Feng-Cheng Chang

Watermarking for Compressive Sampling Applications

Compressive sampling is a newly developed topic in the field of data compression. For current researches, compressive sampling techniques focus on compression performances. There are very few papers aiming at the integration of watermarking into compressive sampling systems. In this paper, we propose an innovative scheme that considers the copyright protection of data with compressive sampling. By carefully utilizing the relationships between coefficients, very few amounts of transmitted coefficients are capable of reconstructing the image to some extent. Moreover, secret information embedded beforehand can be recovered with acceptable rate in correctly extracted bits even experiencing through the lossy channels for data delivery. Simulation results with our algorithm have demonstrated the effectiveness for integrating watermarking into compressive sampling systems.

Prediction-Based Reversible Data Hiding with Content Characteristics

Reversible data hiding is one of the popular topics in watermarking researches, and it belongs to the branch of digital rights management (DRM) applications. Similar to conventional watermarking techniques, for reversible data hiding, at the encoder, secret information can be embedded into original images, and the marked image can be obtained. Different from conventional watermarking, due to the term reversible, it implies that at the decoder, both the original image and embedded secret should be perfectly separated from marked image. It requires keeping the reversibility of proposed algorithm, while looking for good image quality and large amount of secret for embedding. With the prediction-based algorithm, output image can be predicted, and differences between original and predicted images can be altered to make reversible data hiding possible. We also utilize inherent characteristics of original images for reaching better performances. Simulation results reveal that with our algorithm, comparable or better performances can be observed due to the characteristics from different images.

The Design of an XMPP-based Service Integration Scheme

The advances in network-enabled software development raises the demands on the service oriented architecture (SOA). Popular implementations of SOA uses XML-based representations, such as SOAP and WSDL, to enable remote function invocation and service discovery. Our previous experiences show that a sophisticated XML representation is often over-complex for a controlled environment. We also found that the application composition is a high-level operation. Therefore, a proper infrastructure for hosting and delivering various kinds of services is required. We then proposed an XMPP-based solution to overcome these issues. A service is an XMPP client, and focus on how to interact with the other clients, the management of the service can be flexibly arranged through the proposed format of its full JID, and the fundamental security are handled by the XMPP server and the publisher-subscriber pattern.

Electrical Network Frequency as a Tool for Audio Concealment Process

We live in a digital era. Digital contents may be produced by digital equipments or by converting old analog recordings. With the rapid growth of digital contents, digital archiving technology is demanded. Different types of contents require different processing techniques. In this paper, we focus on digital audio contents. The related techniques, such as forensics, authentication, and error concealment, were studied. When converting audio tapes to digital files, sometimes a certain automatic error detection and concealment is needed. However, traditional audio tapes were recorded without any error recovery information. Based on the restriction, we proposed a scheme that incorporates the electrical network frequency (ENF) as a tool for detecting damaged audio segments. The goal is to help people identifying candidate concealment segments. When using in an archiving application, it reduces the manpower as well as increases the accuracy of the generated meta-data.

Adjacent Quad Modification Algorithm for Reversible Data Hiding

Reversible data hiding has been a newly developed branch in data hiding researches. With the definition, when the message is hidden into the original image by some means, the image containing the hidden message is transmitted to the decoder. After decoding, both the original image and the hidden message should both be recovered perfectly. In this paper, we make use of the easily implemented scheme by modifying the histogram of original image, and perform data hiding with the proposed scheme by considering the local statistics in the original image. Simulation results demonstrate the better performance, and hence the effectiveness of the proposed algorithm.

Visual Cryptography for Compressed Sensing of Images with Transmission over Multiple Channels

Visual cryptography is an effective means for ownership protection of multimedia contents. Due to the massive amount of multimedia contents, compression is an inevitable part before transmission over channels. Recently, compressed sensing has attracted much attention with its superior performance in data compression. With compressed sensing, few amounts of coefficients are able to obtain reconstructed image with acceptable quality. In order to protect the ownership of multimedia contents with compressed sensing, we hide visual cryptography information, or shares, into compressed sensing coefficients. Besides, considering the characteristics of visual cryptography, compressed sensing coefficients are split into chunks for delivering over multiple channels. After reception, extracted shares can be verified to protect the ownership. Simulation results have pointed out that our method can be applicable, and integration with robust visual cryptography may lead to enhanced performances.

A Genetic Programming Based Scheme for Combining Image Operators

Sophisticated image processing is usually nonlinear and difficult to model. In addition to the conventional image processing tools, we need some alternatives to bridge the gap between low-level and semantic level computation. This paper presents an idea of image processing scheme. We transform an image into different representations; feed the representations to the proper cellular automaton (CA) components to produce the information images; use the information images as the inputs to the combination program; and finally get the processed result. To identify the needed transforms, the CA transition rules, and the combination expression, we adopt genetic programming (GP) and cellular programming (CP) to search for the configuration. The searched configuration separates the parallelizable and sequential parts of the program. We dont enforce the linearity of the program, and it is likely that the searched result matches to the nonlinear nature of human semantics.

A Study on ENF Discontinuity Detection Techniques

Electrical network frequency (ENF) is known to be useful for audio authentication. In a digital archive system, error concealment of the audio files are sometimes required. Firstly, assuming that traditional recordings were produced by AC-powered devices, the ENF is coupled into the recorded signals. Secondly, assuming that damaging and editing produce similar data error patterns, the ENF phase discontinuities exist in a damaged audio file. To detect phase discontinuity, three phase tracking methods are studied and discussed: DFT over sampling, band pass filtering, and a combined approach. The basic concept of the three methods is to achieve high-precision phase analysis. The cost is the computation complexity. Due to the special properties of ENF, it is likely to reduce the complexity by a certain trade-off design. We then describe the future research directions at the end of this paper.

Preliminary Survey of Multiview Synthesis Technology

With the maturity of digital camera technology, it is feasible to form an array of cameras. The major usage of a camera array is to acquire different views of a scene in one shot. The captured data can be used to analyze the depths of the objects. Once we have the 3D model, we can synthesize virtual views, relight the scene, etc. The potential applications are virtual reality and augmented reality. In order to investigate the multiview technology, we studied the fundamental concepts, including the single lens camera, the eccentric lens camera, the plenoptic camera, and the multiview camera. We also discussed a few application examples for understanding the practical usages. The study showed that the initial depth estimation technology becomes important nowadays in providing photo realistic natural feel to people. The application areas were also extended to entertainment, and also to some crucial tasks such as medical operations and combat missions.

Similarity-Based Watermarking for Sub-Sampling Images with Bacterial Foraging Techniques

Digital watermarking is an effective means for copyright protection or covert communication in the research field of information security. By using watermarking, the watermark bits are hidden into the multimedia content with the designed algorithm. And the watermarked contents, or the watermarked image in this paper, look almost identical to its original counterpart, while it still retains the capability to counteract intentional signal processing by others and keeps the existence of the hidden watermark. We employ the concept of sub-sampling and make use of the frequency coefficient similarities between the sub-images for effectively embedding the watermark. In addition, there must be requirements to keep both the embedded watermark alive and the watermarked contents acceptable. Here we make use of an optimization technique, called bacterial foraging, for obtaining an optimized watermarking algorithm with a properly designed fitness function comprising the possible requirements including the watermarked image quality and the capability for the existence of extracted watermark. Simulation results demonstrate the advantages and effectiveness of the proposed algorithm with the combination of sub-sampling and bacterial foraging.