Kyu Tae Park

A fast algorithm suitable for DCT implementation with integer multiplication

The discrete cosine transform (DCT) has wide applications in speech and image data compression. We propose a novel fast DCT algorithm. The proposed algorithm is based on the symmetric property of the DCT kernel matrix. This decomposes an N-point DCT into an N/2-point DCT and can be iterated further. Because it has fewer multiplication steps, when converting the algorithm into hardware implementation with integer computation, we can reduce the error.

A hierarchical layered model for DVD authoring system

A hierarchical layered model for the implementation of DVD authoring system is proposed in this paper. By employing the hierarchical model, each object can be handled separately in the higher level of the hierarchy and the DVD format-specific bit stream can be generated in the lower levels. An object building tool that maintains a list of the hierarchical dependencies among DVD objects together with the actions to be applied to these objects is also proposed to ensure the consistency of the objects between the different layers of the hierarchy, and for the verification of the produced title sets, a reference player model is proposed. The issues in implementing a model for the DVD authoring system are discussed, and the needs of the authoring system model are addressed from the viewpoint of object encapsulation and data formatting. The proposed model provides an efficient structure in terms of adaptability and flexibility for DVD authoring.

Wavelet transform image compression using human visual characteristics and a tree structure with a height attribute

This paper proposes an image coding algorithm using a tree structure which takes advantage of the geometric similarity among the hierarchical bands of the wavelet transform (WT). A concept of conditional height difference is introduced in order to further reduce the actual number of bits to transmit the tree structure with a height attribute. This paper also proposes a method to appropriately incorporate human visual characteristics into the proposed coding algorithm. This is achieved by determining the human visual weight of each frequency band into which WT decomposes an image. Simulations are used to show the muchimproved performance of the above scheme over the well-known compression algorithms both in terms of peak signal-to-noise ratio (PSNR) and in other senses. For example, the 512x 512 “Lena” image is shown to be compressed at 1 bit/pixel with a PSNR higher than 40 dB, with less hardware complexity.

Wavelet transform image coding using human visual system

Human visual system (HVS) is applied to wavelet transform in this paper. Wavelet transform decomposes the spatial frequency domain to octave band scale and this process is similar to that of HVS. So the wavelet transform and modulation transfer function (MTF) of HVS can be optimally coupled in order to gain coding efficiency. Several quantizing stepsizes are determined for each subband by integral ratio which is from frequency response of wavelet filter and MTF. Compared with JPEG, the result shows superiority.

Filtration and inactivation of aerosolized bacteriophage MS2 by a CNT air filter fabricated using electro-aerodynamic deposition

Abstract Carbon nanotubes (CNTs) were coated on a sample of glass fiber air filter medium at atmospheric pressure and room temperature using electro-aerodynamic deposition (EAD). In the EAD method, CNTs (diameter: 50nm, length: 2–3μm) were aerosolized, electrically charged, and injected through a nozzle. A voltage was applied externally between the ground nozzle and a planar electrode on which the sample was located. The charged CNTs were deposited on the sample in a vertically standing posture even at a low flow velocity. Before the deposition experiment, a calculation was performed to determine the applied voltage by simulating the electric field, flow field, and particle trajectory. Using CNT-coated filter samples, virus aerosol filtration and anti-viral tests were carried out using the aerosol number counting method and the plaque counting method, respectively. For this purpose, bacteriophage MS2 was aerosolized with an atomizer. The particle filtration efficiency was increased to 33.3% in the most penetration particle size zone (100nm) and the antiviral efficiency of the CNT filter was 92% when the coating areal density was 1.5×109 #/cm2. The susceptibility constant of virus to CNTs was 0.2cm2/μg.

CCD noise filtering based on 3-dimensional nonlinear partial differential equation

A three dimensional anisotropic diffusion equation is proposed to remove noise in video sequences. The three dimensional anisotropic diffusion equation utilizes the fact that consecutive frames of high correlation can be obtained in video sequences. It is shown that the three dimensional diffusion equation gives better results than the two dimensional diffusion equation which only deals with still images and slowly smoothes the edge boundaries.

Adaptive approximation bounds for vertex based contour encoding

When approximating the shape of a region, a fixed bound on the tolerable distortion is set for approximating its contour points. An adaptive approximation bound for lossy coding of the contour points is proposed. A function representing the relative significance of the contour points is defined to adjust the distortion bound along the region contour allowing an adaptive approximation of the region shape. The effectiveness of the adaptive contour coding approach for a region-based coding system is verified through experiments.

The Block-based Preprocessing System For The Coding Performance Improvement

In general, most images get to contain unwanted noise in the process of image acquisition and such noise may deteriorate the performance of standard coding systems. Therefore, by pre-reducing inborn noise in images, the coding performance can be improved significantly. In order to improve the coding performance of the standard codecs, a new preprocessing system using spatially adaptive filtering is proposed. In this approach, the preprocessing system is implemented through block region classification and adaptive morphological filtering. Experimental results show that the proposed algorithm obtains significant improvement for standard still and moving images as well as acquired noisy images, as compared with applying no preprocessing. The proposed system can be easily applied to the standard codecs such as JPEG, MPEG, etc., and significantly improve the quality of decoded images without increasing bit-rate.

Maskless, site-selective, nanoaerosol deposition via electro-aerodynamic jet to enhance the performance of flexible Ag-grid transparent electrodes

Grid-type transparent electrodes (TEs) have been fabricated using various techniques, but enhancing the performance of TEs remains challenging. In this study, we developed an aerosol deposition methodology whereby airborne silver nanoparticles of 32 nm diameter were electrically charged and delivered through a nozzle to a pre-patterned silver grid with periodic silver lines (width 18 μm, thickness 360 nm, and pitch 500 μm). A high voltage of 7 kV was applied between the nozzle and the grid, located 10 mm away from the nozzle, such that charged silver nanoparticles were deposited on the grid without the requirement for masking. The pre-patterned silver grid was prepared with AC pulse applied electrohydrodynamic jet printing on a PET film. After 3 minutes of deposition, the grid thickness increased from 360 nm to 587 nm, resulting in decreased TE sheet resistance from 7.38 Ω sq−1 to 1.95 Ω sq−1 while transmittance was kept constant at 84% (σDC/σOP ratio increased from 300 to 1050).

Existence and classification of Hadamard matrices

Hadamard matrices have been widely used in image processing. There are various constructions for Hadamard matrices: Kronecker product, Paley type I and II, Williamson type, etc. We investigate those of the Williamson type, and suggest a search algorithm for Hadamard matrices of the Williamson type. Using this we searched the Hadamard matrices of order 92. It took about 6.8 hours on a DEC Alpha PC. Based on the data for the order 92, we give an estimate for the exhaustive computer search for Hadamard matrices of Williamson type of order 428.