• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.27 no.7A
• /
• pp.705-710
• /
• 2002

In this paper, we present a method to find motion vectors in frequency domain for video data compression. The proposed algorithm is based on the Format Number Transform (FNT), and it declares the most correlated-block as the best matching block, as opposed to declaring the block with least sum of differences between blocks. We show that the proposed method is equivalent to declaring the block with the minimum L2-norm as the best matching block. Unlike other previous fast algorithms, the time requirement for the proposed algorithm does not defend on the image type for finding the optimum solution.

• Proceedings of the Korea Society for Industrial Systems Conference
• /
• 2003.11a
• /
• pp.621-626
• /
• 2003

Mining association rules in web log files can be divided into two steps: 1) discovering frequent item sets in web data; 2) extracting association rules from the frequent item sets found in the previous step. This paper suggests an algorithm for finding frequent item sets efficiently The essence of the proposed algorithm is to transform transaction data files into matrix format. Our experimental results show that the suggested algorithm outperforms the Apriori algorithm, which is widely used to discover frequent item sets, in terms of scan frequency and execution time.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.66 no.10
• /
• pp.1508-1515
• /
• 2017

This paper proposes the defect detection method for automatic inspection machines in printed circuit boards (PCBs) manufacturing system. The defects of PCB such as open, short, pin hole and scratch can be detected by comparing the standard image and the target image. The standard image is obtained from CAD file such as ODB++ format, and the target image is obtained by arranging, filtering and binarization of captured PCB image. Since the PCB size is too large and image resolution is too high, the image processing requires a lot of memory and computational time. The wavelet transform is applied to compress the standard and target images, which results in reducing the memory and computational time. To increase the inspection accuracy, we utilize the he HH-domain as well as LL-domain of the transformed images. Experimental results are finally presented to show the performance improvement of the proposed method.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.37A no.11
• /
• pp.993-1000
• /
• 2012

In this research, we proposed an efficient VLSI architecture of the lifting-based 2D DWT (Discrete Wavelet Transform) filter with 100% hardware utilization. The (9,7) filter structure has been applied and extendable to the filter length. We proposed a new block-based scheduling that computes the DWT for the lower levels on an "as-early-as-possible" basis, which means that the calculation for the lower level will start as soon as the data is ready. Since the proposed 2D DWT computes the outputs of all levels by one row-based scan, the intermediate results for other resolution levels should be kept in storage such as the Data Format Converter (DFC) and the Delay Control Unit (DCU) until they are used. When the size of input image is $N{\times}N$ and m is the filter length, the required storage for the proposed architecture is about 2mN. Since the proposed architecture processes the 2D DWT in horizontal and vertical directions at the same time with 4 input data, the total period for 2D DWT is $N^2(1-2^{-2J})/3$.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.17 no.5
• /
• pp.1413-1432
• /
• 2023

Recommender systems aim to recommend items to users by taking into account their probable interests. This study focuses on creating a model that utilizes multiple sources of information about users and items by employing a multimodality approach. The study addresses the task of how to gather information from different sources (modalities) and transform them into a uniform format, resulting in a multi-modal feature description for users and items. This work also aims to transform and represent the features extracted from different modalities so that the information is in a compatible format for integration and contains important, useful information for the prediction model. To achieve this goal, we propose a novel multi-modal recommendation model, which involves extracting latent features of users and items from a utility matrix using matrix factorization techniques. Various transformation techniques are utilized to extract features from other sources of information such as user reviews, item descriptions, and item categories. We also proposed the use of Principal Component Analysis (PCA) and Feature Selection techniques to reduce the data dimension and extract important features as well as remove noisy features to increase the accuracy of the model. We conducted several different experimental models based on different subsets of modalities on the MovieLens and Amazon sub-category datasets. According to the experimental results, the proposed model significantly enhances the accuracy of recommendations when compared to SVD, which is acknowledged as one of the most effective models for recommender systems. Specifically, the proposed model reduces the RMSE by a range of 4.8% to 21.43% and increases the Precision by a range of 2.07% to 26.49% for the Amazon datasets. Similarly, for the MovieLens dataset, the proposed model reduces the RMSE by 45.61% and increases the Precision by 14.06%. Additionally, the experimental results on both datasets demonstrate that combining information from multiple modalities in the proposed model leads to superior outcomes compared to relying on a single type of information.

• Imaging Science in Dentistry
• /
• v.34 no.3
• /
• pp.117-122
• /
• 2004

Purpose : To determine the proper image compression method and ratio without image quality degradation in intraoral digital radiographic images, comparing the discrete cosine transform (DCT)-based JPEG with the wavelet-based JPEG 2000 algorithm. Materials and Methods : Thirty extracted sound teeth and thirty extracted teeth with occlusal caries were used for this study. Twenty plaster blocks were made with three teeth each. They were radiographically exposed using CDR sensors (Schick Inc., Long Island, USA). Digital images were compressed to JPEG format, using Adobe Photoshop v.7.0 and JPEG 2000 format using Jasper program with compression ratios of 5 : 1,9 : 1, 14 : 1,28 : 1 each. To evaluate the lesion detectability, receiver operating characteristic (ROC) analysis was performed by the three oral and maxillofacial radiologists. To evaluate the image quality, all the compressed images were assessed subjectively using 5 grades, in comparison to the original uncompressed images. Results: Compressed images up to compression ratio of 14 : 1 in JPEG and 28 : 1 in JPEG 2000 showed nearly the same the lesion detectability as the original images. In the subjective assessment of image quality, images up to compression ratio of 9 : 1 in JPEG and 14 : 1 in JPEG 2000 showed minute mean paired differences from the original Images. Conclusion : The results showed that the clinically acceptable compression ratios were up to 9 : 1 for JPEG and 14 : 1 for JPEG 2000. The wavelet-based JPEG 2000 is a better compression method, comparing to DCT-based JPEG for intraoral digital radiographic images.

• Journal of Digital Contents Society
• /
• v.12 no.1
• /
• pp.77-83
• /
• 2011

Providing services in a wireless environment with existing wired-based GIS solutions have many limitations such as slow communication, processing rates, and screen size. Among the components, data is considered to be the most important factor in GIS. In this paper, I describe MVF, the most recent technology pioneered by GIS researchers. MVF is a format for efficient usage of spatial object in order to express, save and transform in mobile GIS. First, I describe the definition of MVF syntax, then the implementation of Header Section Generation Module and Main Section Generation Module through the API. Since the issues of network traffic and Hangul processing in mobile GIS are important, I especially consider development of MVF in terms of spatial data size.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.26 no.12A
• /
• pp.2074-2084
• /
• 2001

la this paper, a new digital watermarking algorithm using wavelet transform in frequency domain is suggested. The wavelet coefficients of low frequency subband are utilized to embed the watermark, After the original image is transformed using discrete wavelet transform, their coefficients are transformed into efficient1y in Sequency domain. DCT and FFT transforms are utilized in this processing. Watermark image of general image format is transformed using DCT and the hiding watermark into wavelet coefficients is equally distributed in frequency domain. Next, these wavelet coefficients are performed with inverse transform. The detection process of watermark is performed with reverse direction to insertion process. In this paper, we developed core watermark technologies which are a data hiding technology to hide unique logo mark which symbolizes the copyright and a robust protection technology to protect logo data from external attack like as compression, filtering, resampling, cropping. The experimental results show that two suggested watermarking technologies are invisible and robust.

• Journal of the Optical Society of Korea
• /
• v.19 no.4
• /
• pp.334-339
• /
• 2015

Terabit-per-second (Tb/s) transmission capacity for the next generation of long-haul communication networks can be achieved using multicarrier optical super-channel technology. In an elastic orthogonal frequency division multiplexing (OFDM) super-channel transmission system, demultiplexing a portion of an entire spectrum in the form of a subband with minimum power is critically required. A major obstacle to achieving this goal is the analog-to-digital converter (ADC), which is power-hungry and extremely expensive. Without a proper ADC that can work with low power, it is unrealistic to design a 100G coherent receiver suitable for a commercially deployable optical network. Discrete Fourier transform (DFT) is often seen as a primary technique for understanding partial demultiplexing, which can be attained either optically or electronically. If fairly comparable performance can be achieved with an all-optical DFT circuit, then a solution independent of data rate and modulation format can be obtained. In this paper, we investigate two distinct OFDM super-channel receiver models, based on electronic and all-optical DFT-technologies, for partial carrier demultiplexing in a multi-Tb/s transmission system. The performance comparison of the receivers is discussed in terms of bit-error-rate (BER) performance.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.18 no.5
• /
• pp.217-223
• /
• 2023

In this paper, we propose a new model for the conditional generation of music, considering key and rhythm, fundamental elements of music. MIDI sheet music is converted into a WAV format, which is then transformed into a Mel Spectrogram using the Short-Time Fourier Transform (STFT). Using this information, key and rhythm details are classified by passing through two Convolutional Neural Networks (CNNs), and this information is again fed into the Music Transformer. The key and rhythm details are combined by differentially multiplying the weights and the embedding vectors of the MIDI events. Several experiments are conducted, including a process for determining the optimal weights. This research represents a new effort to integrate essential elements into music generation and explains the detailed structure and operating principles of the model, verifying its effects and potentials through experiments. In this study, the accuracy for rhythm classification reached 94.7%, the accuracy for key classification reached 92.1%, and the Negative Likelihood based on the weights of the embedding vector resulted in 3.01.