Te-Ming Tu

A Novel Image Steganographic Method Using Tri-way Pixel-Value Differencing

To enlarge the capacity of the hidden secret information and to provide an imperceptible stego-image for human vision, a novel steganographic approach using tri-way pixel-value differencing (TPVD) is proposed in this paper. To upgrade the hiding capacity of original PVD method referring to only one direction, three different directional edges are considered and effectively adopted to design the scheme of tri-way pixel-value differencing. In addition, to reduce the quality distortion of stego-image brought from setting larger embedding capacity, an optimal approach of selecting the reference point and adaptive rules are presented. Theoretical estimation and experimental results demonstrate that the proposed scheme can provide superior embedding capacity and give secrecy protection from dual statistical stego-analysis. Besides, the embedded confidential information can be extracted from stego-images without the assistance of original images.

Best Tradeoff for High-Resolution Image Fusion to Preserve Spatial Details and Minimize Color Distortion

The fast intensity-hue-saturation (IHS) methods can quickly merge massive volumes of data from new satellite imagery. To minimize the color distortion from IHS fusion, Choi used a tradeoff parameter in his new fast IHS approach to control the spatial and spectral resolution of the fused image. However, the method is not general and requires manual adjustment of the tradeoff parameter in order to keep spatial details while preserving spectral characteristics. To tackle this problem, a modified approach that is a tunable IHS-Brovey method is presented in this letter. Experimental results demonstrate that the proposed technique can provide significant improvements over the original method, that is, to achieve a best tradeoff for spatial details and color distortion

Adjustable intensity-hue-saturation and Brovey transform fusion technique for IKONOS/QuickBird imagery

Among various image fusion methods, intensity-hue- saturation IHS and Brovey transforms BT can quickly merge huge amounts of IKONOS/QuickBird imagery. However, spectral degradation often appears in the fused images. Moreover, IHS and BT suffer from individual color distortion on saturation compression and saturation stretching, respectively. To balance these two saturation changes during the fusion process, an adjustable IHS-BT approach with spectral adjust- ment is proposed. Furthermore, to solve the typical bright target recovery BTR problems, a simple procedure of dynamic range adjustment DRA is also presented. By adopting different DRA techniques, the pro- posed IHS-BT method is divided into two different fusion approaches: the model of preserving spectral information and the model of enhancing spatial details. Experimental results demonstrate that the proposed com- bined approaches can achieve significant improvement over other cur- rent approaches.

Adaptive image steganographic scheme based on Tri-way Pixel-Value Differencing

To increase the capacity of the hidden secret information and to provide a stego-image imperceptible for human vision, a novel steganographic approach based on tri-way pixel-value differencing (TPVD) is presented in this paper. This approach uses three different directional edges effectively to design the tri-way differencing scheme and to remove the capacity limitation of the original PVD method using only one direction. In addition, we propose an optimal selection approach for the reference point and adaptive rules to reduce distortion brought from setting larger embedding capacity. The theoretical estimation and experimental results demonstrate that our scheme can provide a superior embedding capacity. Besides, the embedded confidential information can be extracted from stego-images without the assistance of original images.

Using empirical mode decomposition for iris recognition

Iris recognition is known as an inherently reliable technique for human identification. Empirical Mode Decomposition (EMD), an adaptive multi-resolution decomposition technique, appears to be suitable for non-linear, non-stationary data analysis. Based on EMD, a fully data-driven method without using any pre-determined filter or wavelet function, an iris recognition scheme is presented by modifying EMD as a low-pass filter to analyze the iris images. To evaluate the efficacy of the proposed approach, three different similarity measures are used. Experimental results show that three metrics have all achieved promising and similar performance. Therefore, the proposed method demonstrates to be feasible for iris recognition and EMD is suitable for feature extraction.

Iris recognition using local texture analysis

With the increasing needs in security systems, iris recognition is reliable as one of the important solutions for biometrics-based identification systems. This work presents an effective approach for iris recognition by analyzing iris patterns. To improve the rate of recognition, we divide the normalized iris image into several regions to keep the iris image away from several noise factors, such as eyelids, eyelashes, and motion blur. For feature extraction, the local edge pattern (LEP) operator is designed to capture local characteristics of the iris image to produce discriminating texture features in every region. A resulting 2D feature vector is mapped into a low-dimensional subspace using two dimension linear discriminant analysis (2DLDA), and then the minimum distance classifier (MDC) is adopted for recognition. Experiments on the public and freely available iris images taken from the CASIA (Institute of Automation, Chinese Academy of Sciences) and UBIRIS databases confirm the advantage of the proposed approach in terms of speed and accuracy.

A New Vegetation Enhancement/Extraction Technique for IKONOS and QuickBird Imagery

Most image fusion techniques are designed for multispectral (MS) images with natural color responses. If there is an unnatural spectral response, the fused image will be unsuitable for visualization. To enhance the vegetation content in such cases, applying the normalized difference vegetation index (NDVI) to tune the color of MS images is a feasible approach. However, the NDVI is usually used to generate lower resolution vegetation maps, and particularly, the threshold needs to be chosen manually for various scenes. In order to avert this drawback, using VI with a fixed threshold, a method is developed for IKONOS and QuickBird images in this letter, which integrates a fast technique of intensity-hue-saturation fusion. For MS images with unnatural color responses on vegetation areas, experimental results demonstrate that the proposed method can effectively enhance and extract vegetation information from the IKONOS and QuickBird images.

An Empirical Mode Decomposition Approach for Iris Recognition

Biometrics is inherently a reliable technique to identify humans authentication by his or her own physiological or behavioral characteristics. Empirical mode decomposition (EMD), a multiresolution decomposition technique, is adaptive and appears to be suitable for nonlinear, non-stationary data analysis. EMD analyzes the signal locally and separates the component holding locally the highest frequency from the rest into a separate component. In this paper, we adopt the EMD approach to extract residual components from the iris image as the features for recognition. Three different similarity measures have been evaluated. Experimental results show that three metrics have achieved similar performance. Therefore, the proposed method has demonstrated to be promising for iris recognition and EMD is suitable for feature extraction.

A Fixed-Threshold Approach to Generate High-Resolution Vegetation Maps for IKONOS Imagery

Vegetation distribution maps from remote sensors play an important role in urban planning, environmental protecting and related policy making. The normalized difference vegetation index (NDVI) is the most popular approach to generate vegetation maps for remote sensing imagery. However, NDVI is usually used to generate lower resolution vegetation maps, and particularly the threshold needs to be chosen manually for extracting required vegetation information. To tackle this threshold selection problem for IKONOS imagery, a fixed-threshold approach is developed in this work, which integrates with an extended Tasseled Cap transformation and a designed image fusion method to generate high-resolution (1-meter) vegetation maps. Our experimental results are promising and show it can generate more accurate and useful vegetation maps for IKONOS imagery.

Approach to maximize increased details and minimize color distortion for IKONOS and QuickBird image fusion

Both IKONOS and QuickBird offer 11-bit panchromatic and multispectral data in which more details can be extracted from scenes that are very dark under shadows or very washed out due to excessive sun reflectance. In this work, a discrete cosine transform (DCT)-based single-scale Retinex (SSR) technique is presented to extract the highest possible spatial details from the panchromatic image, and a minimum/maximum percent cutoff approach is used to preserve the most spectral information in the multispectral imagery. By fusing those two images, the typical bright target recovery (BTR) response can be avoided. Furthermore, to address the color distortion problem by the fusion process, a fast intensity hue saturation (IHS) image fusion technique with direct saturation stretching in RGB space is also proposed. To verify the efficacy of the proposed techniques, experiments are conducted using real IKONOS and QuickBird imagery covering different areas. Experimental results have shown that the perfect performance, maximizing increased details and minimizing color distortion, is achieved.