• Journal of Korea Society of Industrial Information Systems
• /
• v.22 no.5
• /
• pp.17-24
• /
• 2017

Recently Deep-learning, Technologies using Large or Deep Artificial Neural Networks, have Shown Remarkable Performance, and the Increasing Size of the Network Contributes to its Performance Improvement. However, the Increase in the Size of the Neural Network Leads to an Increase in the Calculation Amount, which Causes Problems Such as Circuit Complexity, Price, Heat Generation, and Real-time Restriction. In This Paper, We Propose and Test a Method to Reduce the Number of Network Connections by Effectively Pruning the Redundancy in the Connection and Showing the Difference between the Performance and the Desired Range of the Original Neural Network. In Particular, we Proposed a Simple Method to Improve the Performance by Re-learning and to Guarantee the Desired Performance by Allocating the Error Rate per Layer in Order to Consider the Difference of each Layer. Experiments have been Performed on a Typical Neural Network Structure such as FCN (full connection network) and CNN (convolution neural network) Structure and Confirmed that the Performance Similar to that of the Original Neural Network can be Obtained by Only about 1/10 Connection.

• Journal of Korea Multimedia Society
• /
• v.20 no.8
• /
• pp.1157-1165
• /
• 2017

2-D complex Gabor filtering has found numerous applications in the fields of computer vision and image processing. Especially, in some applications, it is often needed to compute 2-D complex Gabor filter bank consisting of the 2-D complex Gabor filtering outputs at multiple orientations and frequencies. Although several approaches for fast 2-D complex Gabor filtering have been proposed, they primarily focus on reducing the runtime of performing the 2-D complex Gabor filtering once at specific orientation and frequency. To obtain the 2-D complex Gabor filter bank output, existing methods are repeatedly applied with respect to multiple orientations and frequencies. In this paper, we propose a novel approach that efficiently computes the 2-D complex Gabor filter bank by reducing the computational redundancy that arises when performing the Gabor filtering at multiple orientations and frequencies. The proposed method first decomposes the Gabor basis kernels to allow a fast convolution with the Gaussian kernel in a separable manner. This enables reducing the runtime of the 2-D complex Gabor filter bank by reusing intermediate results of the 2-D complex Gabor filtering computed at a specific orientation. Experimental results demonstrate that our method runs faster than state-of-the-arts methods for fast 2-D complex Gabor filtering, while maintaining similar filtering quality.

• Journal of Korean Institute of Industrial Engineers
• /
• v.37 no.1
• /
• pp.64-73
• /
• 2011

Oil price prediction is an important issue for the regulators of the government and the related industries. When employing the time series techniques for prediction, however, it becomes difficult and challenging since the behavior of the series of oil prices is dominated by quantitatively unexplained irregular external factors, e.g., supply- or demand-side shocks, political conflicts specific to events in the Middle East, and direct or indirect influences from other global economical indices, etc. Identifying and quantifying the relationship between oil price and those external factors may provide more relevant prediction than attempting to unclose the underlying structure of the series itself. Technically, this implies the prediction is to be based on the vectoral data on the degrees of the relationship rather than the series data. This paper proposes a novel method for time series prediction of using Semi-Supervised Learning that was originally designed only for the vector types of data. First, several time series of oil prices and other economical indices are transformed into the multiple dimensional vectors by the various types of technical indicators and the diverse combination of the indicator-specific hyper-parameters. Then, to avoid the curse of dimensionality and redundancy among the dimensions, the wellknown feature extraction techniques, PCA and NLPCA, are employed. With the extracted features, a timepointspecific similarity matrix of oil prices and other economical indices is built and finally, Semi-Supervised Learning generates one-timepoint-ahead prediction. The series of crude oil prices of West Texas Intermediate (WTI) was used to verify the proposed method, and the experiments showed promising results : 0.86 of the average AUC.

• Journal of Korea Multimedia Society
• /
• v.11 no.3
• /
• pp.404-413
• /
• 2008

As toward a ubiquitous society, a lot of methods have been proposed to protect personal privacy. Smart Cards with CPU and Memory are widely being used to implement the methods. The use of Java Card is also gradually getting expanded into more various applications. Because there is no standards in Java Card File System, Generally, Java Card File System follows the standards of Smart Card File System. However, one of disadvantages of the Java Card File System using a standard of Smart Card File System is that inefficient memory use and increasing processing time are caused by redundancy of data and program codes. In this paper, a File Cache method and a Direct Access method are proposed to solve the problems. The proposed methods are providing efficient memory use and reduced processing time by reduce a program codes.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.21 no.4
• /
• pp.774-780
• /
• 2017

This paper proposes an efficient binary arithmetic encoder hardware architecture for CABAC encoding, which is an entropy coding method of HEVC. CABAC is an entropy coding method that is used in HEVC standard. Entropy coding removes statistical redundancy and supports a high compression ratio of images. However, the binary arithmetic encoder causes a delay in real time processing and parallel processing is difficult because of the high dependency between data. The operation of the proposed CABAC BAE hardware structure is to separate the renormalization and process the conventional iterative algorithm in parallel. The new scheme was designed as a four-stage pipeline structure that can reduce critical path optimally. The proposed CABAC BAE hardware architecture was designed with Verilog HDL and implemented in 65nm technology. Its gate count is 8.07K and maximum operating speed of 769MHz. It processes the four bin per clock cycle. Maximum processing speed increased by 26% from existing hardware architectures.

• The KIPS Transactions:PartB
• /
• v.13B no.1 s.104
• /
• pp.7-14
• /
• 2006

This paper presents an efficient face recognition method using both first moment of image and basis images. First moment which is a method for finding centroid of image, is applied to exclude the needless backgrounds in the face recognitions by shifting to the centroid of face image. Basis images which are the face features, are respectively extracted by principal component analysis(PCA) and fixed-point independent component analysis(FP-ICA). This is to improve the recognition performance by excluding the redundancy considering to second- and higher-order statistics of face image. The proposed methods has been applied to the problem for recognizing the 48 face images(12 persons*4 scenes) of 64*64 pixels. The 3 distances such as city-block, Euclidean, negative angle are used as measures when match the probe images to the nearest gallery images. The experimental results show that the proposed methods has a superior recognition performances(speed, rate) than conventional PCA and FP-ICA without preprocessing, the proposed FP-ICA has also better performance than the proposed PCA. The city-block has been relatively achieved more an accurate similarity than Euclidean or negative angle.

• Journal of Korea Multimedia Society
• /
• v.9 no.1
• /
• pp.33-40
• /
• 2006

This paper presents an efficient face recognition method using both first moment of image and eigenvector. First moment is a method for finding centroid of image, which is applied to exclude the needless backgrounds in the face recognitions by shitting to the centroid of face image. Eigenvector which are the basis images as face features, is extracted by principal component analysis(PCA). This is to improve the recognition performance by excluding the redundancy considering to second-order statistics of face image. The proposed methods has been applied to the problem for recognizing the 60 face images(15 persons *4 scenes) of 320*243 pixels. The 3 distances such as city-block, Euclidean, negative angle are used as measures when match the probe images to the nearest gallery images. In case of the 45 face images, the experimental results show that the recognition rate of the proposed methods is about 1.6 times and its the classification is about 5.6 times higher than conventional PCA without preprocessing. The city-block has been relatively achieved more an accurate classification than Euclidean or negative angle.

• Journal of Korea Multimedia Society
• /
• v.5 no.2
• /
• pp.167-175
• /
• 2002

Image Mosaicing creates a new image by composing video frames or still images that are related, and performed by arrangement, composition and redundancy analysis of images. This paper proposes a hierarchical image mosaicing system using camera and object parameters far efficient video database construction. A tree-based image mosiacing has implemented for high-speed computation time and for construction of static and dynamic image mosaic. Camera parameters are measured by using least sum of squared difference and affine model. Dynamic object detection algorithm has proposed for extracting dynamic objects. For object extraction, difference image, macro block, region splitting and 4-split detection methods are proposed and used. Also, a dynamic positioning method is used for presenting dynamic objects and a blurring method is used for creating flexible mosaic image.

• The Journal of the Korea Contents Association
• /
• v.8 no.11
• /
• pp.94-102
• /
• 2008

The development of digital technology constructs meaning through the narrative by digitalized story. Including mobile whole aspect movie, visual contents have the narrative structure and signification based on the digital technology. This study planned to explore the possibility that mobile movie can expand the classical narrative structure and bring up the method of signification. For this purpose, this paper proved the process to cohere narrative through redundancy and a double across of space-time structure according to the principle of causality in mobile movies such as and . According to the findings, mobile movies appeared to construct a consolidated structure through cohesion of meaning and to move the space-time doubly across at a subjective point of view of the characters for continuity of story. Such a result means the narrative structure and the method of signification of mobile movie are modifying or expanding the unique formation of classical films in accordance with mobile devices as digital media.

• Journal of Broadcast Engineering
• /
• v.1 no.1
• /
• pp.67-73
• /
• 1996

In general, the high quality audio coding(HQAC) has the structure of the convertional data compression techniques combined with moodels of human perception. The primary auditory characteristic applied to HQAC is the masking effect in the spectral domain. Therefore spectral techniques such as the subband coding or the transform coding are widely used[1][2]. However no effort has yet been made to apply the temporal masking effect and temporal redundancy removing method in HQAC. The audio data compression method proposed in this paper eliminates statistical and perceptual redundancies in both temporal and spectral domain. Transformed audio signal is divided into packets, which consist of 6 frames. A packet contains 1536 samples($256{\times}6$) :nd redundancies in packet reside in both temporal and spectral domain. Both redundancies are elminated at the same time in each packet. The psychoacoustic model has been improved to give more delicate results by taking into account temporal masking as well as fine spectral masking. For quantization, each packet is divided into subblocks designed to have an analogy with the nonlinear critical bands and to reflect the temporal auditory characteristics. Consequently, high quality of reconstructed audio is conserved at low bit-rates.