our simulation results suggest that ICA can be used to extract exactly watermark that was hidden in image and show that our system performs robustness well under various important types of attacks.

Heuristic reconstructions of neutron penumbral images

Penumbral imaging is a technique of coded aperture imaging proposed for imaging of highly penetrating radiations. To date, the penumbral imaging technique has been successfully applied to neutron imaging in laser fusion experiments. Since the reconstruction of penumbral images is based on linear deconvolution methods, such as inverse filter and Wiener filer, the point spread function of apertures should be space invariant; it is also sensitive to the noise contained in penumbral images. In this article, we propose a new heuristic reconstruction method for neutron penumbral imaging, which can be used for a space-variant imaging system and is also very tolerant to the noise.

A Robust Logo Multiresolution Watermarking Based on Independent Component Analysis Extraction

This paper proposes a novel blind logo multi-resolution watermarking technique based on independent component analysis (ICA) for extraction. To exploit the human visual system (HVS) and the robustness, a perceptual model is applied with a stochastic approach based on noise visibility function (NVF) for adaptive watermarking algorithm. A logo watermark is embedded by modifying middle-frequency sub-bands of wavelet transform. The new detection technique based on ICA is introduced during the extraction phase to ensure a blind watermark. The proposed algorithm is checked for the robustness to several compression algorithms such as Jpeg, jpeg 2000, SPIHT, EZW, and principal components analysis (PCA) based compression and also robust against various image and digital processing operators.

CT image reconstruction by back-propagation

A neural network model is used in CT image reconstruction from four projections. The system is based on the backpropagation algorithm for adaptation of connection weights. Satisfactory agreement between the original and reconstructed images was obtained in simulation, and the results obtained are compared to those obtained by the well-known algebraic reconstruction technique (ART), and it was found that the neural network method is more effective than ART when the number of projection directions is very limited.

Coded penumbral imaging for improvements of signal-to-noise ratio

Penumbral imaging is a power imaging technique for radiations with long mean-free path. Since the reconstruction is based on deconvolution, the technique is sensitive to noise contained in penumbral images. In this article, we proposed a uniformly redundant penumbral array (URPA) technique to improve the signal-to-noise (SN) ratio of penumbral images. In URPA, the penumbral apertures are arranged in m-sequence and the SN ratio of the penumbral image can be improved by a factor of [(r×s+1)/2]1/2, where r×s is the size of two-dimensional m sequence. The effectiveness of URPA has been demonstrated by computer simulations.

Image Retrieval Based on Independent Components of Color Histograms.

Color histograms are effective for representing color visual features. However, the high dimensionality of feature vectors results in high computational cost. Several transformations, including principal component analysis (PCA), have been proposed to reduce the dimensionality. PCA reduce the dimensionality by projecting the data to a subspace which contains most of the variance. It is restricted to an orthogonal transformation and may not be the optimal to represent the intrinsic features of data. In this paper, we apply independent component analysis (ICA) to extract the features in color histograms. PCA is applied to reduce the dimensionality and then ICA is performed on the low-dimensional PCA subspace. Furthermore, spatial information is incoporated by performing ICA on a color coherent vector (CCV). The experimental results show that the proposed method outperform other methods based on SVD of quadratic matrix or PCA, in terms of retrieval accuracy.

Blind deconvolution by genetic algorithms

A genetic algorithm is presented for the blind-deconvolution problem of image restoration. The restoration problem is modeled as an optimization problem, whose cost function is minimized based on mechanics of natural selection and natural genetics. The applicability of GA for blind-deconvolution problem was demonstrated.

Genetic Generation of High-Degree-of-Freedom Feed-Forward Neural Networks

Appropriate definition of neural network architecture prior to data analysis is crucial for successful data mining. This is particularly important when the underlying model of the data is unknown. The proposed algorithm is intended to develop automatically an appropriate neural network (including the number of layers, the number of processing elements per layer, and types of each processing element) needed to solve the given problem. Genetic programming (GP) is used to develop the neural network’s structure and the resilient-back-propagation (RPROP) will be used to train the neural network.

Evolutionary CT image reconstruction

An evolutionary algorithm for reconstructing CT gray images from projections is presented; the algorithm reconstructs two-dimensional unknown images from four one-dimensional projections. A Laplacian constraint term is included in the fitness function of the genetic algorithm for handling smooth images, and the evolutionary process reconstructs images into finer ones by partitioning the images gradually thereby increasing the chromosome size exponentially as the generation proceeds. Results obtained are compared to those obtained by the well-known algebraic reconstruction technique (ART), and it was found that the evolutionary method is more effective than ART when the number of projection directions is very limited.