• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.29B no.10
• /
• pp.90-99
• /
• 1992

In this paper, we propose a neural network model for speech recognition. The model consists of feature extraction parts and recognition parts. The interconnection model based on ${\Delta}^2$G operator was used for frequency analysis. Two features, global feature and local feature, were extracted from this model. Recognition parts consist of global grouping stage and local grouping stage. When the input pattern was coded by slope method, the recognition rate of speakers, A and B, was 100%. When the test was performed with the data of 9 speakers, the recognition rate of 91.4% was obtained.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.30 no.3B
• /
• pp.65-71
• /
• 2005

The primary function of the Internet routers is to forward incoming packets toward their final destinations. IP address lookup is one of the most important functions in evaluating router performance since IP address lookup should be performed in wire-speed for the hundred-millions of incoming packets per second. With CIDR, the IP prefixes of routing table have arbitrary lengths, and hence address lookup by exact match is no longer valid. As a result, when packets arrive, routers compare the destination IP addresses of input packets with all prefixes in its routing table and determine the most specific entry among matching entries, and this is called the longest prefix matching. In this paper, based on parallel multiple hashing and prefix grouping, we have proposed a hardware architecture which performs an address lookup with a single memory access.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.35 no.2A
• /
• pp.112-120
• /
• 2010

In this paper, we propose semi-ML adaptive parallel QRDM (APQRDM) and iterative QRDM (AIQRDM) algorithms based on set grouping. Using the set grouping, the tree-search stage of QRDM algorithm is divided into partial detection phases (PDP). Therefore, when the treesearch stage of QRDM is divided into 4 PDPs, the APQRDM latency is one fourth of that of the QRDM, and the hardware requirements of AIQRDM is approximately one fourth of that of QRDM. Moreover, simulation results show that in $4{\times}4$ system and at Eb/N0 of 12 dB, APQRDM decreases the average computational complexity to approximately 43% of that of the conventional QRDM. Also, at Eb/N0 of 0dB, AIQRDM reduces the computational complexity to about 54% and the average number of metric comparisons to approximately 10% of those required by the conventional QRDM and AQRDM.

• Journal of Convergence for Information Technology
• /
• v.12 no.3
• /
• pp.38-45
• /
• 2022

Due to the development of fintech technology, financial transactions using smart phones are being activated. The password for user authentication during financial transactions is entered through the virtual keypad displayed on the screen of the smart phone. When the password is entered, the attacker can find out the password by capturing it with a high-resolution camera or spying over the shoulder. A virtual keypad with security applied to prevent such an attack is difficult to input on a small touch-screen, and there is still a vulnerability in peeping attacks. In this paper, the entire keypad is divided into several groups and displayed on a small screen, touching the group to which the character to be input belongs, and then touching the corresponding character within the group. The proposed method selects the group to which the character to be input belongs, and displays the keypad in the group on a small screen with no more than 10 keypads, so that the size of the keypad can be enlarged more than twice compared to the existing method, and the location is randomly placed, hence location of the touch attacks can be blocked.

• Journal of the Institute of Electronics Engineers of Korea SD
• /
• v.37 no.8
• /
• pp.86-96
• /
• 2000

The representative area among the ones whose cost increases as the integration ratio increases is the test area. As the relative cost of hardware decreases, the BIST method has been focued on as the future-oriented test method. The biggest drawback of it is the increasing test time to obtain the acceptable fault coverage. This paper proposed a BIST implementation method to reduce the test times. This method uses an input grouping and test point insertion method, in which the definition of test point is different from the previous one. That is, the test points are defined on the basis of the internal nodes which are the reference points of the input grouping and are merging points of the grouped signals. The main algorithms in the proposed method were implemented with C-language, and various circuits were used to apply the proposed method for experiment. The results showed that the test time could be reduced to at most $1/2^{40}$ of the pseudo-random pattern case and the fault coverage were also increased compared with the conventional BIST method. The relative hardware overhead of the proposed method to the circuit under test decreases as th e size of the circuit to be tested increases, and the delay overhead by the BIST utility is negligible compared to that of the original circuit. That means, the proposed method can be applied efficiently to large VLSI circuits.

• Proceedings of the Optical Society of Korea Conference
• /
• 1990.02a
• /
• pp.95-101
• /
• 1990

Parallel optical arithmetic techniques have been developed using the correlation property of optical phase conjugate wave generated by degenerated four wave-mixing. In this paper, conventional rectangular-type coded pattern used for optical logic system is replaced by circular one for effective beam coupling in a photorefractive $BaTiO_3$ material. By adequately adjusting the distance between circular-type pixels of the input pattern and grouping the correlated output, optical binary half addition/subtraction, binary multiplication and, matrix-matrix computation are demonstrated.

• Journal of the Korean Institute of Telematics and Electronics
• /
• v.24 no.6
• /
• pp.1080-1086
• /
• 1987

SLAGEN, a system for symbolic cell layout based on artificial intelligence approach, takes as input a transistor connection description of CMOS standard cells and environment information, and outputs a symbolic layout description. SLAGEN performas transistor grouping by a heuristic search method, in order to minimize the number of separations, and then performs group reordering and transistor reordering with an eye toward minimizing routing. Next, SLAGEN creates a rough initial routing in order to guarantee functionality and correctness, and then improve the initial routing by a rule-based approach.

• 제어로봇시스템학회:학술대회논문집
• /
• 2000.10a
• /
• pp.365-365
• /
• 2000

In this paper, we describe the system to recognize the six digit postal number of mails using neural network. Our zip-code recognition system consists of a preprocessing procedure for the original captured image, a segmentation procedure for separating an address block area with a shape, and recognition procedure for the cognition of a postal number. we extract the feature vectors that are the input of a neural network for the recognition process based on an area optimizing and an image thinning processing. The neural network classifies the zip-code in the mail and the recognized zip-code is verified through the zip-code database.

• Smart Media Journal
• /
• v.6 no.4
• /
• pp.32-40
• /
• 2017

In this paper, a robust text detection method based on edge enhanced contrasting extremal region (CER) is proposed using stroke width transform (SWT) and tensor voting. First, the edge enhanced CER extracts a number of covariant regions, which is a stable connected component from input images. Next, SWT is created by the distance map, which is used to eliminate non-text regions. Then, these candidate text regions are verified based on tensor voting, which uses the input center point in the previous step to compute curve salience values. Finally, the connected component grouping is applied to a cluster closed to characters. The proposed method is evaluated with the ICDAR2003 and ICDAR2013 text detection competition datasets and the experiment results show high accuracy compared to previous methods.

• Proceedings of the IEEK Conference
• /
• 2003.07e
• /
• pp.2455-2458
• /
• 2003

In this paper, the Classification of Energy Labeling has been Proposed. Energy Parameters of input signal which is extracted from each phoneme is labelled. And groups of labelling according to detected energies of input signals are detected. Next, DTW processes in a selected group of labeling. This leads to DTW processing faster than a previous algorithm. In this Method, because an accurate detection of parameters is necessary on the assumption in steps of a detection of speeching duration and a detection of energy parameters, variable windows which are decided by pitch period is used. Extract algorithms don't search for exact frame energy, because 256 frame window-sizes is fixed. For this reason, a new energy extraction method has been proposed. A pitch period is detected firstly; next window scale is decided between 200 frames and 300 frames. The proposed method make it possible to cancel an influence of windows.